{"id":9195,"date":"2020-07-29T01:12:18","date_gmt":"2020-07-29T01:12:18","guid":{"rendered":"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/?p=9195"},"modified":"2020-09-18T15:19:37","modified_gmt":"2020-09-18T15:19:37","slug":"20200729-%e4%bd%9c%e6%a5%ad%e3%83%a1%e3%83%a2-nvlink%e3%82%92%e5%88%a9%e7%94%a8%e3%81%99%e3%82%8b","status":"publish","type":"post","link":"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/?p=9195","title":{"rendered":"(20200729 \u4f5c\u696d\u30e1\u30e2) NVLink\u3092\u5229\u7528\u3059\u308b"},"content":{"rendered":"<p><a href=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/DSCN0874.jpg\"><img loading=\"lazy\" decoding=\"async\" src=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/DSCN0874-300x225.jpg\" alt=\"\" width=\"300\" height=\"225\" class=\"aligncenter size-medium wp-image-9220\" srcset=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/DSCN0874-300x225.jpg 300w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/DSCN0874-768x576.jpg 768w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/DSCN0874-1024x768.jpg 1024w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/DSCN0874-600x450.jpg 600w\" sizes=\"auto, (max-width: 300px) 100vw, 300px\" \/><\/a><\/p>\n<p>NVLink\u3092\u5229\u7528\u3057\uff0c\u30d3\u30c7\u30aa\u30ab\u30fc\u30c9\u30922\u679a\u523a\u3057\u3066\u8a08\u7b97\u3057\u3066\u307f\u305f\u3044\u3002\u4ee5\u4e0b\uff0c\u4e3b\u306a\u30d1\u30fc\u30c4\u3092\u3042\u3052\u307e\u3059\u3002<br \/>\n(20200918 \u7d50\u5c40\u4eca\u306e\u3068\u3053\u308d NVLink \u306f\u4e0a\u624b\u304f\u5229\u7528\u3067\u304d\u3066\u3044\u307e\u305b\u3093\u3002)<\/p>\n<ul>\n<li>CPU : Core i5 9500<\/li>\n<li>MB  : ASRock Z390 Extreme4<\/li>\n<li>GPU : RTX 2070 Super X 2 (ASUS DUAL-RTX2070S-O8G-EVO)<\/li>\n<li>NVLink : ASUS ROG-NVLINK-3<\/li>\n<li>Case : SilverStone SST-RM400<\/li>\n<li>\u96fb\u6e90 : Seasonic SSR-850FX<\/li>\n<\/ul>\n<p>\u3042\u307e\u308a\u306b\u30822\u679a\u306e\u30ab\u30fc\u30c9\u306e\u9593\u306b\u9699\u9593\u304c\u7121\u304b\u3063\u305f\u306e\u3067\uff0c\u30d3\u30c7\u30aa\u30ab\u30fc\u30c9\u306e\u30ab\u30d0\u30fc\u3092\u7247\u65b9\u5916\u3057\u307e\u3057\u305f\u3002\u30b1\u30fc\u30b9\u30d5\u30a1\u30f3\u306e\u98a8\u3092\u76f4\u63a5\u3042\u3066\u3066\u51b7\u5374\u3092\u88dc\u52a9\u3057\u3066\u3044\u307e\u3059\u3002<\/p>\n<p><a href=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/DSCN0861.jpg\"><img loading=\"lazy\" decoding=\"async\" src=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/DSCN0861-300x225.jpg\" alt=\"\" width=\"300\" height=\"225\" class=\"aligncenter size-medium wp-image-9222\" srcset=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/DSCN0861-300x225.jpg 300w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/DSCN0861-768x576.jpg 768w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/DSCN0861-1024x768.jpg 1024w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/DSCN0861-600x450.jpg 600w\" sizes=\"auto, (max-width: 300px) 100vw, 300px\" \/><\/a><\/p>\n<p>\u4ed6\u306e\u89d2\u5ea6\u306e\u3082\u306e\u3082\u3042\u3052\u307e\u3059\u3002<\/p>\n<p><a href=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/DSCN0872.jpg\"><img loading=\"lazy\" decoding=\"async\" src=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/DSCN0872-300x225.jpg\" alt=\"\" width=\"300\" height=\"225\" class=\"aligncenter size-medium wp-image-9226\" srcset=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/DSCN0872-300x225.jpg 300w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/DSCN0872-768x576.jpg 768w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/DSCN0872-1024x768.jpg 1024w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/DSCN0872-600x450.jpg 600w\" sizes=\"auto, (max-width: 300px) 100vw, 300px\" \/><\/a><\/p>\n<p><a href=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/DSCN0864.jpg\"><img loading=\"lazy\" decoding=\"async\" src=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/DSCN0864-300x225.jpg\" alt=\"\" width=\"300\" height=\"225\" class=\"aligncenter size-medium wp-image-9228\" srcset=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/DSCN0864-300x225.jpg 300w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/DSCN0864-768x576.jpg 768w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/DSCN0864-1024x768.jpg 1024w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/DSCN0864-600x450.jpg 600w\" sizes=\"auto, (max-width: 300px) 100vw, 300px\" \/><\/a><\/p>\n<p><a href=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/DSCN0869.jpg\"><img loading=\"lazy\" decoding=\"async\" src=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/DSCN0869-300x225.jpg\" alt=\"\" width=\"300\" height=\"225\" class=\"aligncenter size-medium wp-image-9224\" srcset=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/DSCN0869-300x225.jpg 300w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/DSCN0869-768x576.jpg 768w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/DSCN0869-1024x768.jpg 1024w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/DSCN0869-600x450.jpg 600w\" sizes=\"auto, (max-width: 300px) 100vw, 300px\" \/><\/a><\/p>\n<p>\u4e0b\u8a18\u306e\u30b5\u30a4\u30c8\u306b\u5f93\u3063\u3066\u4f5c\u696d\u3057\u305f\u3002<\/p>\n<li>\u53c2\u8003\u30b5\u30a4\u30c8\uff1a<a href=\"https:\/\/qiita.com\/Brutus\/items\/2db051ec2bcfa726cc2f\">Ubuntu\u3067GPU\u30de\u30b7\u30fc\u30f3\u3092\u69cb\u7bc9\u3059\u308b<\/a><\/li>\n<p><\/p>\n<p>\u3068\u304f\u306b\u4eca\u306e\u3068\u3053\u308d NVLink \u306b\u5bfe\u5fdc\u3057\u305f\u4f5c\u696d\u304c\u7121\u3044\u3002\u8272\u3005\u306a\u30b3\u30de\u30f3\u30c9\u306e\u5fdc\u7b54\u304b\u3089\u306f2\u679a\u3068\u3082\u8a8d\u8b58\u3055\u308c\u3066\u3044\u308b\u3053\u3068\u306f\u5206\u304b\u308b\u304c\uff0cNVLink \u304c\u6a5f\u80fd\u3057\u3066\u3044\u308b\u3053\u3068\u3092\u78ba\u304b\u3081\u308b\u306b\u306f\u3069\u3046\u3057\u305f\u3089\u826f\u3044\u306e\u304b\uff1f<\/p>\n<p>\u6e29\u5ea6\u3092\u30e2\u30cb\u30bf\u30fc\u3057\u305f\u3044\u3068\u601d\u3063\u3066\uff0cPsensor \u3068\u3044\u3046\u30bd\u30d5\u30c8\u3092\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u3057\u305f\u3002<\/p>\n<li>\u53c2\u8003\u30b5\u30a4\u30c8\uff1a<a href=\"https:\/\/dlrecord.hatenablog.com\/entry\/2017\/11\/28\/180027\">ubuntu\u3067GPU, CPU\u6e29\u5ea6\u306a\u3069\u3092\u30b0\u30e9\u30d5\u30a3\u30ab\u30eb\u306b\u76e3\u8996\u3059\u308bpsensor\u3092\u4f7f\u3046<\/a><\/li>\n<p><\/p>\n<p>\u4e0b\u56f3\u304c\u8a08\u7b97\u4e2d\u306e\u89b3\u6e2c\u3002\u6e29\u5ea6\u4e0a\u6607\u306b\u3084\u3084\u9045\u308c\u3066\uff0c\u30d5\u30a1\u30f3\u306e\u56de\u8ee2\u6570\u304c\u4e0a\u6607\u3057\u3066\u3044\u308b\u3002\u8a08\u6e2c\u3067\u304d\u308b\u306e\u306f\uff0c\u30d3\u30c7\u30aa\u30ab\u30fc\u30c9\u306e\u6e29\u5ea6\uff0c\u30d5\u30a1\u30f3\u306e\u56de\u8ee2\u6570\uff0c\u30e1\u30e2\u30ea\u30fc\u306e\u4f7f\u7528\u72b6\u6cc1\u306a\u3069\u3067\u3042\u308b\u3002<\/p>\n<p><a href=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200730a.png\"><img loading=\"lazy\" decoding=\"async\" src=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200730a-300x155.png\" alt=\"\" width=\"300\" height=\"155\" class=\"aligncenter size-medium wp-image-9201\" srcset=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200730a-300x155.png 300w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200730a-768x397.png 768w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200730a-1024x530.png 1024w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200730a-600x310.png 600w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200730a.png 1487w\" sizes=\"auto, (max-width: 300px) 100vw, 300px\" \/><\/a><\/p>\n<p>\u30e1\u30e2\u30ea\u30fc\u3084\u6e29\u5ea6\u306e\u4e0a\u6607\u304b\u3089\u898b\u3066\uff0c2\u679a\u76ee\u306e\u30d3\u30c7\u30aa\u30ab\u30fc\u30c9\u306f\u5229\u7528\u3055\u308c\u3066\u3044\u306a\u3044\u3002\u3082\u3046\u5c11\u3057\u5927\u304d\u306a\u8a08\u7b97\u3092\u3084\u3063\u3066\u307f\u305f\u65b9\u304c\u3088\u3055\u305d\u3046\u3002\u4e0a\u8a18\u306e\u8a08\u7b97\u306f CUPY \u3092\u5229\u7528\u3057\u305f\u3082\u306e\u3067\u3042\u308b\u3002NVLink \u304c\u6a5f\u80fd\u3059\u308b\u3053\u3068\u3068 CUPY \u304cNVLink \u4e0a\u3067\u5229\u7528\u53ef\u80fd\u3068\u306a\u308b\u307e\u3067\u306b\u306f\u304b\u306a\u308a\u9694\u305f\u308a\u304c\u3042\u308b\u306e\u304b\u3082\u3057\u308c\u306a\u3044\u3002<\/p>\n<p>(20200803)<br \/>\nNVLink \u306b\u95a2\u3059\u308b\u5fdc\u7b54\u30e1\u30c3\u30bb\u30fc\u30b8\u3092\u898b\u3064\u3051\u305f\u3002<code>dmesg | grep -i nvidia<\/code> \u3078\u306e\u5fdc\u7b54\u306e\u4e2d\u306b <code>nvidia-nvlink: Nvlink Core is being initialized<\/code> \u3068\u3042\u308b\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">dmesg | grep -i nvidia\r\n[    2.872886] nvidia: loading out-of-tree module taints kernel.\r\n[    2.872892] nvidia: module license 'NVIDIA' taints kernel.\r\n[    2.876488] nvidia: module verification failed: signature and\/or required key missing - tainting kernel\r\n[    2.882683] nvidia-nvlink: Nvlink Core is being initialized, major device number 239\r\n[    2.883223] nvidia 0000:01:00.0: vgaarb: changed VGA decodes: olddecodes=io+mem,decodes=none:owns=io+mem\r\n[    2.927112] nvidia 0000:02:00.0: enabling device (0000 -> 0003)\r\n[    2.927191] nvidia 0000:02:00.0: vgaarb: changed VGA decodes: olddecodes=io+mem,decodes=none:owns=none\r\n[    2.976932] NVRM: loading NVIDIA UNIX x86_64 Kernel Module  450.51.06  Sun Jul 19 20:02:54 UTC 2020\r\n[    2.998138] nvidia-modeset: Loading NVIDIA Kernel Mode Setting Driver for UNIX platforms  450.51.06  Sun Jul 19 20:06:42 UTC 2020\r\n[    2.999991] [drm] [nvidia-drm] [GPU ID 0x00000100] Loading driver\r\n[    2.999992] [drm] Initialized nvidia-drm 0.0.0 20160202 for 0000:01:00.0 on minor 0\r\n[    3.000411] [drm] [nvidia-drm] [GPU ID 0x00000200] Loading driver\r\n[    3.000412] [drm] Initialized nvidia-drm 0.0.0 20160202 for 0000:02:00.0 on minor 1\r\n[    3.012872] nvidia-uvm: Loaded the UVM driver, major device number 236.\r\n[    3.436405] input: HDA NVidia HDMI\/DP,pcm=3 as \/devices\/pci0000:00\/0000:00:01.1\/0000:02:00.1\/sound\/card2\/input15\r\n[    3.436444] input: HDA NVidia HDMI\/DP,pcm=7 as \/devices\/pci0000:00\/0000:00:01.1\/0000:02:00.1\/sound\/card2\/input17\r\n[    3.436506] input: HDA NVidia HDMI\/DP,pcm=8 as \/devices\/pci0000:00\/0000:00:01.1\/0000:02:00.1\/sound\/card2\/input19\r\n[    3.436544] input: HDA NVidia HDMI\/DP,pcm=9 as \/devices\/pci0000:00\/0000:00:01.1\/0000:02:00.1\/sound\/card2\/input21\r\n[    3.436638] input: HDA NVidia HDMI\/DP,pcm=3 as \/devices\/pci0000:00\/0000:00:01.0\/0000:01:00.1\/sound\/card1\/input14\r\n[    3.436663] input: HDA NVidia HDMI\/DP,pcm=7 as \/devices\/pci0000:00\/0000:00:01.0\/0000:01:00.1\/sound\/card1\/input16\r\n[    3.436691] input: HDA NVidia HDMI\/DP,pcm=8 as \/devices\/pci0000:00\/0000:00:01.0\/0000:01:00.1\/sound\/card1\/input18\r\n[    3.436718] input: HDA NVidia HDMI\/DP,pcm=9 as \/devices\/pci0000:00\/0000:00:01.0\/0000:01:00.1\/sound\/card1\/input20\r\n[10301.646637] nvidia-modeset: WARNING: GPU:0: Unable to read EDID for display device Eizo EV2736W (HDMI-0)<\/code><\/pre>\n<p>\u4e0a\u624b\u304f\u884c\u3063\u3066\u3044\u308b\u306e\u3060\u308d\u3046\u304b\uff1f<code>nvidia-smi nvlink -c<\/code> \u306e\u51fa\u529b\u306f\u4e0b\u8a18\u306e\u3088\u3046\u306a\u611f\u3058\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">nvidia-smi nvlink -c\r\nGPU 0: GeForce RTX 2070 SUPER (UUID: GPU-34242219-6f19-24de-f016-787c4f3cb2d0)\r\n\t Link 0, P2P is supported: true\r\n\t Link 0, Access to system memory supported: true\r\n\t Link 0, P2P atomics supported: true\r\n\t Link 0, System memory atomics supported: true\r\n\t Link 0, SLI is supported: true\r\n\t Link 0, Link is supported: false\r\nGPU 1: GeForce RTX 2070 SUPER (UUID: GPU-aed31587-58ba-52cc-a909-6906c3635b30)\r\n\t Link 0, P2P is supported: true\r\n\t Link 0, Access to system memory supported: true\r\n\t Link 0, P2P atomics supported: true\r\n\t Link 0, System memory atomics supported: true\r\n\t Link 0, SLI is supported: true\r\n\t Link 0, Link is supported: false<\/code><\/pre>\n<p><code>nvidia-smi nvlink -s<\/code> \u306e\u51fa\u529b\u306f\u4e0b\u8a18\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">nvidia-smi nvlink -s\r\nGPU 0: GeForce RTX 2070 SUPER (UUID: GPU-34242219-6f19-24de-f016-787c4f3cb2d0)\r\n\t Link 0: 25.781 GB\/s\r\nGPU 1: GeForce RTX 2070 SUPER (UUID: GPU-aed31587-58ba-52cc-a909-6906c3635b30)\r\n\t Link 0: 25.781 GB\/s\r\n<\/code><\/pre>\n<p>CUDA \u30b5\u30f3\u30d7\u30eb\u3092\u8a66\u3057\u3066\u307f\u308b\u3002\u30b5\u30f3\u30d7\u30eb\u306e\u30b3\u30f3\u30d1\u30a4\u30eb\u3092\u3084\u3063\u3066\uff0c\u7aef\u672b\u304b\u3089\u8d77\u52d5\u3059\u308b\u3002\u4e0b\u56f3\u306f nbody \u3068\u3044\u3046\u30b5\u30f3\u30d7\u30eb\u3002\u30d5\u30a1\u30a4\u30eb\u3092\u30c0\u30d6\u30eb\u30af\u30ea\u30c3\u30af\u3067\u306f\u52d5\u304b\u306a\u3044\u3002\u30e9\u30a4\u30d6\u30e9\u30ea\u30fc\u304c\u306a\u3044\u3068\u304b\u6012\u3089\u308c\u308b\u3002<\/p>\n<p><a href=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200803a.png\"><img loading=\"lazy\" decoding=\"async\" src=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200803a-300x230.png\" alt=\"\" width=\"300\" height=\"230\" class=\"aligncenter size-medium wp-image-9237\" srcset=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200803a-300x230.png 300w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200803a-600x460.png 600w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200803a.png 735w\" sizes=\"auto, (max-width: 300px) 100vw, 300px\" \/><\/a><\/p>\n<p><a href=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200803b.png\"><img loading=\"lazy\" decoding=\"async\" src=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200803b-300x221.png\" alt=\"\" width=\"300\" height=\"221\" class=\"aligncenter size-medium wp-image-9238\" srcset=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200803b-300x221.png 300w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200803b-600x442.png 600w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200803b.png 764w\" sizes=\"auto, (max-width: 300px) 100vw, 300px\" \/><\/a><\/p>\n<p>\u30e1\u30e2\u30ea\u30fc\u3092\u307b\u3068\u3093\u3069\u4f7f\u7528\u3057\u306a\u3044\u3002\u3053\u308c\u3067 1 % \u3050\u3089\u3044\u3060\u3002\u3053\u306e\u30c7\u30e2\u3067\u306f\uff0c\u6e29\u5ea6\u304c\u6025\u4e0a\u6607\u3059\u308b\u300285\u5ea6\u3050\u3089\u3044\u307e\u3067\u4e00\u6c17\u306b\u4e0a\u304c\u308b\u3002\u5c11\u3057\u6016\u304f\u306a\u3063\u3066\uff0c\u3069\u3053\u307e\u3067\u4e0a\u6607\u3059\u308b\u304b\u306f\u8a66\u3055\u306a\u304b\u3063\u305f\u3002\u30c7\u30e2\u3092\u6b62\u3081\u308b\u3068\u6025\u964d\u4e0b\u3059\u308b\u3002<\/p>\n<p><a href=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200803c.png\"><img loading=\"lazy\" decoding=\"async\" src=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200803c-300x191.png\" alt=\"\" width=\"300\" height=\"191\" class=\"aligncenter size-medium wp-image-9242\" srcset=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200803c-300x191.png 300w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200803c-768x488.png 768w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200803c-1024x651.png 1024w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200803c-600x381.png 600w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200803c.png 1175w\" sizes=\"auto, (max-width: 300px) 100vw, 300px\" \/><\/a><\/p>\n<p>\uff12\u3064\u306e\u30d3\u30c7\u30aa\u30ab\u30fc\u30c9\u3092\u4f7f\u7528\u3057\u3066\u5b9f\u884c\u3059\u308b\u3088\u3046\u306b\u3001\u30aa\u30d7\u30b7\u30e7\u30f3\u3092\u3064\u3051\u3066 <code>nbody -numdevices=2<\/code> \u3068\u3057\u3066\u307f\u305f\u30022000 GFLOP\/S \u7a0b\u5ea6\u3060\u3063\u305f\u306e\u304c\u30014000 GFLOP\/S \u3042\u305f\u308a\u306b\u500d\u5897\u3057\u3066\u3044\u308b\u3002<\/p>\n<p><a href=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200803d.png\"><img loading=\"lazy\" decoding=\"async\" src=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200803d-300x212.png\" alt=\"\" width=\"300\" height=\"212\" class=\"aligncenter size-medium wp-image-9246\" srcset=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200803d-300x212.png 300w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200803d-600x423.png 600w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200803d.png 719w\" sizes=\"auto, (max-width: 300px) 100vw, 300px\" \/><\/a><\/p>\n<p>\u4e21\u65b9\u306e GPU \u306e\u6e29\u5ea6\u304c\u4e0a\u6607\u3057\u3066\u3044\u308b\u30022\u3064\u306e\u30d3\u30c7\u30aa\u30ab\u30fc\u30c9\u306e\u6e29\u5ea6\u306b\u5dee\u304c\u3042\u308a\u3059\u304e\u308b\u3002<\/p>\n<p><a href=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200803e.png\"><img loading=\"lazy\" decoding=\"async\" src=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200803e-300x190.png\" alt=\"\" width=\"300\" height=\"190\" class=\"aligncenter size-medium wp-image-9247\" srcset=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200803e-300x190.png 300w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200803e-768x487.png 768w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200803e-1024x649.png 1024w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200803e-600x380.png 600w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200803e.png 1177w\" sizes=\"auto, (max-width: 300px) 100vw, 300px\" \/><\/a><\/p>\n<p>(20200804)<br \/>\n\u30d3\u30c7\u30aa\u30ab\u30fc\u30c9\u306e\u30ab\u30d0\u30fc\u3092\u518d\u3073\u4ed8\u3051\u3066\u307f\u305f\u3002\u4e8c\u3064\u306e\u30ab\u30fc\u30c9\u306e\u6e29\u5ea6\u306e\u958b\u304d\u306f20\u5ea6\u3050\u3089\u3044\u3002\u6700\u9ad8\u6e29\u5ea6\u306f85\u5ea6\u3060\u3063\u305f\u3002\u3053\u3061\u3089\u306b\u3059\u308b\u3079\u304d\u304b\uff1f<\/p>\n<p><a href=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200804a.png\"><img loading=\"lazy\" decoding=\"async\" src=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200804a-300x191.png\" alt=\"\" width=\"300\" height=\"191\" class=\"aligncenter size-medium wp-image-9252\" srcset=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200804a-300x191.png 300w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200804a-768x488.png 768w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200804a-1024x651.png 1024w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200804a-600x381.png 600w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200804a.png 1177w\" sizes=\"auto, (max-width: 300px) 100vw, 300px\" \/><\/a><\/p>\n<p>\u540c\u3058\u6761\u4ef6\u3067\uff0c\u30ab\u30d0\u30fc\u7121\u3057\u306e\u65b9\u3092\u8a08\u6e2c\u3057\u3066\u307f\u305f\u3089\uff0c\u6700\u9ad8\u6e29\u5ea6\u304c82\u5ea6\u3060\u3063\u305f\u3002\u5f53\u521d\u306e\u8a08\u753b\u901a\u308a\u30ab\u30d0\u30fc\u7121\u3057\u306e\u65b9\u3067\u884c\u304f\u3053\u3068\u306b\u3059\u308b\u3002<\/p>\n<p><a href=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200804b.png\"><img loading=\"lazy\" decoding=\"async\" src=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200804b-300x190.png\" alt=\"\" width=\"300\" height=\"190\" class=\"aligncenter size-medium wp-image-9254\" srcset=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200804b-300x190.png 300w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200804b-768x487.png 768w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200804b-1024x650.png 1024w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200804b-600x381.png 600w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200804b.png 1176w\" sizes=\"auto, (max-width: 300px) 100vw, 300px\" \/><\/a><\/p>\n<p>\u5f8c\u65e5\uff0c\u5916\u6c17\u3092\u53d6\u308a\u8fbc\u3080\u30b1\u30fc\u30b9\u30d5\u30a1\u30f3\u306e\u30d5\u30a3\u30eb\u30bf\u30fc\u3092\u5916\u3057\u3066\uff0c\u98a8\u91cf\u3092\u3042\u3052\u3066\u307f\u305f\u3089 78 \u5ea6\u307e\u3067\u6e29\u5ea6\u304c\u4e0b\u304c\u3063\u305f\u3002\u3053\u308c\u3050\u3089\u3044\u306a\u3089\u5927\u4e08\u592b\u304b\u306a\uff1f<\/p>\n<p>\u4e0b\u8a18\u306e\u30b5\u30a4\u30c8\u306b\u3057\u305f\u304c\u3063\u3066\u3001\u3044\u304f\u3089\u304b\u78ba\u8a8d\u3057\u3066\u307f\u305f\u3002<\/p>\n<p>\u53c2\u8003\u30b5\u30a4\u30c8\uff1a<a href=\"https:\/\/www.pugetsystems.com\/labs\/hpc\/NVLINK-on-RTX-2080-TensorFlow-and-Peer-to-Peer-Performance-with-Linux-1262\/\">NVLINK on RTX 2080 TensorFlow and Peer-to-Peer Performance with Linux<\/a><\/p>\n<p>CUDA \u30b5\u30f3\u30d7\u30eb\u306b\u3042\u308b <code>simpleP2P<\/code> \u3092\u5b9f\u884c\u3057\u3066\u307f\u305f\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">.\/simpleP2P \r\n[.\/simpleP2P] - Starting...\r\nChecking for multiple GPUs...\r\nCUDA-capable device count: 2\r\n\r\nChecking GPU(s) for support of peer to peer memory access...\r\n> Peer access from GeForce RTX 2070 SUPER (GPU0) -> GeForce RTX 2070 SUPER (GPU1) : Yes\r\n> Peer access from GeForce RTX 2070 SUPER (GPU1) -> GeForce RTX 2070 SUPER (GPU0) : Yes\r\nEnabling peer access between GPU0 and GPU1...\r\nAllocating buffers (64MB on GPU0, GPU1 and CPU Host)...\r\nCreating event handles...\r\ncudaMemcpyPeer \/ cudaMemcpy between GPU0 and GPU1: 22.53GB\/s\r\nPreparing host buffer and memcpy to GPU0...\r\nRun kernel on GPU1, taking source data from GPU0 and writing to GPU1...\r\nRun kernel on GPU0, taking source data from GPU1 and writing to GPU0...\r\nCopy data back to host from GPU0 and verify results...\r\nDisabling peer access...\r\nShutting down...\r\nTest passed<\/code><\/pre>\n<p>CUDA \u30b5\u30f3\u30d7\u30eb\u306b\u3042\u308b <code>p2pBandwidthLatencyTest<\/code> \u3092\u5b9f\u884c\u3057\u3066\u307f\u305f\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">.\/p2pBandwidthLatencyTest \r\n[P2P (Peer-to-Peer) GPU Bandwidth Latency Test]\r\nDevice: 0, GeForce RTX 2070 SUPER, pciBusID: 1, pciDeviceID: 0, pciDomainID:0\r\nDevice: 1, GeForce RTX 2070 SUPER, pciBusID: 2, pciDeviceID: 0, pciDomainID:0\r\nDevice=0 CAN Access Peer Device=1\r\nDevice=1 CAN Access Peer Device=0\r\n\r\n***NOTE: In case a device doesn't have P2P access to other one, it falls back to normal memcopy procedure.\r\nSo you can see lesser Bandwidth (GB\/s) and unstable Latency (us) in those cases.\r\n\r\nP2P Connectivity Matrix\r\n     D\\D     0     1\r\n     0\t     1     1\r\n     1\t     1     1\r\nUnidirectional P2P=Disabled Bandwidth Matrix (GB\/s)\r\n   D\\D     0      1 \r\n     0 377.78   6.09 \r\n     1   6.10 388.04 \r\nUnidirectional P2P=Enabled Bandwidth (P2P Writes) Matrix (GB\/s)\r\n   D\\D     0      1 \r\n     0 388.37  24.23 \r\n     1  24.23 387.52 \r\nBidirectional P2P=Disabled Bandwidth Matrix (GB\/s)\r\n   D\\D     0      1 \r\n     0 385.93   9.21 \r\n     1   9.21 384.11 \r\nBidirectional P2P=Enabled Bandwidth Matrix (GB\/s)\r\n   D\\D     0      1 \r\n     0 387.50  48.38 \r\n     1  48.07 383.19 \r\nP2P=Disabled Latency Matrix (us)\r\n   GPU     0      1 \r\n     0   1.21  13.10 \r\n     1  13.15   1.23 \r\n\r\n   CPU     0      1 \r\n     0   1.99   5.23 \r\n     1   5.22   1.92 \r\nP2P=Enabled Latency (P2P Writes) Matrix (us)\r\n   GPU     0      1 \r\n     0   1.20   0.75 \r\n     1   0.70   1.24 \r\n\r\n   CPU     0      1 \r\n     0   1.95   1.68 \r\n     1   1.82   2.09 \r\n\r\nNOTE: The CUDA Samples are not meant for performance measurements. Results may vary when GPU Boost is enabled.<\/code><\/pre>\n<p>(20200806)<br \/>\nCUPY \u306e\u30de\u30cb\u30e5\u30a2\u30eb\u3092\u8aad\u3093\u3067\u3044\u308b\u3068\u3001GPU \u3092\u5207\u308a\u66ff\u3048\u308b\u3068\u3044\u3046\u30b3\u30de\u30f3\u30c9\u304c\u3042\u3063\u305f\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\"># coding: utf-8\r\nimport sys, os\r\nimport numpy as np\r\nimport cupy as cp\r\n\r\ncp.cuda.Device(1).use()<\/code><\/pre>\n<p>\u4e0a\u8a18\u306e\u6700\u5f8c\u306e\u30b3\u30fc\u30c9\u3067\u4f7f\u7528\u3059\u308b GPU \u3092\u5207\u308a\u66ff\u3048\u308b\u3002\u305d\u3046\u3059\u308b\u3068\u78ba\u304b\u306b2\u756a\u3081\u306e GPU \u3067\u8a08\u7b97\u3057\u59cb\u3081\u305f\u3088\u3046\u3067\u30012\u756a\u3081\u306e\u30d3\u30c7\u30aa\u30ab\u30fc\u30c9\u306e\u6e29\u5ea6\u304c\u4e0a\u6607\u3057\u305f\u3002<\/p>\n<p>(20200810)<br \/>\n<a href=\"https:\/\/www.v-t.co.jp\/product\/gpuphi\/nvlink\/\">https:\/\/www.v-t.co.jp\/product\/gpuphi\/nvlink\/<\/a>\u306b\u306f\u4e0b\u8a18\u306e\u3088\u3046\u306b\u3042\u308b\u3002<\/p>\n<ul><em>Quadro\uff08RTX\/GP\/GV\uff09\u30b7\u30ea\u30fc\u30ba\u306e2\u3064\u306e GPU\u3092\u3001NVLink\u3092\u7528\u3044\u305fNVLink Bridge\u3067\u63a5\u7d9a\u3059\u308b\u3053\u3068\u306b\u3088\u308a\u3001\u5f93\u6765\u306ePCIe\u30d0\u30b9\u9593\u306e\u901a\u4fe1\u3088\u308a\u3082\u9ad8\u901f\u306b\u901a\u4fe1\u3059\u308b\u3053\u3068\u304c\u53ef\u80fd\u3067\u3059\u3002\u5bfe\u5fdc\u3057\u305f\u30a2\u30d7\u30ea\u30b1\u30fc\u30b7\u30e7\u30f3\u3092\u5229\u7528\u3059\u308c\u3070\u3001\u30de\u30eb\u30c1GPU \u69cb\u6210\u3067\u30e1\u30e2\u30ea\u3068\u30d1\u30d5\u30a9\u30fc\u30de\u30f3\u30b9\u3092\u7c21\u5358\u306b\u62e1\u5f35\u3067\u304d\u307e\u3059\u3002<br \/>\n\u4f8b\u3048\u3070\u30012\u3064\u306eQuadro RTX 8000\u3092Quadro RTX 8000 NVLink HB \u30d6\u30ea\u30c3\u30b8\u3067\u63a5\u7d9a\u3057\u3001\u6700\u5927\u6bce\u79d2 100 GB \u306e\u5e2f\u57df\u5e45\u3001\u53ca\u3073\u5408\u8a0896GB\u306eGDDR6 \u30e1\u30e2\u30ea\u3068\u3057\u3066\u3001\u5927\u898f\u6a21\u306a\u30ec\u30f3\u30c0\u30ea\u30f3\u30b0\u3001AI\u3001\u30d0\u30fc\u30c1\u30e3\u30eb \u30ea\u30a2\u30ea\u30c6\u30a3\u3001\u30d3\u30b8\u30e5\u30a2\u30e9\u30a4\u30bc\u30fc\u30b7\u30e7\u30f3\u306e\u30ef\u30fc\u30af\u30ed\u30fc\u30c9\u306b\u5bfe\u5fdc\u3067\u304d\u307e\u3059\u3002<br \/>\n\u305f\u3060\u3057GeForce RTX 20\u30b7\u30ea\u30fc\u30ba\uff082060\u306f\u672a\u5bfe\u5fdc\uff09\u3067\u306f\u3001\u63a5\u7d9a\u30a4\u30f3\u30bf\u30fc\u30d5\u30a7\u30a4\u30b9\u3068\u3057\u3066\u306e\u5229\u7528\u306b\u7559\u307e\u308a\u3001\u30e1\u30e2\u30ea\u3084CUDA\u30b3\u30a2\u3092\u7d71\u5408\u3057\u3066\u6271\u3046\u3053\u3068\u306f\u3067\u304d\u307e\u305b\u3093\u304c\u3001P2P\u3067\u306eGPU\u9593\u306e\u30c7\u30fc\u30bf\u8ee2\u9001\u306f\u9ad8\u901f\u5316\u3055\u308c\u307e\u3059\u3002\u4e00\u65b9\u3067\u3001\u30bd\u30d5\u30c8\u30a6\u30a7\u30a2\u7684\u306a\u4e92\u63db\u6027\u306f\u7dad\u6301\u3055\u308c\u3066\u3044\u308b\u305f\u3081\u3001\u300cSLI\u300d\u5bfe\u5fdc\u30a2\u30d7\u30ea\u30b1\u30fc\u30b7\u30e7\u30f3\u306f\u300cNVLink SLI\u300d\u3067\u3082\u305d\u306e\u307e\u307e\u4f7f\u7528\u3059\u308b\u3053\u3068\u304c\u3067\u304d\u307e\u3059\u3002 <\/em><\/ul>\n<p>\u3053\u308c\u304b\u3089\u3059\u308b\u3068\uff0c\u79c1\u306e\u30ec\u30d9\u30eb\u3067\u306f NVLink \u306e\u6069\u6075\u306b\u306f\u3042\u305a\u304b\u308c\u305d\u3046\u306b\u306a\u3044\u3002\u5148\u306b\u3042\u3052\u305f\u30ea\u30f3\u30af\u5148<\/p>\n<ul><a href=\"https:\/\/www.pugetsystems.com\/labs\/hpc\/NVLINK-on-RTX-2080-TensorFlow-and-Peer-to-Peer-Performance-with-Linux-1262\/\">NVLINK on RTX 2080 TensorFlow and Peer-to-Peer Performance with Linux<\/a><\/ul>\n<p>\u3053\u3053\u306b\u8a18\u8ff0\u3055\u308c\u3066\u3044\u308b\u5185\u5bb9\u306b\u671f\u5f85\u3059\u308b\u3050\u3089\u3044\u3060\u308d\u3046\u304b\u3002<\/p>\n<p>(20200814)<br \/>\n<a href=\"https:\/\/www.oreilly.co.jp\/books\/9784873118345\/\">\u300escikit-learn\u3068TensorFlow\u306b\u3088\u308b\u5b9f\u8df5\u6a5f\u68b0\u5b66\u7fd2\u300f<\/a>\u3068\u3044\u3046\u672c\u304c\u624b\u5143\u306b\u5728\u3063\u3066\uff0c<a href=\"https:\/\/www.pugetsystems.com\/labs\/hpc\/NVLINK-on-RTX-2080-TensorFlow-and-Peer-to-Peer-Performance-with-Linux-1262\/\">\u5148\u306b\u3042\u3052\u305f\u30ea\u30f3\u30af<\/a>\u304c tensorflow \u3067\u30d9\u30f3\u30c1\u30de\u30fc\u30af\u3092\u53d6\u3063\u3066\u3044\u305f\u306e\u3067\uff0c\u5c11\u3057\u8aad\u3093\u3067\u307f\u308b\u3053\u3068\u306b\u3057\u305f\u3002\u3057\u304b\u3057\u672c\u306e\u8a18\u8ff0\u901a\u308a\u306b\u306f\u30b5\u30f3\u30d7\u30eb\u30b3\u30fc\u30c9\u304c\u52d5\u304b\u306a\u3044\u300212\u7ae0\u306bGPU\u5bfe\u5fdc\u306e tensorflow \u3092\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u3059\u308b\u8a71\u304c\u5728\u3063\u3066\uff0c\u305d\u3053\u306e\u8a18\u8ff0\u306b\u5f93\u3063\u3066\u3084\u3063\u3066\u307f\u305f\u304c <code>sess = tf.Session()<\/code> \u3067\u30a8\u30e9\u30fc\u3068\u306a\u308b\u3002\u8abf\u3079\u3066\u307f\u308b\u3068 tensorflow \u306f\u30d0\u30fc\u30b8\u30e7\u30f3\uff12\u3067\u8272\u3005\u3068\u5909\u5316\u304c\u3042\u3063\u305f\u3088\u3046\u3067\uff0csession \u306f\u4f7f\u308f\u306a\u304f\u306a\u3063\u305f\u3088\u3046\u3060\u3002\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u3057\u305f\u30bd\u30d5\u30c8\u306e\u30d0\u30fc\u30b8\u30e7\u30f3\u3092\u8abf\u3079\u305f\u3089\uff0cCUDA \u304c 11 \u3067\uff0ctensorflow \u304c 2.3 \u3060\u3063\u305f\u3002\u3053\u306e\u672c\u306f\u539f\u66f8\u3067\u306f\u6b21\u306e\u7248\u304c\u51fa\u3066\u3044\u305f\u304c\uff0c\u65e5\u672c\u8a9e\u8a33\u306f\u307e\u3060\u3067\u3042\u308b\u3002<a href=\"https:\/\/www.oreilly.com\/library\/view\/hands-on-machine-learning\/9781492032632\/\">\u7b2c2\u7248\u306e\u30b5\u30a4\u30c8<\/a>\u306b\u306f\u307e\u3060\u30b5\u30f3\u30d7\u30eb\u30b3\u30fc\u30c9\u304c\u8f09\u305b\u3089\u308c\u3066\u3044\u306a\u3044\u3002\u304d\u3063\u3068\u5185\u5bb9\u306f tensorflow2 \u306b\u5bfe\u5fdc\u3057\u305f\u3082\u306e\u306b\u306a\u3063\u3066\u3044\u308b\u3068\u601d\u3046\u306e\u3060\u304c\u3002<br \/>\nCUDA 10.1 \u306e\u30de\u30b7\u30f3\u304c\u5728\u3063\u3066\uff0c\u305d\u3061\u3089\u306b tensorflow 1.4 \u3092\u5165\u308c\u3066\u307f\u308b\u3068\u672c\u306e\u30b3\u30fc\u30c9\u304c\u52d5\u3044\u305f\u3002\u305f\u3060\u3057 gpu \u7121\u3057\u306e tensorflow \u306e\u5834\u5408\u3067\u3042\u308b\u3002\u30d0\u30fc\u30b8\u30e7\u30f3\u3092\u6307\u5b9a\u3057\u305f\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u65b9\u6cd5\u306f\uff0c<a href=\"http:\/\/tecsingularity.com\/cuda\/version\/\">\u3053\u306e\u30b5\u30a4\u30c8<\/a>\u3092\u53c2\u8003\u306b\u3057\u305f\u3002\u30b3\u30de\u30f3\u30c9\u3092\u8a18\u9332\u3059\u308b\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">pip3 install tensorflow==1.4<\/code><\/pre>\n<p>tensorflow \u3092\u52d5\u304b\u3057\u306a\u304c\u3089\uff0c\u300escikit-learn\u3068TensorFlow\u306b\u3088\u308b\u5b9f\u8df5\u6a5f\u68b0\u5b66\u7fd2\u300f\u306e9\u7ae0\u3092\u8aad\u3093\u3067\u307f\u305f\u3002<\/p>\n<p>(20200818)<br \/>\n\u4e0b\u8a18\u306e\u8a18\u4e8b\u3092\u591a\u5c11\u8aad\u3093\u3067\u307f\u305f\u3002tensorflow \u306e\u30d0\u30fc\u30b8\u30e7\u30f3\u306f 1.5 \u3067\u3042\u308b\u3002docker \u95a2\u9023\u306e\u60c5\u5831\u304c\u3042\u308b\u3002<\/p>\n<ul>\u53c2\u8003\u30b5\u30a4\u30c8\uff1a<a href=\"https:\/\/www.atmarkit.co.jp\/ait\/subtop\/features\/di\/introtensorflow_index.html\">TensorFlow\u5165\u9580<\/a><\/ul>\n<p>\u3053\u306e\u5148\uff0cdocker \u3092\u4f7f\u3063\u3066 NVLink \u306e\u52b9\u679c\u3092\u8a66\u3059\u3053\u3068\u306b\u306a\u308a\u305d\u3046\u3002<\/p>\n<p>(20200821)<br \/>\n\u4e0b\u8a18\u306e\u8a18\u4e8b\u3092\u8aad\u3080\u3002<\/p>\n<ul>\n<li><a href=\"https:\/\/qiita.com\/ksasaki\/items\/b20a785e1a0f610efa08\">NVIDIA Docker \u3063\u3066\u4eca\u3069\u3046\u306a\u3063\u3066\u308b\u306e\uff1f (19.11\u7248)<\/a><\/li>\n<li><a href=\"https:\/\/knowledge.sakura.ad.jp\/13265\/\">Docker\u5165\u9580\uff08\u7b2c\u4e00\u56de\uff09\uff5eDocker\u3068\u306f\u4f55\u304b\u3001\u4f55\u304c\u826f\u3044\u306e\u304b\uff5e<\/a><\/li>\n<\/ul>\n<p>(20200824)<br \/>\n\u30a2\u30d7\u30e9\u30a4\u30c9\u306e\u30b5\u30a4\u30c8\u300c<a href=\"https:\/\/bto.applied.ne.jp\/c1189-c2208-pm235.html\">TensorFlow \u30d9\u30f3\u30c1\u30fc\u30de\u30fc\u30af\u30c6\u30b9\u30c8\uff08\uff11\uff09GeFoce 2080Ti<\/a>\u300d\u306b\u3042\u308b\u5185\u5bb9\u3092\u8a66\u3057\u3066\u307f\u305f\u3002<\/p>\n<p>\u30d9\u30f3\u30c1\u30de\u30fc\u30af\u306e\u30c0\u30a6\u30f3\u30ed\u30fc\u30c9\u3092\u3059\u308b\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">git clone https:\/\/github.com\/tensorflow\/benchmarks.git<\/code><\/pre>\n<p>\u305d\u306e\u4e2d\u306b\u3042\u308b\u30d5\u30a1\u30a4\u30eb benchmarks\/scripts\/tf_cnn_benchmarks\/tf_cnn_benchmarks.py \u304c\u30bf\u30fc\u30b2\u30c3\u30c8\u3067\u3042\u308b\u3002<br \/>\n\u9694\u96e2\u3055\u308c\u305f\u74b0\u5883(tensorflow 2.3)\u3067\u3001\u4e0a\u8a18\u306e\u30b9\u30af\u30ea\u30d7\u30c8\u3092\u5b9f\u884c\u3057\u305f\u3002\u4e0b\u8a18\u306f\u3001gpu \u30922\u3064\u5229\u7528\u3057\u305f\u5834\u5408\u3067\u3042\u308b\u3002\u30b3\u30de\u30f3\u30c9\u306f\u540c\u3058\u30d5\u30a9\u30eb\u30c0\u30fc\u306b\u3042\u308b READEME.md \u306b\u8a18\u3057\u3066\u3042\u3063\u305f\u3082\u306e\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">(env) friend@z390:~\/benchmarks\/scripts\/tf_cnn_benchmarks$ python3 .\/tf_cnn_benchmarks.py --num_gpus=2 --batch_size=32 --model=resnet50 --variable_update=parameter_server\r\n2020-08-24 17:18:43.949784: I tensorflow\/stream_executor\/platform\/default\/dso_loader.cc:48] Successfully opened dynamic library libcudart.so.10.1\r\nWARNING:tensorflow:From \/home\/friend\/ml_gpu\/env\/lib\/python3.6\/site-packages\/tensorflow\/python\/compat\/v2_compat.py:96: disable_resource_variables (from tensorflow.python.ops.variable_scope) is deprecated and will be removed in a future version.\r\nInstructions for updating:\r\nnon-resource variables are not supported in the long term\r\n2020-08-24 17:18:44.925432: I tensorflow\/core\/platform\/cpu_feature_guard.cc:142] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN)to use the following CPU instructions in performance-critical operations:  AVX2 FMA\r\nTo enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.\r\n2020-08-24 17:18:44.950146: I tensorflow\/core\/platform\/profile_utils\/cpu_utils.cc:104] CPU Frequency: 3000000000 Hz\r\n2020-08-24 17:18:44.950385: I tensorflow\/compiler\/xla\/service\/service.cc:168] XLA service 0x45bfaf0 initialized for platform Host (this does not guarantee that XLA will be used). Devices:\r\n2020-08-24 17:18:44.950404: I tensorflow\/compiler\/xla\/service\/service.cc:176]   StreamExecutor device (0): Host, Default Version\r\n2020-08-24 17:18:44.952309: I tensorflow\/stream_executor\/platform\/default\/dso_loader.cc:48] Successfully opened dynamic library libcuda.so.1\r\n2020-08-24 17:18:45.105237: I tensorflow\/stream_executor\/cuda\/cuda_gpu_executor.cc:982] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\r\n2020-08-24 17:18:45.113618: I tensorflow\/stream_executor\/cuda\/cuda_gpu_executor.cc:982] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\r\n2020-08-24 17:18:45.114043: I tensorflow\/compiler\/xla\/service\/service.cc:168] XLA service 0x45be170 initialized for platform CUDA (this does not guarantee that XLA will be used). Devices:\r\n2020-08-24 17:18:45.114056: I tensorflow\/compiler\/xla\/service\/service.cc:176]   StreamExecutor device (0): GeForce RTX 2070 SUPER, Compute Capability 7.5\r\n2020-08-24 17:18:45.114061: I tensorflow\/compiler\/xla\/service\/service.cc:176]   StreamExecutor device (1): GeForce RTX 2070 SUPER, Compute Capability 7.5\r\n2020-08-24 17:18:45.114440: I tensorflow\/stream_executor\/cuda\/cuda_gpu_executor.cc:982] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\r\n2020-08-24 17:18:45.114778: I tensorflow\/core\/common_runtime\/gpu\/gpu_device.cc:1716] Found device 0 with properties: \r\npciBusID: 0000:01:00.0 name: GeForce RTX 2070 SUPER computeCapability: 7.5\r\ncoreClock: 1.815GHz coreCount: 40 deviceMemorySize: 7.79GiB deviceMemoryBandwidth: 417.29GiB\/s\r\n2020-08-24 17:18:45.114824: I tensorflow\/stream_executor\/cuda\/cuda_gpu_executor.cc:982] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\r\n2020-08-24 17:18:45.115147: I tensorflow\/core\/common_runtime\/gpu\/gpu_device.cc:1716] Found device 1 with properties: \r\npciBusID: 0000:02:00.0 name: GeForce RTX 2070 SUPER computeCapability: 7.5\r\ncoreClock: 1.815GHz coreCount: 40 deviceMemorySize: 7.79GiB deviceMemoryBandwidth: 417.29GiB\/s\r\n2020-08-24 17:18:45.115168: I tensorflow\/stream_executor\/platform\/default\/dso_loader.cc:48] Successfully opened dynamic library libcudart.so.10.1\r\n2020-08-24 17:18:45.116255: I tensorflow\/stream_executor\/platform\/default\/dso_loader.cc:48] Successfully opened dynamic library libcublas.so.10\r\n2020-08-24 17:18:45.117360: I tensorflow\/stream_executor\/platform\/default\/dso_loader.cc:48] Successfully opened dynamic library libcufft.so.10\r\n2020-08-24 17:18:45.117600: I tensorflow\/stream_executor\/platform\/default\/dso_loader.cc:48] Successfully opened dynamic library libcurand.so.10\r\n2020-08-24 17:18:45.118782: I tensorflow\/stream_executor\/platform\/default\/dso_loader.cc:48] Successfully opened dynamic library libcusolver.so.10\r\n2020-08-24 17:18:45.119464: I tensorflow\/stream_executor\/platform\/default\/dso_loader.cc:48] Successfully opened dynamic library libcusparse.so.10\r\n2020-08-24 17:18:45.121877: I tensorflow\/stream_executor\/platform\/default\/dso_loader.cc:48] Successfully opened dynamic library libcudnn.so.7\r\n2020-08-24 17:18:45.121976: I tensorflow\/stream_executor\/cuda\/cuda_gpu_executor.cc:982] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\r\n2020-08-24 17:18:45.122365: I tensorflow\/stream_executor\/cuda\/cuda_gpu_executor.cc:982] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\r\n2020-08-24 17:18:45.122712: I tensorflow\/stream_executor\/cuda\/cuda_gpu_executor.cc:982] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\r\n2020-08-24 17:18:45.123049: I tensorflow\/stream_executor\/cuda\/cuda_gpu_executor.cc:982] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\r\n2020-08-24 17:18:45.123361: I tensorflow\/core\/common_runtime\/gpu\/gpu_device.cc:1858] Adding visible gpu devices: 0, 1\r\n2020-08-24 17:18:45.123385: I tensorflow\/stream_executor\/platform\/default\/dso_loader.cc:48] Successfully opened dynamic library libcudart.so.10.1\r\n2020-08-24 17:18:45.737310: I tensorflow\/core\/common_runtime\/gpu\/gpu_device.cc:1257] Device interconnect StreamExecutor with strength 1 edge matrix:\r\n2020-08-24 17:18:45.737361: I tensorflow\/core\/common_runtime\/gpu\/gpu_device.cc:1263]      0 1 \r\n2020-08-24 17:18:45.737368: I tensorflow\/core\/common_runtime\/gpu\/gpu_device.cc:1276] 0:   N Y \r\n2020-08-24 17:18:45.737374: I tensorflow\/core\/common_runtime\/gpu\/gpu_device.cc:1276] 1:   Y N \r\n2020-08-24 17:18:45.737571: I tensorflow\/stream_executor\/cuda\/cuda_gpu_executor.cc:982] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\r\n2020-08-24 17:18:45.738093: I tensorflow\/stream_executor\/cuda\/cuda_gpu_executor.cc:982] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\r\n2020-08-24 17:18:45.738478: I tensorflow\/stream_executor\/cuda\/cuda_gpu_executor.cc:982] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\r\n2020-08-24 17:18:45.738843: I tensorflow\/core\/common_runtime\/gpu\/gpu_device.cc:1402] Created TensorFlow device (\/job:localhost\/replica:0\/task:0\/device:GPU:0 with 7023 MB memory) -> physical GPU (device: 0, name: GeForce RTX 2070 SUPER, pci bus id: 0000:01:00.0, compute capability: 7.5)\r\n2020-08-24 17:18:45.739364: I tensorflow\/stream_executor\/cuda\/cuda_gpu_executor.cc:982] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\r\n2020-08-24 17:18:45.739713: I tensorflow\/core\/common_runtime\/gpu\/gpu_device.cc:1402] Created TensorFlow device (\/job:localhost\/replica:0\/task:0\/device:GPU:1 with 7269 MB memory) -> physical GPU (device: 1, name: GeForce RTX 2070 SUPER, pci bus id: 0000:02:00.0, compute capability: 7.5)\r\nTensorFlow:  2.3\r\nModel:       resnet50\r\nDataset:     imagenet (synthetic)\r\nMode:        training\r\nSingleSess:  False\r\nBatch size:  64 global\r\n             32 per device\r\nNum batches: 100\r\nNum epochs:  0.00\r\nDevices:     ['\/gpu:0', '\/gpu:1']\r\nNUMA bind:   False\r\nData format: NCHW\r\nOptimizer:   sgd\r\nVariables:   parameter_server\r\n==========\r\nGenerating training model\r\nWARNING:tensorflow:From \/home\/friend\/benchmarks\/scripts\/tf_cnn_benchmarks\/convnet_builder.py:134: conv2d (from tensorflow.python.keras.legacy_tf_layers.convolutional) is deprecated and will be removed in a future version.\r\nInstructions for updating:\r\nUse `tf.keras.layers.Conv2D` instead.\r\nW0824 17:18:45.750814 139659219543872 deprecation.py:323] From \/home\/friend\/benchmarks\/scripts\/tf_cnn_benchmarks\/convnet_builder.py:134: conv2d (from tensorflow.python.keras.legacy_tf_layers.convolutional) is deprecated and will be removed in a future version.\r\nInstructions for updating:\r\nUse `tf.keras.layers.Conv2D` instead.\r\nWARNING:tensorflow:From \/home\/friend\/ml_gpu\/env\/lib\/python3.6\/site-packages\/tensorflow\/python\/keras\/legacy_tf_layers\/convolutional.py:424: Layer.apply (from tensorflow.python.keras.engine.base_layer_v1) is deprecated and will be removed in a future version.\r\nInstructions for updating:\r\nPlease use `layer.__call__` method instead.\r\nW0824 17:18:45.754501 139659219543872 deprecation.py:323] From \/home\/friend\/ml_gpu\/env\/lib\/python3.6\/site-packages\/tensorflow\/python\/keras\/legacy_tf_layers\/convolutional.py:424: Layer.apply (from tensorflow.python.keras.engine.base_layer_v1) is deprecated and will be removed in a future version.\r\nInstructions for updating:\r\nPlease use `layer.__call__` method instead.\r\nWARNING:tensorflow:From \/home\/friend\/benchmarks\/scripts\/tf_cnn_benchmarks\/convnet_builder.py:266: max_pooling2d (from tensorflow.python.keras.legacy_tf_layers.pooling) is deprecated and will be removed in a future version.\r\nInstructions for updating:\r\nUse keras.layers.MaxPooling2D instead.\r\nW0824 17:18:45.776025 139659219543872 deprecation.py:323] From \/home\/friend\/benchmarks\/scripts\/tf_cnn_benchmarks\/convnet_builder.py:266: max_pooling2d (from tensorflow.python.keras.legacy_tf_layers.pooling) is deprecated and will be removed in a future version.\r\nInstructions for updating:\r\nUse keras.layers.MaxPooling2D instead.\r\nInitializing graph\r\nWARNING:tensorflow:From \/home\/friend\/benchmarks\/scripts\/tf_cnn_benchmarks\/benchmark_cnn.py:2268: Supervisor.__init__ (from tensorflow.python.training.supervisor) is deprecated and will be removed in a future version.\r\nInstructions for updating:\r\nPlease switch to tf.train.MonitoredTrainingSession\r\nW0824 17:18:48.420785 139659219543872 deprecation.py:323] From \/home\/friend\/benchmarks\/scripts\/tf_cnn_benchmarks\/benchmark_cnn.py:2268: Supervisor.__init__ (from tensorflow.python.training.supervisor) is deprecated and will be removed in a future version.\r\nInstructions for updating:\r\nPlease switch to tf.train.MonitoredTrainingSession\r\n2020-08-24 17:18:48.689375: I tensorflow\/stream_executor\/cuda\/cuda_gpu_executor.cc:982] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\r\n2020-08-24 17:18:48.689736: I tensorflow\/core\/common_runtime\/gpu\/gpu_device.cc:1716] Found device 0 with properties: \r\npciBusID: 0000:01:00.0 name: GeForce RTX 2070 SUPER computeCapability: 7.5\r\ncoreClock: 1.815GHz coreCount: 40 deviceMemorySize: 7.79GiB deviceMemoryBandwidth: 417.29GiB\/s\r\n2020-08-24 17:18:48.689798: I tensorflow\/stream_executor\/cuda\/cuda_gpu_executor.cc:982] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\r\n2020-08-24 17:18:48.690118: I tensorflow\/core\/common_runtime\/gpu\/gpu_device.cc:1716] Found device 1 with properties: \r\npciBusID: 0000:02:00.0 name: GeForce RTX 2070 SUPER computeCapability: 7.5\r\ncoreClock: 1.815GHz coreCount: 40 deviceMemorySize: 7.79GiB deviceMemoryBandwidth: 417.29GiB\/s\r\n2020-08-24 17:18:48.690139: I tensorflow\/stream_executor\/platform\/default\/dso_loader.cc:48] Successfully opened dynamic library libcudart.so.10.1\r\n2020-08-24 17:18:48.690155: I tensorflow\/stream_executor\/platform\/default\/dso_loader.cc:48] Successfully opened dynamic library libcublas.so.10\r\n2020-08-24 17:18:48.690165: I tensorflow\/stream_executor\/platform\/default\/dso_loader.cc:48] Successfully opened dynamic library libcufft.so.10\r\n2020-08-24 17:18:48.690175: I tensorflow\/stream_executor\/platform\/default\/dso_loader.cc:48] Successfully opened dynamic library libcurand.so.10\r\n2020-08-24 17:18:48.690184: I tensorflow\/stream_executor\/platform\/default\/dso_loader.cc:48] Successfully opened dynamic library libcusolver.so.10\r\n2020-08-24 17:18:48.690193: I tensorflow\/stream_executor\/platform\/default\/dso_loader.cc:48] Successfully opened dynamic library libcusparse.so.10\r\n2020-08-24 17:18:48.690202: I tensorflow\/stream_executor\/platform\/default\/dso_loader.cc:48] Successfully opened dynamic library libcudnn.so.7\r\n2020-08-24 17:18:48.690237: I tensorflow\/stream_executor\/cuda\/cuda_gpu_executor.cc:982] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\r\n2020-08-24 17:18:48.690566: I tensorflow\/stream_executor\/cuda\/cuda_gpu_executor.cc:982] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\r\n2020-08-24 17:18:48.690896: I tensorflow\/stream_executor\/cuda\/cuda_gpu_executor.cc:982] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\r\n2020-08-24 17:18:48.691223: I tensorflow\/stream_executor\/cuda\/cuda_gpu_executor.cc:982] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\r\n2020-08-24 17:18:48.691566: I tensorflow\/core\/common_runtime\/gpu\/gpu_device.cc:1858] Adding visible gpu devices: 0, 1\r\n2020-08-24 17:18:48.691616: I tensorflow\/core\/common_runtime\/gpu\/gpu_device.cc:1257] Device interconnect StreamExecutor with strength 1 edge matrix:\r\n2020-08-24 17:18:48.691637: I tensorflow\/core\/common_runtime\/gpu\/gpu_device.cc:1263]      0 1 \r\n2020-08-24 17:18:48.691641: I tensorflow\/core\/common_runtime\/gpu\/gpu_device.cc:1276] 0:   N Y \r\n2020-08-24 17:18:48.691645: I tensorflow\/core\/common_runtime\/gpu\/gpu_device.cc:1276] 1:   Y N \r\n2020-08-24 17:18:48.691757: I tensorflow\/stream_executor\/cuda\/cuda_gpu_executor.cc:982] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\r\n2020-08-24 17:18:48.692168: I tensorflow\/stream_executor\/cuda\/cuda_gpu_executor.cc:982] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\r\n2020-08-24 17:18:48.692665: I tensorflow\/stream_executor\/cuda\/cuda_gpu_executor.cc:982] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\r\n2020-08-24 17:18:48.693011: I tensorflow\/core\/common_runtime\/gpu\/gpu_device.cc:1402] Created TensorFlow device (\/job:localhost\/replica:0\/task:0\/device:GPU:0 with 7023 MB memory) -> physical GPU (device: 0, name: GeForce RTX 2070 SUPER, pci bus id: 0000:01:00.0, compute capability: 7.5)\r\n2020-08-24 17:18:48.693165: I tensorflow\/stream_executor\/cuda\/cuda_gpu_executor.cc:982] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\r\n2020-08-24 17:18:48.693480: I tensorflow\/core\/common_runtime\/gpu\/gpu_device.cc:1402] Created TensorFlow device (\/job:localhost\/replica:0\/task:0\/device:GPU:1 with 7269 MB memory) -> physical GPU (device: 1, name: GeForce RTX 2070 SUPER, pci bus id: 0000:02:00.0, compute capability: 7.5)\r\nINFO:tensorflow:Running local_init_op.\r\nI0824 17:18:50.304418 139659219543872 session_manager.py:505] Running local_init_op.\r\nINFO:tensorflow:Done running local_init_op.\r\nI0824 17:18:50.338307 139659219543872 session_manager.py:508] Done running local_init_op.\r\nRunning warm up\r\n2020-08-24 17:18:51.596903: I tensorflow\/stream_executor\/platform\/default\/dso_loader.cc:48] Successfully opened dynamic library libcublas.so.10\r\n2020-08-24 17:18:51.931339: I tensorflow\/stream_executor\/platform\/default\/dso_loader.cc:48] Successfully opened dynamic library libcudnn.so.7\r\nDone warm up\r\nStep\tImg\/sec\ttotal_loss\r\n1\timages\/sec: 367.7 +\/- 0.0 (jitter = 0.0)\t7.829\r\n10\timages\/sec: 384.5 +\/- 4.4 (jitter = 12.5)\t8.001\r\n20\timages\/sec: 379.3 +\/- 4.2 (jitter = 13.3)\t7.946\r\n30\timages\/sec: 381.3 +\/- 3.4 (jitter = 8.0)\t7.869\r\n40\timages\/sec: 378.8 +\/- 3.1 (jitter = 14.3)\t7.727\r\n50\timages\/sec: 378.4 +\/- 2.8 (jitter = 14.8)\t7.751\r\n60\timages\/sec: 377.6 +\/- 2.6 (jitter = 24.8)\t7.933\r\n70\timages\/sec: 376.3 +\/- 2.4 (jitter = 29.1)\t7.915\r\n80\timages\/sec: 375.5 +\/- 2.3 (jitter = 28.1)\t7.894\r\n90\timages\/sec: 374.7 +\/- 2.1 (jitter = 23.9)\t7.830\r\n100\timages\/sec: 374.5 +\/- 2.0 (jitter = 23.9)\t8.156\r\n----------------------------------------------------------------\r\ntotal images\/sec: 374.28\r\n----------------------------------------------------------------\r\n(env) friend@z390:~\/benchmarks\/scripts\/tf_cnn_benchmarks$ <\/code><\/pre>\n<p>\u6e29\u5ea6\u4e0a\u6607\u3068\u30e1\u30e2\u30ea\u30fc\u306e\u4f7f\u7528\u7387\u306f\u4e0b\u56f3\u3002\u30e1\u30e2\u30ea\u30fc\u306f\u4e21\u65b9\u306e gpu \u3068\u3082 70% \u3050\u3089\u3044\u3060\u3063\u305f\u3002\u6e29\u5ea6\u304c\u4e0b\u304c\u308b\u306e\u306b\u306f\u6642\u9593\u304c\u304b\u304b\u3063\u3066\u3044\u308b\u3002\u30e1\u30e2\u30ea\u30fc\u306e\u4f7f\u7528\u7387\u304c\u4e0b\u304c\u3063\u305f\u3068\u3053\u308d\u3042\u305f\u308a\u304c\u8a08\u7b97\u306e\u7d42\u4e86\u6642\u523b\u3092\u793a\u3057\u3066\u3044\u308b\u3002<br \/>\n<a href=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/08\/20200824a.png\"><img loading=\"lazy\" decoding=\"async\" src=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/08\/20200824a-300x190.png\" alt=\"\" width=\"300\" height=\"190\" class=\"aligncenter size-medium wp-image-9339\" srcset=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/08\/20200824a-300x190.png 300w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/08\/20200824a-768x487.png 768w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/08\/20200824a-1024x649.png 1024w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/08\/20200824a-600x380.png 600w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/08\/20200824a.png 1179w\" sizes=\"auto, (max-width: 300px) 100vw, 300px\" \/><\/a><\/p>\n<p>gpu \u304c\u4e00\u3064\u306e\u5834\u5408\u306f\u4e0b\u8a18\u3002\u3053\u308c\u306f\u51fa\u529b\u306e\u6700\u5f8c\u306e\u90e8\u5206\u3060\u3051\u3042\u3052\u307e\u3059\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">Step\tImg\/sec\ttotal_loss\r\n1\timages\/sec: 196.7 +\/- 0.0 (jitter = 0.0)\t7.765\r\n10\timages\/sec: 199.5 +\/- 0.7 (jitter = 1.7)\t8.049\r\n20\timages\/sec: 199.7 +\/- 0.5 (jitter = 1.7)\t7.808\r\n30\timages\/sec: 199.5 +\/- 0.6 (jitter = 1.3)\t7.976\r\n40\timages\/sec: 199.1 +\/- 0.6 (jitter = 1.5)\t7.591\r\n50\timages\/sec: 199.3 +\/- 0.5 (jitter = 1.5)\t7.549\r\n60\timages\/sec: 199.4 +\/- 0.4 (jitter = 1.5)\t7.819\r\n70\timages\/sec: 198.3 +\/- 0.7 (jitter = 1.9)\t7.820\r\n80\timages\/sec: 194.8 +\/- 1.1 (jitter = 2.7)\t7.847\r\n90\timages\/sec: 192.0 +\/- 1.2 (jitter = 4.2)\t8.028\r\n100\timages\/sec: 190.0 +\/- 1.3 (jitter = 5.8)\t8.028\r\n----------------------------------------------------------------\r\ntotal images\/sec: 189.92\r\n----------------------------------------------------------------<\/code><\/pre>\n<p>\u6570\u5024\u306f\u534a\u5206\u3050\u3089\u3044\u306b\u306a\u3063\u3066\u3044\u308b\u3002\u3053\u306e\u30d9\u30f3\u30c1\u30de\u30fc\u30af\u306f tensorflow 1 \u7528\u3060\u3068\u66f8\u3044\u3066\u3042\u3063\u305f\u304c\u3001\u4e00\u5fdc\u52d5\u3044\u3066\u3044\u308b\u3088\u3046\u306b\u898b\u3048\u308b\u3002<\/p>\n<p>(20200901)<br \/>\ndocker \u306e\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u3002virtualbox \u4e0a\u306e ubuntu 20.04 \u3078\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u3057\u3066\u307f\u305f\u3002GPU \u304c\u4f7f\u3048\u308b\u308f\u3051\u3067\u306f\u306a\u3044\u304c\u3001\u3068\u308a\u3042\u3048\u305a\uff0cdocker \u306b\u6163\u308c\u308b\u3053\u3068\u304c\u76ee\u7684\u3002<\/p>\n<p>\u53c2\u8003\uff1a<a href=\"https:\/\/qiita.com\/kei0425\/items\/73623947921e84366fb1\">Ubuntu 20.04 LTS \u306b docker \u3092\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u3059\u308b<\/a><\/p>\n<p>\u30b3\u30de\u30f3\u30c9\u3092\u8a18\u9332\u3059\u308b\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">sudo apt install docker-compose<\/code><\/pre>\n<p>\u4e0b\u8a18\u306f root \u30e6\u30fc\u30b6\u30fc\u3067\u306a\u304f\u3066\u3082\u5b9f\u884c\u3067\u304d\u308b\u3088\u3046\u306b\uff0c\u81ea\u5206\u3092\u30b0\u30eb\u30fc\u30d7\u306b\u5165\u308c\u308b\u305f\u3081\u306e\u3082\u306e\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">sudo gpasswd -a \u30e6\u30fc\u30b6\u30fc\u540d docker<\/code><\/pre>\n<p>(20200902)<br \/>\ndocker \u3067 wordpress \u3092\u8d77\u52d5\u3059\u308b\u3002<\/p>\n<p>\u53c2\u8003\u30b5\u30a4\u30c8\uff1a<a href=\"https:\/\/qiita.com\/maimax\/items\/5090fd37f9832fe3c5fd\">\u4eca\u66f4\u3060\u3051\u3069Docker\u3067WordPress\u74b0\u5883\u3092\u7528\u610f\u3057\u3066\u307f\u305f\u3089\u8d85\u7c21\u5358\u3060\u3063\u305f<\/a><\/p>\n<p>\u30a4\u30e1\u30fc\u30b8\u3092\u6301\u3063\u3066\u304f\u308b\u3002<code>docker pull<\/code> \u3067\u3001mysql \u3068 WORDPRESS \u306e2\u3064\u306e\u30d5\u30a1\u30a4\u30eb\u3092\u6301\u3063\u3066\u304f\u308b\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">~$ docker pull mysql:5.7.25\r\n5.7.25: Pulling from library\/mysql\r\nDigest: sha256:dba5fed182e64064b688ccd22b2f9cad4ee88608c82f8cff21e17bab8da72b81\r\nStatus: Image is up to date for mysql:5.7.25\r\ndocker.io\/library\/mysql:5.7.25\r\n~$ docker pull wordpress\r\nUsing default tag: latest\r\nlatest: Pulling from library\/wordpress\r\nbf5952930446: Pull complete \r\na409b57eb464: Pull complete \r\n3192e6c84ad0: Pull complete \r\n43553740162b: Pull complete \r\nd8b8bba42dea: Pull complete \r\neb10907c0110: Pull complete \r\n10568906f34e: Pull complete \r\n03fe17709781: Pull complete \r\n98171b7166c8: Pull complete \r\n4a1bb352c362: Pull complete \r\ncfbcb1b22459: Pull complete \r\n9c47da96c73c: Pull complete \r\nd5ff66b2340d: Pull complete \r\n1a9d629afb81: Pull complete \r\n7491b4c1cf25: Pull complete \r\ncfd1d61e1215: Pull complete \r\n9dc8914ad89c: Pull complete \r\n5e36ed3f63b0: Pull complete \r\ndfdac20bfc12: Pull complete \r\n5221e8aad98a: Pull complete \r\nDigest: sha256:37f77cf9a9cd50291b3550a745872603370b569d4b74eaea4e08f22753ea4179\r\nStatus: Downloaded newer image for wordpress:latest\r\ndocker.io\/library\/wordpress:latest<\/code><\/pre>\n<p>\u30b3\u30f3\u30c6\u30ca\u306e\u8d77\u52d5\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">~$ docker run --name my_mysql -e MYSQL_ROOT_PASSWORD=xxxxx -d mysql:5.7.25\r\n6a7ca88bccc13ad287a9be44cd81c6c22c49628e7c28c81ece33191845f6cd8e<\/code><\/pre>\n<p>\u7d9a\u3051\u3066\u3001\u3082\u3046\u3072\u3068\u3064\u306e\u30b3\u30f3\u30c6\u30ca\u3082\u8d77\u52d5\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">~$ docker run -e WORDPRESS_DB_PASSWORD=xxxxx --link my_mysql:mysql -d -p 8080:80 wordpress\r\nee5fa915cc7a26b32f6a4847b6616be18c01d3a579d9f79739f54acc16280b9b<\/code><\/pre>\n<p><code>localhost:8080<\/code> \u306b\u30a2\u30af\u30bb\u30b9\u3059\u308b\u3068 wordpress \u304c\u52d5\u3044\u3066\u3044\u308b\u3002<\/p>\n<p><a href=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-03-13-24-52.png\"><img loading=\"lazy\" decoding=\"async\" src=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-03-13-24-52-300x225.png\" alt=\"\" width=\"300\" height=\"225\" class=\"aligncenter size-medium wp-image-9366\" srcset=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-03-13-24-52-300x225.png 300w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-03-13-24-52-768x576.png 768w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-03-13-24-52-1024x768.png 1024w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-03-13-24-52-600x450.png 600w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-03-13-24-52.png 1400w\" sizes=\"auto, (max-width: 300px) 100vw, 300px\" \/><\/a><\/p>\n<p>(20200903)<br \/>\n\u9806\u4e0d\u540c\u3067\u3042\u308b\u304c\u3001\u30b3\u30de\u30f3\u30c9\u3092\u8a18\u9332\u3002<\/p>\n<p>\u4e00\u5ea6\u3001ubuntu \u3092\u7d42\u4e86\u3057\u3066\u3001\u518d\u5ea6 docker \u3067 wordpress \u3092\u8d77\u52d5\u3057\u3088\u3046\u3068\u3057\u305f\u3089\u3001\u4e0b\u8a18\u306e\u30a8\u30e9\u30fc\u3002\u540d\u524d\u304c\u91cd\u306a\u3063\u3066\u3044\u308b\u3068\u3044\u3046\u3088\u3046\u306a\u5185\u5bb9\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">docker: Error response from daemon: Conflict. The container name \"\/my_mysql\" is already in use by container \"1d4939229ad1edbc4b2e4ad30dc54e8a4ee2c93155c4a76d838cea739fa5a241\". You have to remove (or rename) that container to be able to reuse that name.\r\nSee 'docker run --help'.<\/code><\/pre>\n<p>\u305d\u3053\u3067\u3001\u73fe\u5728\u306e\u72b6\u6cc1\u3092\u8abf\u67fb\u3059\u308b\uff08\u53c2\u8003\u30b5\u30a4\u30c8\uff1a<a href=\"https:\/\/qiita.com\/tifa2chan\/items\/e9aa408244687a63a0ae\">Docker\u30a4\u30e1\u30fc\u30b8\u3068\u30b3\u30f3\u30c6\u30ca\u306e\u524a\u9664\u65b9\u6cd5<\/a>\uff09<\/p>\n<p>\u505c\u6b62\u3057\u3066\u3044\u308b\u30b3\u30f3\u30c6\u30ca\u306e\u78ba\u8a8d\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">docker ps -a\r\nCONTAINER ID        IMAGE               COMMAND                  CREATED             STATUS                      PORTS               NAMES\r\na22352706942        wordpress           \"docker-entrypoint.s\u2026\"   16 minutes ago      Exited (0) 10 minutes ago                       busy_gates\r\n1d4939229ad1        mysql:5.7.25        \"docker-entrypoint.s\u2026\"   17 minutes ago      Exited (0) 10 minutes ago                       my_mysql<\/code><\/pre>\n<p>2\u3064\u3042\u308b\u306e\u3067\u3001\u4e21\u65b9\u3068\u3082\u524a\u9664\u3059\u308b\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">~$ docker rm a22352706942\r\na22352706942\r\n~$ docker rm 1d4939229ad1\r\n1d4939229ad1<\/code><\/pre>\n<p>\u3053\u306e\u3042\u3068\u3001\u524d\u56de\u3068\u540c\u3058\u3088\u3046\u306b\uff08\u540c\u3058\u540d\u524d\u3067\uff09\u30b3\u30f3\u30c6\u30ca\u3092\u8d77\u52d5\u3067\u304d\u305f\u3002<\/p>\n<p>\u30b3\u30f3\u30c6\u30ca\u306e\u505c\u6b62\u3002\u540d\u524d\u3067\u6307\u5b9a\u3067\u304d\u308b\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">~$ sudo docker stop tender_kare\r\ntender_kare\r\n~$ sudo docker stop my_mysql \r\nmy_mysql<\/code><\/pre>\n<p>docker \u30a4\u30e1\u30fc\u30b8\u306e\u524a\u9664\u3092\u3059\u308b\u3002\u5148\u306b\u30a4\u30e1\u30fc\u30b8\u3092\u5229\u7528\u3057\u3066\u3044\u308b\u30b3\u30f3\u30c6\u30ca\u3092\u524a\u9664\u3057\u3066\u304a\u304f\uff08\u4e0a\u306e\u65b9\u306e\u8a18\u8ff0\u3092\u53c2\u7167\uff09\u3002\u305d\u306e\u5f8c\u3001\u30a4\u30e1\u30fc\u30b8\u306e\u78ba\u8a8d\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">~$ sudo docker images\r\nREPOSITORY          TAG                 IMAGE ID            CREATED             SIZE\r\nwordpress           latest              6158ccbb8924        28 hours ago        546MB\r\nmysql               5.7.25              98455b9624a9        17 months ago       372MB<\/code><\/pre>\n<p>ID \u3092\u6307\u5b9a\u3057\u3066\u524a\u9664\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">~$ sudo docker rmi 6158ccbb8924\r\nUntagged: wordpress:latest\r\nUntagged: wordpress@sha256:93ee786387237f25705610977d5f506c87ea99b1f207aa2441a027b2b5f8a7a2\r\nDeleted: sha256:6158ccbb892411687b23ac4fd6d7c8f4e35fce2caa8118f8e3ce934cbd99b4dd\r\nDeleted: sha256:0dd131695f295133a6f1f2b90f066d8a222768e3b72a73912c4adf04c52396b4\r\nDeleted: sha256:109467c335d2553c688db740d686bc4cc71198a5c4f859118ca197757c9ae703\r\nDeleted: sha256:a20b304f3c4c87a25f8c5ea1dc582e0ed3cb3b16f85ae65de322ca505109eae4\r\nDeleted: sha256:4a0c89e004a146c22de1a97e17fe857989b51946f59dce2d83fc17496f9a14bb\r\nDeleted: sha256:a8d2a9f9cf582707834fc9025fa48aad65ba7d4d7b4999f530b99450dec3929a\r\nDeleted: sha256:bf9b010ae680378a86fbc021505bd69922c4942010b2f56e7dec2b59599eebb2\r\nDeleted: sha256:6166b82101bfae46ad9b3fa68ed0b2c008ab3da54c4692daf34b40139b55563e\r\nDeleted: sha256:8324edb0d0acf5ea5343f9a3d4bd5a6596711265d2598256c2eb1037cd633466\r\nDeleted: sha256:a54b1508ba4eadc13394c78555693d3a89771448ee71da33de64a5928759a4e7\r\nDeleted: sha256:df1f5315ab04605d1a13d64ff4b892be69d4d89336b85a1214b30cdd47c39b8f\r\nDeleted: sha256:27e2fcb0c233ca7730507439deea516e1192dcc1468a878775e7acb6c82df77b\r\nDeleted: sha256:56a340d84f17f5b4e83bd02ffd40a3b1273912dc977e450500fc224f61d43eb9\r\nDeleted: sha256:f8899bfcfa880bd6caa5078c27b488a37f6abdce21df829ad9a7b831589bac28\r\nDeleted: sha256:08b50110935f318104dd652795f0b0a2c6d007b368230adb8779c0a235c0b0f4\r\nDeleted: sha256:3bcc29238c6f1ce1602b88b2317fd7429228c311b624274a54e05e8192569ead\r\nDeleted: sha256:e7854f919e1ba15b6ac0b0e70cd2cb1eac7ca4cdcdd876c341f80b7b319ac395\r\nDeleted: sha256:45c01c7d544fd67d36753507065e5458be16dfcf82e0e31e321c2f84e81d7d4d\r\nDeleted: sha256:8e311722d5028e3cb61f5dea8786ed5fbc54a3e3ed4da94833a62749ebb6e7e8\r\nDeleted: sha256:2f245d38723aa315787ea341d92c1abbcccc9cb6b0b32fde7dd02555def26970\r\nDeleted: sha256:d0f104dc0a1f9c744b65b23b3fd4d4d3236b4656e67f776fe13f8ad8423b955c<\/code><\/pre>\n<p>\u3044\u308d\u3044\u308d\u306a\u30d5\u30a1\u30a4\u30eb\u304c\u4e00\u7dd2\u306b\u6d88\u3055\u308c\u308b\u3002<\/p>\n<p>(20200908)<br \/>\ndocker \u3067 wordpress \u3092\u52d5\u304b\u3059\u7d9a\u304d\u3002wordpress \u306e\u8a2d\u5b9a\u5909\u66f4\u7b49\u3092\u7dad\u6301\u3057\u3066\u304a\u304f\u3088\u3046\u306b\u8a66\u307f\u308b\u3002<\/p>\n<p>\u53c2\u8003\u30b5\u30a4\u30c8\uff1a<a href=\"https:\/\/super-yusuke.gitbook.io\/wordpress-memo\/dtano-docker-volume-ri\">\u30c7\u30fc\u30bf\u306e\u6c38\u7d9a\u5316 docker volume \u5468\u308a<\/a><\/p>\n<p>\u4e0a\u8a18\u306e\u30b5\u30a4\u30c8\u3092\u53c2\u8003\u306b\u3057\u3066\u3001\u307e\u305a\u306f\u7df4\u7fd2\u306a\u3069\u3002\u30b3\u30de\u30f3\u30c9\u3092\u8a18\u9332\u3057\u307e\u3059\u3002\u4e8b\u524d\u306b\u3001\u30b3\u30f3\u30c6\u30ca\u5185\u306b\u7528\u610f\u3059\u308b\u30d5\u30a9\u30eb\u30c0\u30fc\u3068\u5bfe\u5fdc\u3059\u308b volume \u306a\u308b\u3082\u306e\u3092\u4f5c\u6210\u3059\u308b\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">$ docker volume create --name mysqldata\r\nmysqldata\r\n$ docker volume ls\r\nDRIVER              VOLUME NAME\r\nlocal               3a2cbd5317d0a7515fce6a69563b431ddba4acec5a03834bae6afb4ba2fc6a4e\r\nlocal               7c6b8cb4ffcba1f4786af18808c456e94d74be8bfc5e8b12518b174803efb2f4\r\nlocal               30dc886e128e24e23068df50e84f408ca1d2445d7f96da184bfd4e2fe356df47\r\nlocal               5833df75afe37f9a16bf2fcad7f66a454e46feec04a76195d776e6efdfbba818\r\nlocal               ad84ae9e906dac7e2f614088387dd8aca7e1014f7251a74017416ed924c3c803\r\nlocal               e3205bd67a0dcb8678186239f7b9d3a1ec891671e056b9c65b4ce4685c27e8be\r\nlocal               mysqldata<\/code><\/pre>\n<p>\u3044\u3064\u306e\u9593\u306b\u304b\u8272\u3005\u306a volume \u304c\u4f5c\u3089\u308c\u3066\u3044\u307e\u3059\u304c\u3001mysqldata \u3082\u5728\u308a\u307e\u3059\u3002mysqldata \u306e\u8a73\u7d30\u3092\u898b\u308b\u3068<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">$ docker volume inspect mysqldata \r\n[\r\n    {\r\n        \"CreatedAt\": \"2020-09-08T13:13:36+09:00\",\r\n        \"Driver\": \"local\",\r\n        \"Labels\": {},\r\n        \"Mountpoint\": \"\/var\/lib\/docker\/volumes\/mysqldata\/_data\",\r\n        \"Name\": \"mysqldata\",\r\n        \"Options\": {},\r\n        \"Scope\": \"local\"\r\n    }\r\n]<\/code><\/pre>\n<p>\u3068\u308a\u3042\u3048\u305a\u3001ubuntu \u3067\u8a66\u3057\u307e\u3059\u3002\u4e0a\u8a18\u306e volume \u3092\u30de\u30a6\u30f3\u30c8\u3057\u305f\u3001ubuntu \u30b3\u30f3\u30c6\u30ca\u3092\u8d77\u52d5\u3057\u307e\u3059\u3002\u30b3\u30f3\u30c6\u30ca\u3067\u306e\u30de\u30a6\u30f3\u30c8\u5148\u3092\u6307\u5b9a\u3057\u307e\u3059\u3002\u3061\u3083\u3093\u3068\u3042\u308b\u304b\u3069\u3046\u304b\u63a2\u3057\u3066\u307f\u307e\u3057\u305f\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">$ docker run -it -v mysqldata:\/home\/mysqldata ubuntu:latest \/bin\/bash\r\nroot@ca95c4a8e194:\/# ls\r\nbin  boot  dev  etc  home  lib  lib32  lib64  libx32  media  mnt  opt  proc  root  run  sbin  srv  sys  tmp  usr  var\r\nroot@ca95c4a8e194:\/# cd home\/\r\nroot@ca95c4a8e194:\/home# ls\r\nmysqldata<\/code><\/pre>\n<p>\u30d5\u30a9\u30eb\u30c0\u30fc\u304c\u5728\u308a\u307e\u3057\u305f\u3002<\/p>\n<p>\u7d9a\u3044\u3066\u3001wordpress \u306e\u8d77\u52d5\u3067\u3059\u3002docker-compose.yml \u3068\u3044\u3046\u30d5\u30a1\u30a4\u30eb\u3092\u4f5c\u6210\u3002\u5185\u5bb9\u306f\u4e0b\u8a18\u3002\u6700\u521d\u30a8\u30e9\u30fc\u304c\u751f\u3058\u3066\u4e0a\u624b\u304f\u4f7f\u7528\u3067\u304d\u306a\u304b\u3063\u305f\u3002\u534a\u89d2\u30b9\u30da\u30fc\u30b9\u3092\u633f\u5165\u3059\u308b\u304b\u3069\u3046\u304b\u304c\u554f\u984c\u3060\u3063\u305f\uff08\u53c2\u8003\u30b5\u30a4\u30c8\uff1a<a href=\"http:\/\/programs.blog.jp\/archives\/1043461932.html\">YAMLError: mapping values are not allowed here<\/a>\uff09\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">version: '3.3'\r\n\r\nservices:\r\n   db:\r\n     image: mysql:5.7.25\r\n     volumes: # \u4e0a\u306e\u968e\u5c64\u3067\u6307\u5b9a\u3057\u305f db_data \u3092\u4f7f\u3044\u307e\u3059\u3088\r\n       - db_data:\/var\/lib\/mysql\r\n     restart: always\r\n     environment:\r\n       MYSQL_ROOT_PASSWORD: rootpw\r\n       MYSQL_DATABASE: wordpress\r\n       MYSQL_USER: wordpress\r\n       MYSQL_PASSWORD: wordpresspw\r\n\r\n   wordpress:\r\n     depends_on:\r\n       - db\r\n     image: wordpress:latest\r\n     ports:\r\n       - \"8080:80\"\r\n     restart: always\r\n     volumes:\r\n       - .\/htmldata\/:\/var\/www\/html\/\r\n     environment:\r\n       WORDPRESS_DB_HOST: db:3306\r\n       WORDPRESS_DB_USER: wordpress\r\n       WORDPRESS_DB_PASSWORD: wordpresspw\r\nvolumes: \r\n    db_data: # \u307e\u305a\u3053\u306e\u540d\u524d\u3067 volume \u3092\u4f5c\u308b<\/code><\/pre>\n<p>docker \u306e volume \u306f\u4e8b\u524d\u306b\u6e96\u5099\u3057\u306a\u304b\u3063\u305f\u3002\u4e0b\u8a18\u306e\u30b3\u30de\u30f3\u30c9\u3067\u30b3\u30f3\u30c6\u30ca\u3092\u8d77\u52d5\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">docker-compose up<\/code><\/pre>\n<p>\u3053\u306e\u8fd4\u4e8b\u306f\u76f8\u5f53\u9577\u3044\u3082\u306e\u3060\u3063\u305f\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">$ docker-compose up\r\nCreating network \"docker_default\" with the default driver\r\nCreating docker_db_1 ... done\r\nCreating docker_wordpress_1 ... done\r\nAttaching to docker_db_1, docker_wordpress_1\r\ndb_1         | 2020-09-08T07:18:14.490866Z 0 [Warning] TIMESTAMP with implicit DEFAULT value is deprecated. Please use --explicit_defaults_for_timestamp server option (see documentation for more details).\r\ndb_1         | 2020-09-08T07:18:14.491939Z 0 [Note] mysqld (mysqld 5.7.25) starting as process 1 ...\r\ndb_1         | 2020-09-08T07:18:14.495653Z 0 [Note] InnoDB: PUNCH HOLE support available\r\ndb_1         | 2020-09-08T07:18:14.495692Z 0 [Note] InnoDB: Mutexes and rw_locks use GCC atomic builtins\r\ndb_1         | 2020-09-08T07:18:14.495704Z 0 [Note] InnoDB: Uses event mutexes\r\ndb_1         | 2020-09-08T07:18:14.495715Z 0 [Note] InnoDB: GCC builtin __atomic_thread_fence() is used for memory barrier\r\ndb_1         | 2020-09-08T07:18:14.495725Z 0 [Note] InnoDB: Compressed tables use zlib 1.2.11\r\ndb_1         | 2020-09-08T07:18:14.495735Z 0 [Note] InnoDB: Using Linux native AIO\r\ndb_1         | 2020-09-08T07:18:14.496020Z 0 [Note] InnoDB: Number of pools: 1\r\ndb_1         | 2020-09-08T07:18:14.496161Z 0 [Note] InnoDB: Using CPU crc32 instructions\r\ndb_1         | 2020-09-08T07:18:14.498570Z 0 [Note] InnoDB: Initializing buffer pool, total size = 128M, instances = 1, chunk size = 128M\r\ndb_1         | 2020-09-08T07:18:14.508490Z 0 [Note] InnoDB: Completed initialization of buffer pool\r\ndb_1         | 2020-09-08T07:18:14.510028Z 0 [Note] InnoDB: If the mysqld execution user is authorized, page cleaner thread priority can be changed. See the man page of setpriority().\r\ndb_1         | 2020-09-08T07:18:14.529597Z 0 [Note] InnoDB: Highest supported file format is Barracuda.\r\ndb_1         | 2020-09-08T07:18:14.622849Z 0 [Note] InnoDB: Creating shared tablespace for temporary tables\r\ndb_1         | 2020-09-08T07:18:14.622896Z 0 [Note] InnoDB: Setting file '.\/ibtmp1' size to 12 MB. Physically writing the file full; Please wait ...\r\ndb_1         | 2020-09-08T07:18:14.835131Z 0 [Note] InnoDB: File '.\/ibtmp1' size is now 12 MB.\r\ndb_1         | 2020-09-08T07:18:14.836191Z 0 [Note] InnoDB: 96 redo rollback segment(s) found. 96 redo rollback segment(s) are active.\r\ndb_1         | 2020-09-08T07:18:14.836216Z 0 [Note] InnoDB: 32 non-redo rollback segment(s) are active.\r\ndb_1         | 2020-09-08T07:18:14.836617Z 0 [Note] InnoDB: 5.7.25 started; log sequence number 13734298\r\ndb_1         | 2020-09-08T07:18:14.836749Z 0 [Note] InnoDB: Loading buffer pool(s) from \/var\/lib\/mysql\/ib_buffer_pool\r\ndb_1         | 2020-09-08T07:18:14.836838Z 0 [Note] Plugin 'FEDERATED' is disabled.\r\ndb_1         | 2020-09-08T07:18:14.841254Z 0 [Note] Found ca.pem, server-cert.pem and server-key.pem in data directory. Trying to enable SSL support using them.\r\ndb_1         | 2020-09-08T07:18:14.841525Z 0 [Warning] CA certificate ca.pem is self signed.\r\ndb_1         | 2020-09-08T07:18:14.842984Z 0 [Note] Server hostname (bind-address): '*'; port: 3306\r\ndb_1         | 2020-09-08T07:18:14.843226Z 0 [Note] IPv6 is available.\r\ndb_1         | 2020-09-08T07:18:14.843239Z 0 [Note]   - '::' resolves to '::';\r\ndb_1         | 2020-09-08T07:18:14.843256Z 0 [Note] Server socket created on IP: '::'.\r\ndb_1         | 2020-09-08T07:18:14.844321Z 0 [Note] InnoDB: Buffer pool(s) load completed at 200908  7:18:14\r\ndb_1         | 2020-09-08T07:18:14.847380Z 0 [Warning] Insecure configuration for --pid-file: Location '\/var\/run\/mysqld' in the path is accessible to all OS users. Consider choosing a different directory.\r\ndb_1         | 2020-09-08T07:18:14.848291Z 0 [Warning] 'user' entry 'root@localhost' ignored in --skip-name-resolve mode.\r\ndb_1         | 2020-09-08T07:18:14.848315Z 0 [Warning] 'user' entry 'mysql.session@localhost' ignored in --skip-name-resolve mode.\r\ndb_1         | 2020-09-08T07:18:14.848323Z 0 [Warning] 'user' entry 'mysql.sys@localhost' ignored in --skip-name-resolve mode.\r\ndb_1         | 2020-09-08T07:18:14.848371Z 0 [Warning] 'db' entry 'performance_schema mysql.session@localhost' ignored in --skip-name-resolve mode.\r\ndb_1         | 2020-09-08T07:18:14.848377Z 0 [Warning] 'db' entry 'sys mysql.sys@localhost' ignored in --skip-name-resolve mode.\r\ndb_1         | 2020-09-08T07:18:14.848388Z 0 [Warning] 'proxies_priv' entry '@ root@localhost' ignored in --skip-name-resolve mode.\r\ndb_1         | 2020-09-08T07:18:14.861849Z 0 [Warning] 'tables_priv' entry 'user mysql.session@localhost' ignored in --skip-name-resolve mode.\r\ndb_1         | 2020-09-08T07:18:14.861869Z 0 [Warning] 'tables_priv' entry 'sys_config mysql.sys@localhost' ignored in --skip-name-resolve mode.\r\ndb_1         | 2020-09-08T07:18:14.890137Z 0 [Note] Event Scheduler: Loaded 0 events\r\ndb_1         | 2020-09-08T07:18:14.901595Z 0 [Note] mysqld: ready for connections.\r\ndb_1         | Version: '5.7.25'  socket: '\/var\/run\/mysqld\/mysqld.sock'  port: 3306  MySQL Community Server (GPL)\r\nwordpress_1  | AH00558: apache2: Could not reliably determine the server's fully qualified domain name, using 172.19.0.3. Set the 'ServerName' directive globally to suppress this message\r\nwordpress_1  | AH00558: apache2: Could not reliably determine the server's fully qualified domain name, using 172.19.0.3. Set the 'ServerName' directive globally to suppress this message\r\nwordpress_1  | [Tue Sep 08 07:18:15.507048 2020] [mpm_prefork:notice] [pid 1] AH00163: Apache\/2.4.38 (Debian) PHP\/7.4.9 configured -- resuming normal operations\r\nwordpress_1  | [Tue Sep 08 07:18:15.507097 2020] [core:notice] [pid 1] AH00094: Command line: 'apache2 -D FOREGROUND'<\/code><\/pre>\n<p>\u30b5\u30a4\u30c8\u306b\u30a2\u30af\u30bb\u30b9\u3059\u308b\u3068\u4e0a\u8a18\u306e\u30ed\u30b0\u306b\u8ffd\u52a0\u3055\u308c\u308b\u3002\u30ed\u30b0\u304c\u6a19\u6e96\u753b\u9762\u306b\u8868\u793a\u3055\u308c\u308b\u3088\u3046\u3060\u3002localhost:8080 \u306a\u3089\u52d5\u3044\u3066\u3044\u308b\u3002The Box \u3068\u3044\u3046\u30c6\u30fc\u30de\u3092\u5229\u7528\u3057\u3066\u307f\u305f\u3002<\/p>\n<p><a href=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200908a.png\"><img loading=\"lazy\" decoding=\"async\" src=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200908a-300x235.png\" alt=\"\" width=\"300\" height=\"235\" class=\"aligncenter size-medium wp-image-9376\" srcset=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200908a-300x235.png 300w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200908a-768x603.png 768w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200908a-1024x803.png 1024w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200908a-600x471.png 600w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200908a.png 1300w\" sizes=\"auto, (max-width: 300px) 100vw, 300px\" \/><\/a><\/p>\n<p>\u3044\u3063\u305f\u3093\u7d42\u4e86\u3059\u308b\u3002ctrl+c \u3067\u6b62\u3081\u3066\u304b\u3089\u3001\u4e0b\u8a18\u306e\u30b3\u30de\u30f3\u30c9\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">$ docker-compose down\r\nRemoving docker_wordpress_1 ... done\r\nRemoving docker_db_1        ... done\r\nRemoving network docker_default<\/code><\/pre>\n<p>\u3082\u3046\u4e00\u56de\u3001\u8d77\u52d5\u3057\u3066\u307f\u308b\u3002\u7d50\u679c\u3001The Box \u30c6\u30fc\u30de\u3092\u9069\u7528\u3055\u308c\u305f\u72b6\u614b\u3067\u8868\u793a\u3055\u308c\u305f\u3002\u5168\u304f\u540c\u3058\u753b\u9762\u306a\u306e\u3067\u30ad\u30e3\u30d7\u30c1\u30e3\u30fc\u306f\u7701\u7565\u3057\u307e\u3059\u3002\u4e00\u5fdc\u6c38\u7d9a\u5316\u306f\u3067\u304d\u305f\u3088\u3046\u3060\u3002<\/p>\n<p>localhost \u4ee5\u5916\u304b\u3089\u3060\u3068\uff0c\u8868\u793a\u304c\u304a\u304b\u3057\u3044\u3002css \u304c\u8aad\u307f\u8fbc\u307e\u308c\u3066\u3044\u306a\u3044\u3088\u3046\u306a\u8868\u793a\u306b\u306a\u3063\u305f\u3002<\/p>\n<p><a href=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200908b.png\"><img loading=\"lazy\" decoding=\"async\" src=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200908b-300x180.png\" alt=\"\" width=\"300\" height=\"180\" class=\"aligncenter size-medium wp-image-9383\" srcset=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200908b-300x180.png 300w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200908b-768x460.png 768w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200908b-1024x614.png 1024w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200908b-600x360.png 600w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200908b.png 1442w\" sizes=\"auto, (max-width: 300px) 100vw, 300px\" \/><\/a><\/p>\n<p>\u4e0a\u8a18\u306e\u75c7\u72b6\u306f wordpress \u306e\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u6642\uff0c\u30b5\u30fc\u30d0\u30fc\u306e\u30a2\u30c9\u30ec\u30b9\u304c\u30b3\u30f3\u30d5\u30a3\u30b0\u30d5\u30a1\u30a4\u30eb\u306b\u66f8\u304d\u8fbc\u307e\u308c\u308b\u969b\u306b\uff0clocalhost \u3068\u8a18\u5165\u3055\u308c\u3066\u3044\u308b\u6545\u3067\u306f\u306a\u3044\u304b\u3068\u8003\u3048\u3066\uff0c\u4e00\u65e6\uff0cmysql \u7528\u306e volume \u3092\u524a\u9664\u3057\u3066\u30c7\u30fc\u30bf\u30fc\u30d9\u30fc\u30b9\u3092\u521d\u671f\u5316\u3057\uff0c\u518d\u3073\u30b3\u30f3\u30c6\u30ca\u3092\u8d77\u52d5\u3057\u3066\uff0cwordpress \u306e\u518d\u8a2d\u5b9a\u3092\u8a66\u307f\u305f\u3002\u4eca\u56de\u306f localhost \u3067\u306f\u306a\u304f\uff0c\u5916\u90e8\u304b\u3089 IP \u30a2\u30c9\u30ec\u30b9\u3067\u30a2\u30af\u30bb\u30b9\u3059\u308b\u3002\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u3092\u7d42\u3048\u308b\u3068\uff0c\u5916\u90e8\u304b\u3089\u306e\u30a2\u30af\u30bb\u30b9\u306b\u5bfe\u3057\u3066\u3001\u4e0b\u56f3\u306e\u3088\u3046\u306b\u305d\u308c\u3089\u3057\u3044\u8868\u793a\u3068\u306a\u3063\u305f\u3002<\/p>\n<p><a href=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200909a.png\"><img loading=\"lazy\" decoding=\"async\" src=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200909a-300x190.png\" alt=\"\" width=\"300\" height=\"190\" class=\"aligncenter size-medium wp-image-9389\" srcset=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200909a-300x190.png 300w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200909a-768x486.png 768w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200909a-1024x649.png 1024w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200909a-600x380.png 600w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/20200909a.png 1413w\" sizes=\"auto, (max-width: 300px) 100vw, 300px\" \/><\/a><\/p>\n<p>(20200909)<br \/>\ndocker \u3067 GPU \u3092\u5229\u7528\u3059\u3079\u304f\u3001\u3082\u3068\u306e RTX2070super x 2 \u679a\u523a\u3057\u306e\u30d1\u30bd\u30b3\u30f3\u306b\u623b\u308b\u3002\u74b0\u5883\u3092\u78ba\u8a8d\u3057\u307e\u3059\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">$ nvidia-smi\r\nFri Sep 11 14:26:39 2020       \r\n+-----------------------------------------------------------------------------+\r\n| NVIDIA-SMI 450.51.06    Driver Version: 450.51.06    CUDA Version: 11.0     |\r\n|-------------------------------+----------------------+----------------------+\r\n| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |\r\n| Fan  Temp  Perf  Pwr:Usage\/Cap|         Memory-Usage | GPU-Util  Compute M. |\r\n|                               |                      |               MIG M. |\r\n|===============================+======================+======================|\r\n|   0  GeForce RTX 207...  On   | 00000000:01:00.0  On |                  N\/A |\r\n|  0%   33C    P8    12W \/ 215W |    152MiB \/  7979MiB |      1%      Default |\r\n|                               |                      |                  N\/A |\r\n+-------------------------------+----------------------+----------------------+\r\n|   1  GeForce RTX 207...  On   | 00000000:02:00.0 Off |                  N\/A |\r\n|  0%   32C    P8     7W \/ 215W |      1MiB \/  7982MiB |      0%      Default |\r\n|                               |                      |                  N\/A |\r\n+-------------------------------+----------------------+----------------------+\r\n                                                                               \r\n+-----------------------------------------------------------------------------+\r\n| Processes:                                                                  |\r\n|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |\r\n|        ID   ID                                                   Usage      |\r\n|=============================================================================|\r\n|    0   N\/A  N\/A      1092      G   \/usr\/lib\/xorg\/Xorg                 82MiB |\r\n|    0   N\/A  N\/A      1228      G   \/usr\/bin\/gnome-shell               68MiB |\r\n+-----------------------------------------------------------------------------+<\/code><\/pre>\n<p>\u307e\u305a\u306f docker \u306e\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u304b\u3089\u3002\u4e0b\u8a18\u306e\u30b5\u30a4\u30c8\u306b\u5f93\u3063\u3066\u4f5c\u696d\u3002<\/p>\n<p>\u53c2\u8003\u30b5\u30a4\u30c8\uff1a<a href=\"https:\/\/medium.com\/nvidiajapan\/nvidia-docker-%E3%81%A3%E3%81%A6%E4%BB%8A%E3%81%A9%E3%81%86%E3%81%AA%E3%81%A3%E3%81%A6%E3%82%8B%E3%81%AE-20-09-%E7%89%88-558fae883f44\">NVIDIA Docker \u3063\u3066\u4eca\u3069\u3046\u306a\u3063\u3066\u308b\u306e\uff1f (20.09 \u7248)<\/a><\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">$ curl https:\/\/get.docker.com | sh\r\n  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current\r\n                                 Dload  Upload   Total   Spent    Left  Speed\r\n100 13857  100 13857    0     0  80563      0 --:--:-- --:--:-- --:--:-- 80563\r\n# Executing docker install script, commit: 26ff363bcf3b3f5a00498ac43694bf1c7d9ce16c\r\n+ sudo -E sh -c apt-get update -qq >\/dev\/null\r\n+ sudo -E sh -c DEBIAN_FRONTEND=noninteractive apt-get install -y -qq apt-transport-https ca-certificates curl >\/dev\/null\r\n+ sudo -E sh -c curl -fsSL \"https:\/\/download.docker.com\/linux\/ubuntu\/gpg\" | apt-key add -qq - >\/dev\/null\r\nWarning: apt-key output should not be parsed (stdout is not a terminal)\r\n+ sudo -E sh -c echo \"deb [arch=amd64] https:\/\/download.docker.com\/linux\/ubuntu bionic stable\" > \/etc\/apt\/sources.list.d\/docker.list\r\n+ sudo -E sh -c apt-get update -qq >\/dev\/null\r\n+ [ -n  ]\r\n+ sudo -E sh -c apt-get install -y -qq --no-install-recommends docker-ce >\/dev\/null\r\n+ sudo -E sh -c docker version\r\nClient: Docker Engine - Community\r\n Version:           19.03.12\r\n API version:       1.40\r\n Go version:        go1.13.10\r\n Git commit:        48a66213fe\r\n Built:             Mon Jun 22 15:45:36 2020\r\n OS\/Arch:           linux\/amd64\r\n Experimental:      false\r\n\r\nServer: Docker Engine - Community\r\n Engine:\r\n  Version:          19.03.12\r\n  API version:      1.40 (minimum version 1.12)\r\n  Go version:       go1.13.10\r\n  Git commit:       48a66213fe\r\n  Built:            Mon Jun 22 15:44:07 2020\r\n  OS\/Arch:          linux\/amd64\r\n  Experimental:     false\r\n containerd:\r\n  Version:          1.2.13\r\n  GitCommit:        7ad184331fa3e55e52b890ea95e65ba581ae3429\r\n runc:\r\n  Version:          1.0.0-rc10\r\n  GitCommit:        dc9208a3303feef5b3839f4323d9beb36df0a9dd\r\n docker-init:\r\n  Version:          0.18.0\r\n  GitCommit:        fec3683\r\nIf you would like to use Docker as a non-root user, you should now consider\r\nadding your user to the \"docker\" group with something like:\r\n\r\n  sudo usermod -aG docker friend\r\n\r\nRemember that you will have to log out and back in for this to take effect!\r\n\r\nWARNING: Adding a user to the \"docker\" group will grant the ability to run\r\n         containers which can be used to obtain root privileges on the\r\n         docker host.\r\n         Refer to https:\/\/docs.docker.com\/engine\/security\/security\/#docker-daemon-attack-surface\r\n         for more information.<\/code><\/pre>\n<p>\u7d9a\u3051\u3066\u4e0b\u8a18\u306e\u30b3\u30de\u30f3\u30c9\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">$ sudo systemctl start docker && sudo systemctl enable docker\r\nSynchronizing state of docker.service with SysV service script with \/lib\/systemd\/systemd-sysv-install.\r\nExecuting: \/lib\/systemd\/systemd-sysv-install enable docker<\/code><\/pre>\n<p>\u901a\u5e38\u306e\u30e6\u30fc\u30b6\u30fc\u3067 docker \u3092\u5229\u7528\u3067\u304d\u308b\u3088\u3046\u306b\u3001\u30b0\u30eb\u30fc\u30d7\u306b\u81ea\u5206\u3092\u8ffd\u52a0\u3059\u308b\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">$ sudo usermod -aG docker \u81ea\u5206<\/code><\/pre>\n<p>\u3042\u3068\u306f<a href=\"https:\/\/docs.nvidia.com\/datacenter\/cloud-native\/container-toolkit\/install-guide.html#installing-docker-ce\">\u3053\u3053<\/a>\u3092\u53c2\u7167\u3057\u306a\u304c\u3089\u4e0b\u8a18\u306e\u30b3\u30de\u30f3\u30c9\u3092\u6253\u3063\u305f\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">distribution=$(. \/etc\/os-release;echo $ID$VERSION_ID)\r\n\r\ncurl -s -L https:\/\/nvidia.github.io\/nvidia-docker\/gpgkey | sudo apt-key add -\r\n\r\ncurl -s -L https:\/\/nvidia.github.io\/nvidia-docker\/$distribution\/nvidia-docker.list | sudo tee \/etc\/apt\/sources.list.d\/nvidia-docker.list\r\n\r\nsudo apt-get update\r\n\r\nsudo apt-get install -y nvidia-docker2\r\n\r\nsudo systemctl restart docker<\/code><\/pre>\n<p>\u30c6\u30b9\u30c8\u7528\u306e\u30b3\u30f3\u30c6\u30ca\u3092\u52d5\u304b\u3059\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">$ sudo docker run --rm --gpus all nvidia\/cuda:11.0-base nvidia-smi\r\nUnable to find image 'nvidia\/cuda:11.0-base' locally\r\n11.0-base: Pulling from nvidia\/cuda\r\n54ee1f796a1e: Pull complete \r\nf7bfea53ad12: Pull complete \r\n46d371e02073: Pull complete \r\nb66c17bbf772: Pull complete \r\n3642f1a6dfb3: Pull complete \r\ne5ce55b8b4b9: Pull complete \r\n155bc0332b0a: Pull complete \r\nDigest: sha256:774ca3d612de15213102c2dbbba55df44dc5cf9870ca2be6c6e9c627fa63d67a\r\nStatus: Downloaded newer image for nvidia\/cuda:11.0-base\r\nWed Sep  9 09:15:49 2020       \r\n+-----------------------------------------------------------------------------+\r\n| NVIDIA-SMI 450.51.06    Driver Version: 450.51.06    CUDA Version: 11.0     |\r\n|-------------------------------+----------------------+----------------------+\r\n| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |\r\n| Fan  Temp  Perf  Pwr:Usage\/Cap|         Memory-Usage | GPU-Util  Compute M. |\r\n|                               |                      |               MIG M. |\r\n|===============================+======================+======================|\r\n|   0  GeForce RTX 207...  On   | 00000000:01:00.0  On |                  N\/A |\r\n|  0%   34C    P8    13W \/ 215W |    174MiB \/  7979MiB |     10%      Default |\r\n|                               |                      |                  N\/A |\r\n+-------------------------------+----------------------+----------------------+\r\n|   1  GeForce RTX 207...  On   | 00000000:02:00.0 Off |                  N\/A |\r\n|  0%   33C    P8     7W \/ 215W |      1MiB \/  7982MiB |      0%      Default |\r\n|                               |                      |                  N\/A |\r\n+-------------------------------+----------------------+----------------------+\r\n                                                                               \r\n+-----------------------------------------------------------------------------+\r\n| Processes:                                                                  |\r\n|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |\r\n|        ID   ID                                                   Usage      |\r\n|=============================================================================|\r\n+-----------------------------------------------------------------------------+<\/code><\/pre>\n<p>\u4e0a\u624b\u304f\u3044\u3063\u305f\uff1f<\/p>\n<p><a href=\"https:\/\/www.tensorflow.org\/install\/\">\u3053\u306e\u30b5\u30a4\u30c8<\/a>\u3092\u53c2\u8003\u306b\u3057\u3066\u3001\u3068\u308a\u3042\u3048\u305a\u4f55\u304b\u30b3\u30f3\u30c6\u30ca\u3092\u52d5\u304b\u3057\u3066\u307f\u308b\u3002\u30b3\u30de\u30f3\u30c9\u306f\u4e0b\u8a18\u3067\u3001jyupyter \u304c\u5229\u7528\u3067\u304d\u308b\u3082\u306e\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\"> docker pull tensorflow\/tensorflow:latest-py3\r\n docker run -it -p 8888:8888 tensorflow\/tensorflow:latest-py3-jupyter<\/code><\/pre>\n<p>\u5b9f\u884c\u3057\u305f\u3068\u304d\u306e\u30ec\u30b9\u30dd\u30f3\u30b9\u3002sudo \u3067\u5b9f\u884c\u3057\u305f\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">$ sudo docker run -it -p 8888:8888 tensorflow\/tensorflow:latest-py3-jupyter\r\nUnable to find image 'tensorflow\/tensorflow:latest-py3-jupyter' locally\r\nlatest-py3-jupyter: Pulling from tensorflow\/tensorflow\r\n2746a4a261c9: Already exists \r\n4c1d20cdee96: Already exists \r\n0d3160e1d0de: Already exists \r\nc8e37668deea: Already exists \r\ne52cad4ccd83: Already exists \r\ne97116da5f98: Already exists \r\n75c61371a2e3: Already exists \r\n8592f093fc78: Already exists \r\ndccb0709d7fb: Already exists \r\n107f0b841886: Already exists \r\nedc69fe5c6be: Already exists \r\n3d7f9e997aed: Pull complete \r\n1575375ec2e9: Pull complete \r\na574cd2a2ef5: Pull complete \r\na1565ebf3379: Pull complete \r\naf0d84cd6cdc: Pull complete \r\n8c1a10281be2: Pull complete \r\n649bf527b9db: Pull complete \r\n62895ac313e8: Pull complete \r\n0d2cfdddc1a6: Pull complete \r\na315501e4ca9: Pull complete \r\n146e7ce36cb8: Pull complete \r\ne638992c0d5d: Pull complete \r\nea6d34ce743b: Pull complete \r\n3bf310c11c24: Pull complete \r\ne4e0bb9d2283: Pull complete \r\nDigest: sha256:37709ed9fcb2e57132710d521b5a6f826bc022e9f137750cc19728a1533f08e1\r\nStatus: Downloaded newer image for tensorflow\/tensorflow:latest-py3-jupyter\r\n\r\n________                               _______________                \r\n___  __\/__________________________________  ____\/__  \/________      __\r\n__  \/  _  _ \\_  __ \\_  ___\/  __ \\_  ___\/_  \/_   __  \/_  __ \\_ | \/| \/ \/\r\n_  \/   \/  __\/  \/ \/ \/(__  )\/ \/_\/ \/  \/   _  __\/   _  \/ \/ \/_\/ \/_ |\/ |\/ \/ \r\n\/_\/    \\___\/\/_\/ \/_\/\/____\/ \\____\/\/_\/    \/_\/      \/_\/  \\____\/____\/|__\/\r\n\r\n\r\nWARNING: You are running this container as root, which can cause new files in\r\nmounted volumes to be created as the root user on your host machine.\r\n\r\nTo avoid this, run the container by specifying your user's userid:\r\n\r\n$ docker run -u $(id -u):$(id -g) args...\r\n\r\n[I 16:03:10.363 NotebookApp] Writing notebook server cookie secret to \/root\/.local\/share\/jupyter\/runtime\/notebook_cookie_secret\r\njupyter_http_over_ws extension initialized. Listening on \/http_over_websocket\r\n[I 16:03:10.503 NotebookApp] Serving notebooks from local directory: \/tf\r\n[I 16:03:10.503 NotebookApp] The Jupyter Notebook is running at:\r\n[I 16:03:10.503 NotebookApp] http:\/\/5e4fea31b956:8888\/?token=d7f9b426c52afcd12333f0aab8b132f30397a00d50a12240\r\n[I 16:03:10.503 NotebookApp]  or http:\/\/127.0.0.1:8888\/?token=d7f9b426c52afcd12333f0aab8b132f30397a00d50a12240\r\n[I 16:03:10.503 NotebookApp] Use Control-C to stop this server and shut down all kernels (twice to skip confirmation).\r\n[C 16:03:10.506 NotebookApp] \r\n    \r\n    To access the notebook, open this file in a browser:\r\n        file:\/\/\/root\/.local\/share\/jupyter\/runtime\/nbserver-1-open.html\r\n    Or copy and paste one of these URLs:\r\n        http:\/\/5e4fea31b956:8888\/?token=d7f9b426c52afcd12333f0aab8b132f30397a00d50a12240\r\n     or http:\/\/127.0.0.1:8888\/?token=d7f9b426c52afcd12333f0aab8b132f30397a00d50a12240\r\n[I 16:03:58.024 NotebookApp] 302 GET \/ (172.17.0.1) 0.42ms\r\n[I 16:03:58.026 NotebookApp] 302 GET \/tree? (172.17.0.1) 0.48ms\r\n[W 16:04:51.383 NotebookApp] 401 POST \/login?next=%2Ftree%3F (172.17.0.1) 0.94ms referer=http:\/\/localhost:8888\/login?next=%2Ftree%3F\r\n[W 16:07:36.535 NotebookApp] 401 POST \/login?next=%2Ftree%3F (172.17.0.1) 1.05ms referer=http:\/\/localhost:8888\/login?next=%2Ftree%3F\r\n[W 16:07:55.703 NotebookApp] 401 POST \/login?next=%2Ftree%3F (172.17.0.1) 0.89ms referer=http:\/\/localhost:8888\/login?next=%2Ftree%3F\r\n[W 16:08:25.930 NotebookApp] 401 POST \/login?next=%2Ftree%3F (172.17.0.1) 0.94ms referer=http:\/\/localhost:8888\/login?next=%2Ftree%3F\r\n[W 16:08:43.800 NotebookApp] 401 POST \/login?next=%2Ftree%3F (172.17.0.1) 0.90ms referer=http:\/\/localhost:8888\/login?next=%2Ftree%3F\r\n[W 16:09:21.194 NotebookApp] 401 POST \/login?next=%2Ftree%3F (172.17.0.1) 0.92ms referer=http:\/\/localhost:8888\/login?next=%2Ftree%3F\r\n[I 16:09:56.678 NotebookApp] 302 GET \/?token=c8de56fa...%20::%20\/Users\/you\/notebooks (172.17.0.1) 0.37ms\r\n[I 16:09:56.680 NotebookApp] 302 GET \/tree?token=c8de56fa...%20::%20\/Users\/you\/notebooks (172.17.0.1) 0.52ms\r\n[I 16:11:23.088 NotebookApp] 302 GET \/?token=d7f9b426c52afcd12333f0aab8b132f30397a00d50a12240 (172.17.0.1) 0.39ms\r\n[I 16:11:38.011 NotebookApp] Creating new notebook in \r\n[I 16:11:38.019 NotebookApp] Writing notebook-signing key to \/root\/.local\/share\/jupyter\/notebook_secret\r\n[I 16:11:38.539 NotebookApp] Kernel started: c0ee8cda-44be-4035-87d9-cad7cbc9c76a\r\n[I 16:13:38.534 NotebookApp] Saving file at \/Untitled.ipynb\r\n[I 02:03:39.300 NotebookApp] Saving file at \/Untitled.ipynb<\/code><\/pre>\n<p>\u51fa\u529b\u306e\u4e2d\u306b\u3042\u308b\u30ea\u30f3\u30af <code>http:\/\/127.0.0.1:8888\/?token=d7f9b426c52afcd12333f0aab8b132f30397a00d50a12240<\/code> \u3092\u30d6\u30e9\u30a6\u30b6\u3067\u958b\u304f\u3068\u3001jupyter notebook \u304c\u8868\u793a\u3055\u308c\u305f\u3002<\/p>\n<p><a href=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-10-11-17-17.png\"><img loading=\"lazy\" decoding=\"async\" src=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-10-11-17-17-300x216.png\" alt=\"\" width=\"300\" height=\"216\" class=\"aligncenter size-medium wp-image-9407\" srcset=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-10-11-17-17-300x216.png 300w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-10-11-17-17-768x553.png 768w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-10-11-17-17-1024x737.png 1024w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-10-11-17-17-600x432.png 600w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-10-11-17-17.png 1213w\" sizes=\"auto, (max-width: 300px) 100vw, 300px\" \/><\/a><\/p>\n<p><a href=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-11-13-31-17.png\"><img loading=\"lazy\" decoding=\"async\" src=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-11-13-31-17-300x216.png\" alt=\"\" width=\"300\" height=\"216\" class=\"aligncenter size-medium wp-image-9414\" srcset=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-11-13-31-17-300x216.png 300w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-11-13-31-17-768x553.png 768w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-11-13-31-17-1024x737.png 1024w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-11-13-31-17-600x432.png 600w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-11-13-31-17.png 1213w\" sizes=\"auto, (max-width: 300px) 100vw, 300px\" \/><\/a><\/p>\n<p>\u3053\u306e\u5148\u306f\u3001\u672c\u306b\u8f09\u3063\u3066\u3044\u308b\u30b3\u30fc\u30c9\u304c\u5b9f\u884c\u3067\u304d\u308b\u3088\u3046\u306a\u30b3\u30f3\u30c6\u30ca\u3092\u7528\u610f\u3059\u308b\u3053\u3068\u304c\u76ee\u6a19\u304b\u306a\uff1f tensorflow \u306e version 1 \u306e\u30b3\u30fc\u30c9\u304c\u52d5\u304f\u74b0\u5883\u304c\u3042\u308c\u3070\u3001\u672c\u306e\u30b3\u30fc\u30c9\u304c\u8a66\u305b\u308b\u3068\u601d\u3046\u3002<\/p>\n<p>(20200911)<\/p>\n<p>\u4e00\u65e6\u3001ubuntu18 \u3092\u5165\u308c\u76f4\u3059\u3002<a href=\"https:\/\/www.tensorflow.org\/install\/docker\">\u3053\u306e\u30da\u30fc\u30b8<\/a>\u306b\u3088\u308b\u3068\u3001\u30db\u30b9\u30c8\u306b\u306f\u30d3\u30c7\u30aa\u30ab\u30fc\u30c9\u306e\u30c9\u30e9\u30a4\u30d0\u30fc\u3060\u3051\u3092\u5165\u308c\u3066\u304a\u3051\u3070\u826f\u3044\u3068\u3044\u3046\u3053\u3068\u306a\u306e\u3067\u3001\u305d\u308c\u306b\u5408\u308f\u305b\u3066\u5165\u308c\u76f4\u3057\u3066\u307f\u308b\u3002\u30d3\u30c7\u30aa\u30ab\u30fc\u30c9\u306e\u30c9\u30e9\u30a4\u30d0\u30fc\u306f\u4e0b\u56f3\u306e\u8ffd\u52a0\u306e\u30c9\u30e9\u30a4\u30d0\u30fc\u304b\u3089\u5165\u308c\u3066\u307f\u305f\u3002<\/p>\n<p><a href=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-12-01-00-03.png\"><img loading=\"lazy\" decoding=\"async\" src=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-12-01-00-03-300x175.png\" alt=\"\" width=\"300\" height=\"175\" class=\"aligncenter size-medium wp-image-9421\" srcset=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-12-01-00-03-300x175.png 300w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-12-01-00-03-768x447.png 768w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-12-01-00-03-600x349.png 600w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-12-01-00-03.png 856w\" sizes=\"auto, (max-width: 300px) 100vw, 300px\" \/><\/a><\/p>\n<p>\u3053\u306e\u3042\u3068\u306f\u4ee5\u524d\u306e\u30b3\u30de\u30f3\u30c9\u306e\u7e70\u308a\u8fd4\u3057\u3068\u306a\u308b\u304c\u3001\u307e\u3068\u3081\u3066\u304a\u304d\u307e\u3059\uff08<a href=\"https:\/\/docs.nvidia.com\/datacenter\/cloud-native\/container-toolkit\/install-guide.html#installing-docker-c\">\u53c2\u8003\u30b5\u30a4\u30c8<\/a>\uff09\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">$ sudo apt install curl\r\n$ curl https:\/\/get.docker.com | sh\r\n$ sudo systemctl start docker && sudo systemctl enable docker\r\n$ sudo usermod -aG docker \u30e6\u30fc\u30b6\u30fc\u540d\r\n$ distribution=$(. \/etc\/os-release;echo $ID$VERSION_ID)\r\n$ curl -s -L https:\/\/nvidia.github.io\/nvidia-docker\/gpgkey | sudo apt-key add -\r\n$ curl -s -L https:\/\/nvidia.github.io\/nvidia-docker\/$distribution\/nvidia-docker.list | sudo tee \/etc\/apt\/sources.list.d\/nvidia-docker.list\r\n$ sudo apt-get update\r\n$ sudo apt-get install -y nvidia-docker2\r\n$ sudo systemctl restart docker<\/code><\/pre>\n<p>\u78ba\u8a8d\u3092\u3057\u3066\u307f\u308b\u3002CUDA \u3092\u5165\u308c\u305f\u61b6\u3048\u306f\u306a\u3044\u306e\u3060\u3051\u3069\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">$ nvidia-smi\r\nFri Sep 11 18:27:22 2020       \r\n+-----------------------------------------------------------------------------+\r\n| NVIDIA-SMI 440.100      Driver Version: 440.100      CUDA Version: 10.2     |\r\n|-------------------------------+----------------------+----------------------+\r\n| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |\r\n| Fan  Temp  Perf  Pwr:Usage\/Cap|         Memory-Usage | GPU-Util  Compute M. |\r\n|===============================+======================+======================|\r\n|   0  GeForce RTX 207...  Off  | 00000000:01:00.0  On |                  N\/A |\r\n|  0%   34C    P8    13W \/ 215W |    189MiB \/  7974MiB |      2%      Default |\r\n+-------------------------------+----------------------+----------------------+\r\n|   1  GeForce RTX 207...  Off  | 00000000:02:00.0 Off |                  N\/A |\r\n|  0%   33C    P8     7W \/ 215W |      1MiB \/  7982MiB |      0%      Default |\r\n+-------------------------------+----------------------+----------------------+\r\n                                                                               \r\n+-----------------------------------------------------------------------------+\r\n| Processes:                                                       GPU Memory |\r\n|  GPU       PID   Type   Process name                             Usage      |\r\n|=============================================================================|\r\n|    0       900      G   \/usr\/lib\/xorg\/Xorg                           113MiB |\r\n|    0      1083      G   \/usr\/bin\/gnome-shell                          70MiB |\r\n|    0      1903      G   \/usr\/lib\/firefox\/firefox                       2MiB |\r\n+-----------------------------------------------------------------------------+\r\n<\/code><\/pre>\n<p>docker \u306e\u30d0\u30fc\u30b8\u30e7\u30f3\u3092\u78ba\u8a8d\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">$ docker -v\r\nDocker version 19.03.12, build 48a66213fe<\/code><\/pre>\n<p>CUDA \u306e\u30d0\u30fc\u30b8\u30e7\u30f3\u3092\u78ba\u8a8d\u3059\u308b\u30b3\u30de\u30f3\u30c9\u3068\u3057\u3066 <code>nvcc -V<\/code> \u304c\u3042\u308b\u304c\u3001\u3053\u306e\u30b3\u30de\u30f3\u30c9\u306f\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u3055\u308c\u3066\u3044\u306a\u3044\u3088\u3046\u3067\u3042\u308b\u3002\u3064\u307e\u308a CUDA \u306f\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u3055\u308c\u3066\u3044\u306a\u3044\u3068\u3044\u3046\u3053\u3068\u3060\u308d\u3046\u304b\u3002<\/p>\n<p>\u3053\u308c\u3082\u7e70\u308a\u8fd4\u3057\u306b\u306a\u308b\u304c<a href=\"https:\/\/www.tensorflow.org\/install\/\">\u3053\u306e\u30b5\u30a4\u30c8<\/a>\u3092\u53c2\u8003\u306b\u3057\u3066\u3001jyupyter \u304c\u5229\u7528\u3067\u304d\u308b\u30b3\u30f3\u30c6\u30ca\u3092\u52d5\u304b\u3057\u3066\u307f\u308b\u3002\u30b3\u30de\u30f3\u30c9\u306f\u4e0b\u8a18\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\"> docker pull tensorflow\/tensorflow:latest-py3\r\n docker run -it -p 8888:8888 tensorflow\/tensorflow:latest-py3-jupyter<\/code><\/pre>\n<p>\u30b3\u30de\u30f3\u30c9\u306e\u51fa\u529b\u3092\u3042\u3052\u307e\u3059\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">$ docker pull tensorflow\/tensorflow:latest-py3\r\nlatest-py3: Pulling from tensorflow\/tensorflow\r\n2746a4a261c9: Pull complete \r\n4c1d20cdee96: Pull complete \r\n0d3160e1d0de: Pull complete \r\nc8e37668deea: Pull complete \r\ne52cad4ccd83: Pull complete \r\ne97116da5f98: Pull complete \r\n75c61371a2e3: Pull complete \r\n8592f093fc78: Pull complete \r\ndccb0709d7fb: Pull complete \r\n107f0b841886: Pull complete \r\nedc69fe5c6be: Pull complete \r\nDigest: sha256:14ec674cefd622aa9d45f07485500da254acaf8adfef80bd0f279db03c735689\r\nStatus: Downloaded newer image for tensorflow\/tensorflow:latest-py3\r\ndocker.io\/tensorflow\/tensorflow:latest-py3\r\n\r\n$ docker run -it -p 8888:8888 tensorflow\/tensorflow:latest-py3-jupyter\r\nUnable to find image 'tensorflow\/tensorflow:latest-py3-jupyter' locally\r\nlatest-py3-jupyter: Pulling from tensorflow\/tensorflow\r\n2746a4a261c9: Already exists \r\n4c1d20cdee96: Already exists \r\n0d3160e1d0de: Already exists \r\nc8e37668deea: Already exists \r\ne52cad4ccd83: Already exists \r\ne97116da5f98: Already exists \r\n75c61371a2e3: Already exists \r\n8592f093fc78: Already exists \r\ndccb0709d7fb: Already exists \r\n107f0b841886: Already exists \r\nedc69fe5c6be: Already exists \r\n3d7f9e997aed: Pull complete \r\n1575375ec2e9: Pull complete \r\na574cd2a2ef5: Pull complete \r\na1565ebf3379: Pull complete \r\naf0d84cd6cdc: Pull complete \r\n8c1a10281be2: Pull complete \r\n649bf527b9db: Pull complete \r\n62895ac313e8: Pull complete \r\n0d2cfdddc1a6: Pull complete \r\na315501e4ca9: Pull complete \r\n146e7ce36cb8: Pull complete \r\ne638992c0d5d: Pull complete \r\nea6d34ce743b: Pull complete \r\n3bf310c11c24: Pull complete \r\ne4e0bb9d2283: Pull complete \r\nDigest: sha256:37709ed9fcb2e57132710d521b5a6f826bc022e9f137750cc19728a1533f08e1\r\nStatus: Downloaded newer image for tensorflow\/tensorflow:latest-py3-jupyter\r\n\r\n________                               _______________                \r\n___  __\/__________________________________  ____\/__  \/________      __\r\n__  \/  _  _ \\_  __ \\_  ___\/  __ \\_  ___\/_  \/_   __  \/_  __ \\_ | \/| \/ \/\r\n_  \/   \/  __\/  \/ \/ \/(__  )\/ \/_\/ \/  \/   _  __\/   _  \/ \/ \/_\/ \/_ |\/ |\/ \/ \r\n\/_\/    \\___\/\/_\/ \/_\/\/____\/ \\____\/\/_\/    \/_\/      \/_\/  \\____\/____\/|__\/\r\n\r\n\r\nWARNING: You are running this container as root, which can cause new files in\r\nmounted volumes to be created as the root user on your host machine.\r\n\r\nTo avoid this, run the container by specifying your user's userid:\r\n\r\n$ docker run -u $(id -u):$(id -g) args...\r\n\r\n[I 17:01:04.729 NotebookApp] Writing notebook server cookie secret to \/root\/.local\/share\/jupyter\/runtime\/notebook_cookie_secret\r\njupyter_http_over_ws extension initialized. Listening on \/http_over_websocket\r\n[I 17:01:04.869 NotebookApp] Serving notebooks from local directory: \/tf\r\n[I 17:01:04.869 NotebookApp] The Jupyter Notebook is running at:\r\n[I 17:01:04.869 NotebookApp] http:\/\/4dbeafa44de8:8888\/?token=fbddcfe328f0511dff608cbcb182a58827ca9573930cb069\r\n[I 17:01:04.869 NotebookApp]  or http:\/\/127.0.0.1:8888\/?token=fbddcfe328f0511dff608cbcb182a58827ca9573930cb069\r\n[I 17:01:04.869 NotebookApp] Use Control-C to stop this server and shut down all kernels (twice to skip confirmation).\r\n[C 17:01:04.872 NotebookApp] \r\n    \r\n    To access the notebook, open this file in a browser:\r\n        file:\/\/\/root\/.local\/share\/jupyter\/runtime\/nbserver-1-open.html\r\n    Or copy and paste one of these URLs:\r\n        http:\/\/4dbeafa44de8:8888\/?token=fbddcfe328f0511dff608cbcb182a58827ca9573930cb069\r\n     or http:\/\/127.0.0.1:8888\/?token=fbddcfe328f0511dff608cbcb182a58827ca9573930cb069\r\n[I 17:02:15.433 NotebookApp] 302 GET \/?token=fbddcfe328f0511dff608cbcb182a58827ca9573930cb069 (172.17.0.1) 0.48ms\r\n<\/code><\/pre>\n<p>\u4e0a\u8a18\u306e\u51fa\u529b\u306e\u6700\u5f8c\u306e\u65b9\u306b\u3042\u308b URL <code>http:\/\/127.0.0.1:8888\/?token=fbddcfe328f0511dff608cbcb182a58827ca9573930cb069<\/code> \u3092\u30d6\u30e9\u30a6\u30b6\u3067\u958b\u304f\u3068\u3001jupyter \u304c\u52d5\u3044\u3066\u3044\u305f\u3002<\/p>\n<p><a href=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-12-02-07-33.png\"><img loading=\"lazy\" decoding=\"async\" src=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-12-02-07-33-300x222.png\" alt=\"\" width=\"300\" height=\"222\" class=\"aligncenter size-medium wp-image-9432\" srcset=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-12-02-07-33-300x222.png 300w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-12-02-07-33-768x568.png 768w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-12-02-07-33-1024x758.png 1024w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-12-02-07-33-600x444.png 600w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-12-02-07-33.png 1280w\" sizes=\"auto, (max-width: 300px) 100vw, 300px\" \/><\/a><\/p>\n<p>tensorflow \u306e\u30d0\u30fc\u30b8\u30e7\u30f3\u3092\u78ba\u8a8d\u3059\u308b\u3002bash \u3067\u6253\u3064\u306a\u3089\u4e0b\u8a18\u3002tensorflow2 \u7528\u306e\u3082\u306e\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">python3 -c 'import tensorflow as tf; print(tf.__version__)'<\/code><\/pre>\n<p>jyupyter \u4e0a\u3067\u5b9f\u884c\u3057\u3066\u307f\u305f\u3002<\/p>\n<p><a href=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-12-02-10-34.png\"><img loading=\"lazy\" decoding=\"async\" src=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-12-02-10-34-300x222.png\" alt=\"\" width=\"300\" height=\"222\" class=\"aligncenter size-medium wp-image-9433\" srcset=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-12-02-10-34-300x222.png 300w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-12-02-10-34-768x568.png 768w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-12-02-10-34-1024x758.png 1024w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-12-02-10-34-600x444.png 600w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-12-02-10-34.png 1280w\" sizes=\"auto, (max-width: 300px) 100vw, 300px\" \/><\/a><\/p>\n<p><a href=\"https:\/\/codelabo.com\/posts\/20200229081221\">Ubuntu 18.04 \u3067CUDA, Cudnn, Tensorflow GPU \u306e\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb<\/a>\u306b\u3042\u308b\u306e\u3092\u53c2\u8003\u306b\u3057\u3066 GPU \u3092\u8a8d\u8b58\u3067\u304d\u3066\u3044\u308b\u304b\u3069\u3046\u304b\u8abf\u3079\u3066\u307f\u305f\u3002<\/p>\n<p>\u30b3\u30de\u30f3\u30c9\u306f\u4e0b\u8a18\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">from tensorflow.python.client import device_lib\r\ndevice_lib.list_local_devices()<\/code><\/pre>\n<p><a href=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-12-02-46-01.png\"><img loading=\"lazy\" decoding=\"async\" src=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-12-02-46-01-300x222.png\" alt=\"\" width=\"300\" height=\"222\" class=\"aligncenter size-medium wp-image-9440\" srcset=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-12-02-46-01-300x222.png 300w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-12-02-46-01-768x568.png 768w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-12-02-46-01-1024x758.png 1024w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-12-02-46-01-600x444.png 600w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-12-02-46-01.png 1280w\" sizes=\"auto, (max-width: 300px) 100vw, 300px\" \/><\/a><\/p>\n<p>GPU \u306e\u6587\u5b57\u304c\u898b\u5f53\u305f\u3089\u306a\u3044\u3002\u30b3\u30f3\u30c6\u30ca\u304c\u3044\u3051\u306a\u304b\u3063\u305f\u304b\uff1f<\/p>\n<p>\u4ed6\u306e\u30a4\u30e1\u30fc\u30b8\u3092 pull \u3057\u3066\u307f\u308b\uff08<a href=\"https:\/\/www.tensorflow.org\/install\/docker\">\u53c2\u8003\u30b5\u30a4\u30c8<\/a>\uff09\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">docker pull tensorflow\/tensorflow:latest-gpu-jupyter<\/code><\/pre>\n<p>\u30b3\u30f3\u30c6\u30ca\u306e\u8d77\u52d5\u306f\u4e0b\u8a18\u306e\u69d8\u306b\u3057\u305f\u3002<code>--gpus all<\/code> \u3092\u4ed8\u3051\u3066\u307f\u305f\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">$ docker run -it -p 8888:8888 --gpus all tensorflow\/tensorflow:latest-gpu-jupyter<\/code><\/pre>\n<p>\u5168\u90e8\u306e\u51fa\u529b\u306f\u9577\u3059\u304e\u3066\u5165\u3063\u3066\u3044\u306a\u3044\u304c\u3001GPU \u306e\u6587\u5b57\u304c\u73fe\u308c\u305f\u3002<\/p>\n<p><a href=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-13-02-05-43.png\"><img loading=\"lazy\" decoding=\"async\" src=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-13-02-05-43-300x246.png\" alt=\"\" width=\"300\" height=\"246\" class=\"aligncenter size-medium wp-image-9451\" srcset=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-13-02-05-43-300x246.png 300w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-13-02-05-43-768x630.png 768w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-13-02-05-43-1024x840.png 1024w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-13-02-05-43-600x492.png 600w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-13-02-05-43.png 1264w\" sizes=\"auto, (max-width: 300px) 100vw, 300px\" \/><\/a><\/p>\n<p>\u51fa\u529b\u3092\u30c6\u30ad\u30b9\u30c8\u3067\u62fe\u3063\u3066\u307f\u307e\u3057\u305f\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">[name: \"\/device:CPU:0\"\r\n device_type: \"CPU\"\r\n memory_limit: 268435456\r\n locality {\r\n }\r\n incarnation: 13088651263261040163,\r\n name: \"\/device:XLA_CPU:0\"\r\n device_type: \"XLA_CPU\"\r\n memory_limit: 17179869184\r\n locality {\r\n }\r\n incarnation: 642595996897465730\r\n physical_device_desc: \"device: XLA_CPU device\",\r\n name: \"\/device:XLA_GPU:0\"\r\n device_type: \"XLA_GPU\"\r\n memory_limit: 17179869184\r\n locality {\r\n }\r\n incarnation: 14919872198234234354\r\n physical_device_desc: \"device: XLA_GPU device\",\r\n name: \"\/device:XLA_GPU:1\"\r\n device_type: \"XLA_GPU\"\r\n memory_limit: 17179869184\r\n locality {\r\n }\r\n incarnation: 15134617895178856787\r\n physical_device_desc: \"device: XLA_GPU device\",\r\n name: \"\/device:GPU:0\"\r\n device_type: \"GPU\"\r\n memory_limit: 7398066752\r\n locality {\r\n   bus_id: 1\r\n   links {\r\n     link {\r\n       device_id: 1\r\n       type: \"StreamExecutor\"\r\n       strength: 1\r\n     }\r\n   }\r\n }\r\n incarnation: 9709212083290167696\r\n physical_device_desc: \"device: 0, name: GeForce RTX 2070 SUPER, pci bus id: 0000:01:00.0, compute capability: 7.5\",\r\n name: \"\/device:GPU:1\"\r\n device_type: \"GPU\"\r\n memory_limit: 7614296224\r\n locality {\r\n   bus_id: 1\r\n   links {\r\n     link {\r\n       type: \"StreamExecutor\"\r\n       strength: 1\r\n     }\r\n   }\r\n }\r\n incarnation: 13142821548236644970\r\n physical_device_desc: \"device: 1, name: GeForce RTX 2070 SUPER, pci bus id: 0000:02:00.0, compute capability: 7.5\"]<\/code><\/pre>\n<p>\u3042\u3068\u306f\u3001\u53e4\u3044 tensorflow \u304c\u52d5\u304f\u3088\u3046\u306b\u3057\u305f\u3044\u3002<\/p>\n<p>(20200913)<br \/>\n\u4ed6\u306e\u53e4\u3044\u30a4\u30e1\u30fc\u30b8\u3092<a href=\"https:\/\/hub.docker.com\/r\/tensorflow\/tensorflow\/tags\">\u3053\u306e\u30b5\u30a4\u30c8<\/a>\u304b\u3089\u9078\u3093\u3067 pull \u3057\u3066\u307f\u308b\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">$ docker pull tensorflow\/tensorflow:1.14.0-gpu-py3-jupyter<\/code><\/pre>\n<p>\u30b3\u30f3\u30c6\u30ca\u3092\u8d77\u52d5\u3059\u308b\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">$ docker run -it -p 8888:8888 --gpus all tensorflow\/tensorflow:1.14.0-gpu-py3-jupyter<\/code><\/pre>\n<p>jyupyter \u3092\u958b\u3044\u3066\u307f\u308b\u3002tensorflow \u306e\u30d0\u30fc\u30b8\u30e7\u30f3\u3068 GPU \u3092\u8a8d\u8b58\u3057\u3066\u3044\u308b\u304b\u3069\u3046\u304b\u3092\u8a66\u3059\u3002<\/p>\n<p><a href=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-14-00-13-54.png\"><img loading=\"lazy\" decoding=\"async\" src=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-14-00-13-54-300x246.png\" alt=\"\" width=\"300\" height=\"246\" class=\"aligncenter size-medium wp-image-9458\" srcset=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-14-00-13-54-300x246.png 300w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-14-00-13-54-768x630.png 768w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-14-00-13-54-1024x840.png 1024w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-14-00-13-54-600x492.png 600w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-14-00-13-54.png 1264w\" sizes=\"auto, (max-width: 300px) 100vw, 300px\" \/><\/a><\/p>\n<p>\u5927\u4e08\u592b\u306a\u3088\u3046\u3060\u3002<a href=\"https:\/\/www.oreilly.co.jp\/books\/9784873118345\/\">\u300escikit-learn\u3068tensorflow\u306b\u3088\u308b\u5b9f\u8df5\u6a5f\u68b0\u5b66\u7fd2\u300f<\/a>\u306e\u30b3\u30fc\u30c9\u3092\u8a66\u3057\u3066\u307f\u305f\u3002<\/p>\n<p><a href=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-14-00-22-16.png\"><img loading=\"lazy\" decoding=\"async\" src=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-14-00-22-16-300x246.png\" alt=\"\" width=\"300\" height=\"246\" class=\"aligncenter size-medium wp-image-9460\" srcset=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-14-00-22-16-300x246.png 300w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-14-00-22-16-768x630.png 768w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-14-00-22-16-1024x840.png 1024w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-14-00-22-16-600x492.png 600w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-14-00-22-16.png 1264w\" sizes=\"auto, (max-width: 300px) 100vw, 300px\" \/><\/a><\/p>\n<p>Session \u304c\u5229\u7528\u3067\u304d\u308b\u3002\u3053\u308c\u3067\u3001\u3053\u306e\u672c\u304c\u8aad\u3081\u308b\u306e\u3067\u306f\u306a\u3044\u304b\u306a\uff1f<\/p>\n<p>(20200914)<\/p>\n<p><a href=\"https:\/\/www.pugetsystems.com\/labs\/hpc\/2-x-RTX2070-Super-with-NVLINK-TensorFlow-Performance-Comparison-1551\/\">\u3053\u3053<\/a>\u306b\u3042\u308b\u3088\u3046\u306a\u30d9\u30f3\u30c1\u30de\u30fc\u30af\u3092\u52d5\u304b\u3057\u3066\u307f\u305f\u3044\u3002\u307e\u305a\u306f\u30a4\u30e1\u30fc\u30b8\u306e\u53d6\u5f97\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">$ docker run --runtime=nvidia --rm -it -v $HOME\/projects:\/projects nvcr.io\/nvidia\/tensorflow:19.02-py3<\/code><\/pre>\n<p>\u4e0a\u8a18\u3067\u306f\u81ea\u5206\u306e\u30db\u30fc\u30e0\u30c7\u30a3\u30ec\u30af\u30c8\u30ea\u30fc\u306b projects \u3068\u3044\u3046\u30d5\u30a9\u30eb\u30c0\u30fc\u3092\u4f5c\u3063\u3066\u304a\u3051\u3070\u3001\u8d77\u52d5\u3057\u305f\u30b3\u30f3\u30c6\u30ca\u306e \/projects \u30d5\u30a9\u30eb\u30c0\u30fc\u3068\u3064\u306a\u304c\u308b\u3088\u3046\u306b\u306a\u3063\u3066\u3044\u308b\u3002bash \u304c\u8d77\u52d5\u3055\u308c\u3066\u3001\u3069\u306e python \u30b3\u30fc\u30c9\u3092\u5b9f\u884c\u3059\u308b\u306e\u304b\u3088\u304f\u5206\u304b\u3089\u306a\u304b\u3063\u305f\u304c\u3001<a href=\"https:\/\/qiita.com\/ksasaki\/items\/618b2f6c0cacbc7f9584\">\u3053\u3053<\/a>\u306b\u3042\u308b\u5185\u5bb9\u3092\u3084\u3063\u3066\u307f\u305f\u3002\u30b3\u30de\u30f3\u30c9\u306f\u4e0b\u8a18\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">mpiexec --allow-run-as-root --bind-to socket -np 2 python \/opt\/tensorflow\/nvidia-examples\/cnn\/resnet.py --layers=50 --precision=fp16 --batch_size=128<\/code><\/pre>\n<p>\u304b\u306a\u308a\u306e\u91cf\u306e\u5fdc\u7b54\u30e1\u30c3\u30bb\u30fc\u30b8\u304c\u3042\u308b\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\"># mpiexec --allow-run-as-root --bind-to socket -np 2 python \/opt\/tensorflow\/nvidia-examples\/cnn\/resnet.py --layers=50 --precision=fp16 --batch_size=128\r\n--------------------------------------------------------------------------\r\nWARNING: Open MPI tried to bind a process but failed.  This is a\r\nwarning only; your job will continue, though performance may\r\nbe degraded.\r\n\r\n  Local host:        2bbf45ca6ffc\r\n  Application name:  \/usr\/bin\/python\r\n  Error message:     failed to bind memory\r\n  Location:          rtc_hwloc.c:445\r\n\r\n--------------------------------------------------------------------------\r\nPY 3.5.2 (default, Nov 12 2018, 13:43:14) \r\n[GCC 5.4.0 20160609]\r\nTF 1.13.0-rc0\r\nPY 3.5.2 (default, Nov 12 2018, 13:43:14) \r\n[GCC 5.4.0 20160609]\r\nTF 1.13.0-rc0\r\nScript arguments:\r\n  --predict False\r\n  --batch_size 128\r\n  --display_every 10\r\n  --iter_unit epoch\r\n  --num_iter 90\r\n  --layers 50\r\n  --precision fp16\r\nWARNING:tensorflow:Using temporary folder as model directory: \/tmp\/tmpr7czshlr\r\nTraining\r\nWARNING:tensorflow:Using temporary folder as model directory: \/tmp\/tmpkt9_9irw\r\nTraining\r\nWARNING:tensorflow:From \/usr\/local\/lib\/python3.5\/dist-packages\/tensorflow\/python\/framework\/op_def_library.py:263: colocate_with (from tensorflow.python.framework.ops) is deprecated and will be removed in a future version.\r\nInstructions for updating:\r\nColocations handled automatically by placer.\r\nWARNING:tensorflow:From \/usr\/local\/lib\/python3.5\/dist-packages\/tensorflow\/python\/framework\/op_def_library.py:263: colocate_with (from tensorflow.python.framework.ops) is deprecated and will be removed in a future version.\r\nInstructions for updating:\r\nColocations handled automatically by placer.\r\nWARNING:tensorflow:From \/usr\/local\/lib\/python3.5\/dist-packages\/tensorflow_estimator\/python\/estimator\/util.py:104: DatasetV1.make_initializable_iterator (from tensorflow.python.data.ops.dataset_ops) is deprecated and will be removed in a future version.\r\nInstructions for updating:\r\nUse `for ... in dataset:` to iterate over a dataset. If using `tf.estimator`, return the `Dataset` object directly from your input function. As a last resort, you can use `tf.compat.v1.data.make_initializable_iterator(dataset)`.\r\nWARNING:tensorflow:From \/usr\/local\/lib\/python3.5\/dist-packages\/tensorflow_estimator\/python\/estimator\/util.py:104: DatasetV1.make_initializable_iterator (from tensorflow.python.data.ops.dataset_ops) is deprecated and will be removed in a future version.\r\nInstructions for updating:\r\nUse `for ... in dataset:` to iterate over a dataset. If using `tf.estimator`, return the `Dataset` object directly from your input function. As a last resort, you can use `tf.compat.v1.data.make_initializable_iterator(dataset)`.\r\nWARNING:tensorflow:From \/opt\/tensorflow\/nvidia-examples\/cnn\/nvutils\/builder.py:25: conv2d (from tensorflow.python.layers.convolutional) is deprecated and will be removed in a future version.\r\nInstructions for updating:\r\nUse keras.layers.conv2d instead.\r\nWARNING:tensorflow:From \/opt\/tensorflow\/nvidia-examples\/cnn\/nvutils\/builder.py:25: conv2d (from tensorflow.python.layers.convolutional) is deprecated and will be removed in a future version.\r\nInstructions for updating:\r\nUse keras.layers.conv2d instead.\r\nWARNING:tensorflow:From \/opt\/tensorflow\/nvidia-examples\/cnn\/nvutils\/builder.py:58: max_pooling2d (from tensorflow.python.layers.pooling) is deprecated and will be removed in a future version.\r\nInstructions for updating:\r\nUse keras.layers.max_pooling2d instead.\r\nWARNING:tensorflow:From \/opt\/tensorflow\/nvidia-examples\/cnn\/nvutils\/builder.py:58: max_pooling2d (from tensorflow.python.layers.pooling) is deprecated and will be removed in a future version.\r\nInstructions for updating:\r\nUse keras.layers.max_pooling2d instead.\r\nWARNING:tensorflow:From \/opt\/tensorflow\/nvidia-examples\/cnn\/nvutils\/builder.py:90: average_pooling2d (from tensorflow.python.layers.pooling) is deprecated and will be removed in a future version.\r\nInstructions for updating:\r\nUse keras.layers.average_pooling2d instead.\r\nWARNING:tensorflow:From \/opt\/tensorflow\/nvidia-examples\/cnn\/nvutils\/runner.py:116: dense (from tensorflow.python.layers.core) is deprecated and will be removed in a future version.\r\nInstructions for updating:\r\nUse keras.layers.dense instead.\r\nWARNING:tensorflow:From \/usr\/local\/lib\/python3.5\/dist-packages\/tensorflow\/python\/ops\/losses\/losses_impl.py:209: to_float (from tensorflow.python.ops.math_ops) is deprecated and will be removed in a future version.\r\nInstructions for updating:\r\nUse tf.cast instead.\r\nWARNING:tensorflow:From \/opt\/tensorflow\/nvidia-examples\/cnn\/nvutils\/builder.py:90: average_pooling2d (from tensorflow.python.layers.pooling) is deprecated and will be removed in a future version.\r\nInstructions for updating:\r\nUse keras.layers.average_pooling2d instead.\r\nWARNING:tensorflow:From \/opt\/tensorflow\/nvidia-examples\/cnn\/nvutils\/runner.py:116: dense (from tensorflow.python.layers.core) is deprecated and will be removed in a future version.\r\nInstructions for updating:\r\nUse keras.layers.dense instead.\r\nWARNING:tensorflow:From \/usr\/local\/lib\/python3.5\/dist-packages\/tensorflow\/python\/ops\/losses\/losses_impl.py:209: to_float (from tensorflow.python.ops.math_ops) is deprecated and will be removed in a future version.\r\nInstructions for updating:\r\nUse tf.cast instead.\r\nWARNING:tensorflow:From \/usr\/local\/lib\/python3.5\/dist-packages\/tensorflow\/python\/training\/learning_rate_decay_v2.py:321: div (from tensorflow.python.ops.math_ops) is deprecated and will be removed in a future version.\r\nInstructions for updating:\r\nDeprecated in favor of operator or tf.math.divide.\r\nWARNING:tensorflow:From \/usr\/local\/lib\/python3.5\/dist-packages\/tensorflow\/python\/training\/learning_rate_decay_v2.py:321: div (from tensorflow.python.ops.math_ops) is deprecated and will be removed in a future version.\r\nInstructions for updating:\r\nDeprecated in favor of operator or tf.math.divide.\r\n[2bbf45ca6ffc:00059] 1 more process has sent help message help-orte-odls-default.txt \/ memory not bound\r\n[2bbf45ca6ffc:00059] Set MCA parameter \"orte_base_help_aggregate\" to 0 to see all help \/ error messages\r\n2020-09-14 02:02:10.195593: I tensorflow\/core\/platform\/profile_utils\/cpu_utils.cc:94] CPU Frequency: 3000000000 Hz\r\n2020-09-14 02:02:10.195972: I tensorflow\/compiler\/xla\/service\/service.cc:161] XLA service 0xaeb5b70 executing computations on platform Host. Devices:\r\n2020-09-14 02:02:10.195991: I tensorflow\/compiler\/xla\/service\/service.cc:168]   StreamExecutor device (0): <undefined>, <undefined>\r\n2020-09-14 02:02:10.277787: I tensorflow\/core\/platform\/profile_utils\/cpu_utils.cc:94] CPU Frequency: 3000000000 Hz\r\n2020-09-14 02:02:10.278051: I tensorflow\/compiler\/xla\/service\/service.cc:161] XLA service 0xa411b20 executing computations on platform Host. Devices:\r\n2020-09-14 02:02:10.278067: I tensorflow\/compiler\/xla\/service\/service.cc:168]   StreamExecutor device (0): <undefined>, <undefined>\r\n2020-09-14 02:02:10.281113: I tensorflow\/stream_executor\/cuda\/cuda_gpu_executor.cc:998] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\r\n2020-09-14 02:02:10.282426: I tensorflow\/compiler\/xla\/service\/service.cc:161] XLA service 0xac01180 executing computations on platform CUDA. Devices:\r\n2020-09-14 02:02:10.282441: I tensorflow\/compiler\/xla\/service\/service.cc:168]   StreamExecutor device (0): GeForce RTX 2070 SUPER, Compute Capability 7.5\r\n2020-09-14 02:02:10.282596: I tensorflow\/core\/common_runtime\/gpu\/gpu_device.cc:1433] Found device 0 with properties: \r\nname: GeForce RTX 2070 SUPER major: 7 minor: 5 memoryClockRate(GHz): 1.815\r\npciBusID: 0000:01:00.0\r\ntotalMemory: 7.79GiB freeMemory: 7.42GiB\r\n2020-09-14 02:02:10.282623: I tensorflow\/core\/common_runtime\/gpu\/gpu_device.cc:1512] Adding visible gpu devices: 0\r\n2020-09-14 02:02:10.366559: I tensorflow\/stream_executor\/cuda\/cuda_gpu_executor.cc:998] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\r\n2020-09-14 02:02:10.367156: I tensorflow\/compiler\/xla\/service\/service.cc:161] XLA service 0x967ba80 executing computations on platform CUDA. Devices:\r\n2020-09-14 02:02:10.367176: I tensorflow\/compiler\/xla\/service\/service.cc:168]   StreamExecutor device (0): GeForce RTX 2070 SUPER, Compute Capability 7.5\r\n2020-09-14 02:02:10.367271: I tensorflow\/core\/common_runtime\/gpu\/gpu_device.cc:1433] Found device 0 with properties: \r\nname: GeForce RTX 2070 SUPER major: 7 minor: 5 memoryClockRate(GHz): 1.815\r\npciBusID: 0000:02:00.0\r\ntotalMemory: 7.79GiB freeMemory: 7.69GiB\r\n2020-09-14 02:02:10.367283: I tensorflow\/core\/common_runtime\/gpu\/gpu_device.cc:1512] Adding visible gpu devices: 1\r\n2020-09-14 02:02:10.559057: I tensorflow\/core\/common_runtime\/gpu\/gpu_device.cc:984] Device interconnect StreamExecutor with strength 1 edge matrix:\r\n2020-09-14 02:02:10.559086: I tensorflow\/core\/common_runtime\/gpu\/gpu_device.cc:990]      0 \r\n2020-09-14 02:02:10.559092: I tensorflow\/core\/common_runtime\/gpu\/gpu_device.cc:1003] 0:   N \r\n2020-09-14 02:02:10.559208: I tensorflow\/core\/common_runtime\/gpu\/gpu_device.cc:1115] Created TensorFlow device (\/job:localhost\/replica:0\/task:0\/device:GPU:0 with 5582 MB memory) -> physical GPU (device: 0, name: GeForce RTX 2070 SUPER, pci bus id: 0000:01:00.0, compute capability: 7.5)\r\n2020-09-14 02:02:10.603602: I tensorflow\/core\/common_runtime\/gpu\/gpu_device.cc:984] Device interconnect StreamExecutor with strength 1 edge matrix:\r\n2020-09-14 02:02:10.603630: I tensorflow\/core\/common_runtime\/gpu\/gpu_device.cc:990]      1 \r\n2020-09-14 02:02:10.603636: I tensorflow\/core\/common_runtime\/gpu\/gpu_device.cc:1003] 1:   N \r\n2020-09-14 02:02:10.603764: I tensorflow\/core\/common_runtime\/gpu\/gpu_device.cc:1115] Created TensorFlow device (\/job:localhost\/replica:0\/task:0\/device:GPU:0 with 5587 MB memory) -> physical GPU (device: 1, name: GeForce RTX 2070 SUPER, pci bus id: 0000:02:00.0, compute capability: 7.5)\r\n  Step Epoch Img\/sec   Loss  LR\r\n2020-09-14 02:02:19.418561: I tensorflow\/stream_executor\/dso_loader.cc:152] successfully opened CUDA library libcublas.so.10.0 locally\r\n2020-09-14 02:02:19.491472: I tensorflow\/stream_executor\/dso_loader.cc:152] successfully opened CUDA library libcublas.so.10.0 locally\r\n2020-09-14 02:02:21.279530: W tensorflow\/core\/common_runtime\/bfc_allocator.cc:211] Allocator (GPU_0_bfc) ran out of memory trying to allocate 2.92GiB. The caller indicates that this is not a failure, but may mean that there could be performance gains if more memory were available.\r\n2020-09-14 02:02:21.279570: W tensorflow\/core\/common_runtime\/bfc_allocator.cc:211] Allocator (GPU_0_bfc) ran out of memory trying to allocate 2.92GiB. The caller indicates that this is not a failure, but may mean that there could be performance gains if more memory were available.\r\n2020-09-14 02:02:21.414634: W tensorflow\/core\/common_runtime\/bfc_allocator.cc:211] Allocator (GPU_0_bfc) ran out of memory trying to allocate 2.92GiB. The caller indicates that this is not a failure, but may mean that there could be performance gains if more memory were available.\r\n2020-09-14 02:02:21.414693: W tensorflow\/core\/common_runtime\/bfc_allocator.cc:211] Allocator (GPU_0_bfc) ran out of memory trying to allocate 2.92GiB. The caller indicates that this is not a failure, but may mean that there could be performance gains if more memory were available.\r\n2020-09-14 02:02:21.617778: W tensorflow\/core\/common_runtime\/bfc_allocator.cc:211] Allocator (GPU_0_bfc) ran out of memory trying to allocate 865.00MiB. The caller indicates that this is not a failure, but may mean that there could be performance gains if more memory were available.\r\n2020-09-14 02:02:21.617829: W tensorflow\/core\/common_runtime\/bfc_allocator.cc:211] Allocator (GPU_0_bfc) ran out of memory trying to allocate 865.00MiB. The caller indicates that this is not a failure, but may mean that there could be performance gains if more memory were available.\r\n2020-09-14 02:02:21.685731: W tensorflow\/core\/common_runtime\/bfc_allocator.cc:211] Allocator (GPU_0_bfc) ran out of memory trying to allocate 865.00MiB. The caller indicates that this is not a failure, but may mean that there could be performance gains if more memory were available.\r\n2020-09-14 02:02:21.685779: W tensorflow\/core\/common_runtime\/bfc_allocator.cc:211] Allocator (GPU_0_bfc) ran out of memory trying to allocate 865.00MiB. The caller indicates that this is not a failure, but may mean that there could be performance gains if more memory were available.\r\n2020-09-14 02:02:21.714468: W tensorflow\/core\/common_runtime\/bfc_allocator.cc:211] Allocator (GPU_0_bfc) ran out of memory trying to allocate 649.00MiB. The caller indicates that this is not a failure, but may mean that there could be performance gains if more memory were available.\r\n2020-09-14 02:02:21.714498: W tensorflow\/core\/common_runtime\/bfc_allocator.cc:211] Allocator (GPU_0_bfc) ran out of memory trying to allocate 649.00MiB. The caller indicates that this is not a failure, but may mean that there could be performance gains if more memory were available.\r\n2020-09-14 02:02:21.722806: W tensorflow\/core\/common_runtime\/bfc_allocator.cc:211] Allocator (GPU_0_bfc) ran out of memory trying to allocate 865.00MiB. The caller indicates that this is not a failure, but may mean that there could be performance gains if more memory were available.\r\n2020-09-14 02:02:21.722857: W tensorflow\/core\/common_runtime\/bfc_allocator.cc:211] Allocator (GPU_0_bfc) ran out of memory trying to allocate 865.00MiB. The caller indicates that this is not a failure, but may mean that there could be performance gains if more memory were available.\r\n2020-09-14 02:02:21.743530: W tensorflow\/core\/common_runtime\/bfc_allocator.cc:211] Allocator (GPU_0_bfc) ran out of memory trying to allocate 649.00MiB. The caller indicates that this is not a failure, but may mean that there could be performance gains if more memory were available.\r\n2020-09-14 02:02:21.743554: W tensorflow\/core\/common_runtime\/bfc_allocator.cc:211] Allocator (GPU_0_bfc) ran out of memory trying to allocate 649.00MiB. The caller indicates that this is not a failure, but may mean that there could be performance gains if more memory were available.\r\n2020-09-14 02:02:21.753277: W tensorflow\/core\/common_runtime\/bfc_allocator.cc:211] Allocator (GPU_0_bfc) ran out of memory trying to allocate 865.00MiB. The caller indicates that this is not a failure, but may mean that there could be performance gains if more memory were available.\r\n2020-09-14 02:02:21.753299: W tensorflow\/core\/common_runtime\/bfc_allocator.cc:211] Allocator (GPU_0_bfc) ran out of memory trying to allocate 865.00MiB. The caller indicates that this is not a failure, but may mean that there could be performance gains if more memory were available.\r\n2020-09-14 02:02:21.802270: W tensorflow\/core\/common_runtime\/bfc_allocator.cc:211] Allocator (GPU_0_bfc) ran out of memory trying to allocate 2.92GiB. The caller indicates that this is not a failure, but may mean that there could be performance gains if more memory were available.\r\n2020-09-14 02:02:21.802343: W tensorflow\/core\/common_runtime\/bfc_allocator.cc:211] Allocator (GPU_0_bfc) ran out of memory trying to allocate 2.92GiB. The caller indicates that this is not a failure, but may mean that there could be performance gains if more memory were available.\r\n2020-09-14 02:02:21.822577: W tensorflow\/core\/common_runtime\/bfc_allocator.cc:211] Allocator (GPU_0_bfc) ran out of memory trying to allocate 729.00MiB. The caller indicates that this is not a failure, but may mean that there could be performance gains if more memory were available.\r\n2020-09-14 02:02:21.822607: W tensorflow\/core\/common_runtime\/bfc_allocator.cc:211] Allocator (GPU_0_bfc) ran out of memory trying to allocate 729.00MiB. The caller indicates that this is not a failure, but may mean that there could be performance gains if more memory were available.\r\n     1   1.0    32.3  7.525  8.496 2.00000\r\n    10  10.0   228.8  4.313  5.286 1.62000\r\n    20  20.0   647.2  0.048  1.025 1.24469\r\n    30  30.0   683.5  0.003  0.982 0.91877\r\n    40  40.0   672.6  0.023  1.002 0.64222\r\n    50  50.0   665.2  0.112  1.092 0.41506\r\n    60  60.0   676.2  0.087  1.068 0.23728\r\n    70  70.0   665.4  0.082  1.064 0.10889\r\n    80  80.0   677.2  0.001  0.983 0.02988\r\n    90  90.0   574.4  0.000  0.983 0.00025<\/code><\/pre>\n<p>\u4e0a\u624b\u304f\u3044\u3063\u305f\u306e\u3060\u308d\u3046\u304b\uff1f\u6bce\u79d2670\u30a4\u30e1\u30fc\u30b8\u3068\u3044\u3046\u306e\u306f\u3001\u5148\u306b\u3042\u3052\u305f\u30b5\u30a4\u30c8\u304b\u3089\u3059\u308b\u3068\u3001\u305d\u308c\u3089\u3057\u3044\u5024\u3067\u306f\u306a\u3044\u3060\u308d\u3046\u304b\u3002\u3067\u3082\u30c7\u30fc\u30bf\u306f\u3069\u3053\u306b\u3042\u308b\u3093\u3060\uff1f<\/p>\n<p>\u540c\u3058\u3088\u3046\u306a\u30b3\u30de\u30f3\u30c9\u304c<a href=\"https:\/\/www.pugetsystems.com\/labs\/hpc\/TensorFlow-Performance-with-1-4-GPUs----RTX-Titan-2080Ti-2080-2070-GTX-1660Ti-1070-1080Ti-and-Titan-V-1386\/\">\u3053\u3053<\/a>\u306b\u3082\u3042\u3063\u305f\u3002\u3057\u304b\u3057 LSTM \u306e\u65b9\u306f\u4e0a\u624b\u304f\u52d5\u304b\u305b\u306a\u3044\u3002<\/p>\n<p><a href=\"https:\/\/ngc.nvidia.com\/\">NVIDIA GPU Cloud\uff08NGC\uff09<\/a>\u306b\u767b\u9332\u3057\u305f\u3002\u5fc5\u8981\u3060\u3063\u305f\u306e\u304b\u306a\uff1f<\/p>\n<p>(20200915)<\/p>\n<p><a href=\"https:\/\/www.pugetsystems.com\/labs\/hpc\/2-x-RTX2070-Super-with-NVLINK-TensorFlow-Performance-Comparison-1551\/\">2 x RTX2070 Super with NVLINK TensorFlow Performance Comparison<\/a> \u3067\u306f Big-LSTM \u306e\u30d9\u30f3\u30c1\u30de\u30fc\u30af\u3082\u3042\u308b\u3002\u3053\u308c\u3092\u52d5\u304b\u3057\u3066\u307f\u308b\u3002docker \u30a4\u30e1\u30fc\u30b8\u306f 19.02-py3 \u3067\u3042\u308b\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">docker run --runtime=nvidia --rm -it -v $HOME\/projects:\/projects nvcr.io\/nvidia\/tensorflow:19.02-py3<\/code><\/pre>\n<p><code>\/opt\/tensorflow\/nvidia-examples\/big_lstm<\/code> \u306b\u30b3\u30fc\u30c9\u304c\u3042\u308b\u3002\u6700\u521d\u306b\u30c7\u30fc\u30bf\u306e\u7528\u610f\u3067\u3042\u308b\u3002<code>download_1b_words_data.sh<\/code> \u3092\u5b9f\u884c\u3059\u308b\u3002\u305d\u3046\u3059\u308b\u3068\u3001\u540c\u3058\u30d5\u30a9\u30eb\u30c0\u30fc\u5185\u306b<code>1-billion-word-language-modeling-benchmark-r13output<\/code> \u3068\u3044\u3046\u30d5\u30a9\u30eb\u30c0\u30fc\u304c\u4f5c\u6210\u3055\u308c\u308b\u3002\u3053\u306e\u4e2d\u306b\u30c7\u30fc\u30bf\u304c\u3042\u308b\u306e\u3060\u304c\u3001\u3053\u308c\u3092 <code>projects<\/code> \u30d5\u30a9\u30eb\u30c0\u30fc\u306b\u79fb\u52d5\u3055\u305b\u305f\u3002\u3053\u306e\u4ed6\u306b\u3001log \u7528\u306e\u30d5\u30a9\u30eb\u30c0\u30fc\u3092\u3001\u3053\u308c\u3082 <code>projects<\/code> \u306b\u7528\u610f\u3057\u305f\u3002\u3053\u308c\u306b\u5408\u308f\u305b\u3066\u3001\u30b3\u30de\u30f3\u30c9\u3092\u4e0b\u8a18\u306e\u3088\u3046\u306b\u4fee\u6b63\u3057\u305f\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">python single_lm_train.py --mode=train --logdir=\/projects\/logs --num_gpus=2 --datadir=\/projects\/1-billion-word-language-modeling-benchmark-r13output --hpconfig  run_profiler=False,max_time=240,num_steps=20,num_shards=8,num_layers=2, learning_rate=0.2,max_grad_norm=1,keep_prob=0.9,emb_size=1024,projected_size=1024,  state_size=8192,num_sampled=8192,batch_size=448<\/code><\/pre>\n<p>\u5b9f\u884c\u7d50\u679c\u3092\u4e0b\u8a18\u306b\u8a18\u3059\u3002<\/p>\n<pre><code style=\"color:midnightblue; font-size:12px;white-space:pre-wrap\">root@c35a8402c135:\/opt\/tensorflow\/nvidia-examples\/big_lstm# python single_lm_train.py --mode=train --logdir=\/projects\/logs --num_gpus=2 --datadir=\/projects\/1-billion-word-language-modeling-benchmark-r13output --hpconfig  run_profiler=False,max_time=240,num_steps=20,num_shards=8,num_layers=2, learning_rate=0.2,max_grad_norm=1,keep_prob=0.9,emb_size=1024,projected_size=1024,  state_size=8192,num_sampled=8192,batch_size=448\r\n\r\nWARNING: The TensorFlow contrib module will not be included in TensorFlow 2.0.\r\nFor more information, please see:\r\n  * https:\/\/github.com\/tensorflow\/community\/blob\/master\/rfcs\/20180907-contrib-sunset.md\r\n  * https:\/\/github.com\/tensorflow\/addons\r\nIf you depend on functionality not listed there, please file an issue.\r\n\r\n*****HYPER PARAMETERS*****\r\n{'emb_size': 512, 'num_delayed_steps': 150, 'max_grad_norm': 10.0, 'learning_rate': 0.2, 'vocab_size': 793470, 'batch_size': 128, 'num_gpus': 2, 'keep_prob': 0.9, 'average_params': True, 'num_shards': 8, 'num_steps': 20, 'state_size': 2048, 'max_time': 240, 'run_profiler': False, 'num_layers': 2, 'optimizer': 0, 'do_summaries': False, 'projected_size': 512, 'num_sampled': 8192}\r\n**************************\r\nWARNING:tensorflow:From \/usr\/local\/lib\/python3.5\/dist-packages\/tensorflow\/python\/framework\/op_def_library.py:263: colocate_with (from tensorflow.python.framework.ops) is deprecated and will be removed in a future version.\r\nInstructions for updating:\r\nColocations handled automatically by placer.\r\nWARNING:tensorflow:From \/opt\/tensorflow\/nvidia-examples\/big_lstm\/model_utils.py:33: UniformUnitScaling.__init__ (from tensorflow.python.ops.init_ops) is deprecated and will be removed in a future version.\r\nInstructions for updating:\r\nUse tf.initializers.variance_scaling instead with distribution=uniform to get equivalent behavior.\r\nWARNING:tensorflow:From \/opt\/tensorflow\/nvidia-examples\/big_lstm\/language_model.py:75: calling dropout (from tensorflow.python.ops.nn_ops) with keep_prob is deprecated and will be removed in a future version.\r\nInstructions for updating:\r\nPlease use `rate` instead of `keep_prob`. Rate should be set to `rate = 1 - keep_prob`.\r\nWARNING:tensorflow:From \/opt\/tensorflow\/nvidia-examples\/big_lstm\/language_model.py:107: to_float (from tensorflow.python.ops.math_ops) is deprecated and will be removed in a future version.\r\nInstructions for updating:\r\nUse tf.cast instead.\r\nWARNING:tensorflow:From \/usr\/local\/lib\/python3.5\/dist-packages\/tensorflow\/python\/ops\/nn_impl.py:1444: sparse_to_dense (from tensorflow.python.ops.sparse_ops) is deprecated and will be removed in a future version.\r\nInstructions for updating:\r\nCreate a `tf.sparse.SparseTensor` and use `tf.sparse.to_dense` instead.\r\nWARNING:tensorflow:From \/usr\/local\/lib\/python3.5\/dist-packages\/tensorflow\/python\/ops\/array_grad.py:425: to_int32 (from tensorflow.python.ops.math_ops) is deprecated and will be removed in a future version.\r\nInstructions for updating:\r\nUse tf.cast instead.\r\nCurrent time: 1600138841.891555\r\nALL VARIABLES\r\nWARNING:tensorflow:From \/opt\/tensorflow\/nvidia-examples\/big_lstm\/run_utils.py:18: all_variables (from tensorflow.python.ops.variables) is deprecated and will be removed after 2017-03-02.\r\nInstructions for updating:\r\nPlease use tf.global_variables instead.\r\nmodel\/emb_0:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/emb_1:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/emb_2:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/emb_3:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/emb_4:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/emb_5:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/emb_6:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/emb_7:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/lstm_0\/LSTMCell\/W_0:0 (1024, 8192) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/lstm_0\/LSTMCell\/B:0 (8192,) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/lstm_0\/LSTMCell\/W_P_0:0 (2048, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/lstm_1\/LSTMCell\/W_0:0 (1024, 8192) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/lstm_1\/LSTMCell\/B:0 (8192,) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/lstm_1\/LSTMCell\/W_P_0:0 (2048, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/softmax_w_0:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/softmax_w_1:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/softmax_w_2:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/softmax_w_3:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/softmax_w_4:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/softmax_w_5:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/softmax_w_6:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/softmax_w_7:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/softmax_b:0 (793470,) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/global_step:0 () <dtype: 'int32_ref'> \r\nmodel\/model\/emb_0\/Adagrad:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/model\/emb_1\/Adagrad:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/model\/emb_2\/Adagrad:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/model\/emb_3\/Adagrad:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/model\/emb_4\/Adagrad:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/model\/emb_5\/Adagrad:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/model\/emb_6\/Adagrad:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/model\/emb_7\/Adagrad:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/model\/lstm_0\/LSTMCell\/W_0\/Adagrad:0 (1024, 8192) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/model\/lstm_0\/LSTMCell\/B\/Adagrad:0 (8192,) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/model\/lstm_0\/LSTMCell\/W_P_0\/Adagrad:0 (2048, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/model\/lstm_1\/LSTMCell\/W_0\/Adagrad:0 (1024, 8192) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/model\/lstm_1\/LSTMCell\/B\/Adagrad:0 (8192,) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/model\/lstm_1\/LSTMCell\/W_P_0\/Adagrad:0 (2048, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/model\/softmax_w_0\/Adagrad:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/model\/softmax_w_1\/Adagrad:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/model\/softmax_w_2\/Adagrad:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/model\/softmax_w_3\/Adagrad:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/model\/softmax_w_4\/Adagrad:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/model\/softmax_w_5\/Adagrad:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/model\/softmax_w_6\/Adagrad:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/model\/softmax_w_7\/Adagrad:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/model\/softmax_b\/Adagrad:0 (793470,) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/model\/lstm_0\/LSTMCell\/W_0\/ExponentialMovingAverage:0 (1024, 8192) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/model\/lstm_0\/LSTMCell\/B\/ExponentialMovingAverage:0 (8192,) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/model\/lstm_0\/LSTMCell\/W_P_0\/ExponentialMovingAverage:0 (2048, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/model\/lstm_1\/LSTMCell\/W_0\/ExponentialMovingAverage:0 (1024, 8192) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/model\/lstm_1\/LSTMCell\/B\/ExponentialMovingAverage:0 (8192,) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/model\/lstm_1\/LSTMCell\/W_P_0\/ExponentialMovingAverage:0 (2048, 512) <dtype: 'float32_ref'> \/gpu:0\r\nTRAINABLE VARIABLES\r\nmodel\/emb_0:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/emb_1:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/emb_2:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/emb_3:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/emb_4:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/emb_5:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/emb_6:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/emb_7:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/lstm_0\/LSTMCell\/W_0:0 (1024, 8192) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/lstm_0\/LSTMCell\/B:0 (8192,) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/lstm_0\/LSTMCell\/W_P_0:0 (2048, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/lstm_1\/LSTMCell\/W_0:0 (1024, 8192) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/lstm_1\/LSTMCell\/B:0 (8192,) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/lstm_1\/LSTMCell\/W_P_0:0 (2048, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/softmax_w_0:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/softmax_w_1:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/softmax_w_2:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/softmax_w_3:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/softmax_w_4:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/softmax_w_5:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/softmax_w_6:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/softmax_w_7:0 (99184, 512) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/softmax_b:0 (793470,) <dtype: 'float32_ref'> \/gpu:0\r\nLOCAL VARIABLES\r\nmodel\/model\/state_0_0:0 (128, 2560) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/model\/state_0_1:0 (128, 2560) <dtype: 'float32_ref'> \/gpu:0\r\nmodel\/model_1\/state_1_0:0 (128, 2560) <dtype: 'float32_ref'> \/gpu:1\r\nmodel\/model_1\/state_1_1:0 (128, 2560) <dtype: 'float32_ref'> \/gpu:1\r\nWARNING:tensorflow:From \/opt\/tensorflow\/nvidia-examples\/big_lstm\/run_utils.py:32: Supervisor.__init__ (from tensorflow.python.training.supervisor) is deprecated and will be removed in a future version.\r\nInstructions for updating:\r\nPlease switch to tf.train.MonitoredTrainingSession\r\n2020-09-15 03:00:42.467627: I tensorflow\/core\/platform\/profile_utils\/cpu_utils.cc:94] CPU Frequency: 3000000000 Hz\r\n2020-09-15 03:00:42.468021: I tensorflow\/compiler\/xla\/service\/service.cc:161] XLA service 0xe8051e0 executing computations on platform Host. Devices:\r\n2020-09-15 03:00:42.468039: I tensorflow\/compiler\/xla\/service\/service.cc:168]   StreamExecutor device (0): <undefined>, <undefined>\r\n2020-09-15 03:00:42.652862: I tensorflow\/stream_executor\/cuda\/cuda_gpu_executor.cc:998] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\r\n2020-09-15 03:00:42.661208: I tensorflow\/stream_executor\/cuda\/cuda_gpu_executor.cc:998] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\r\n2020-09-15 03:00:42.661766: I tensorflow\/compiler\/xla\/service\/service.cc:161] XLA service 0xbf707d0 executing computations on platform CUDA. Devices:\r\n2020-09-15 03:00:42.661782: I tensorflow\/compiler\/xla\/service\/service.cc:168]   StreamExecutor device (0): GeForce RTX 2070 SUPER, Compute Capability 7.5\r\n2020-09-15 03:00:42.661788: I tensorflow\/compiler\/xla\/service\/service.cc:168]   StreamExecutor device (1): GeForce RTX 2070 SUPER, Compute Capability 7.5\r\n2020-09-15 03:00:42.661980: I tensorflow\/core\/common_runtime\/gpu\/gpu_device.cc:1433] Found device 0 with properties: \r\nname: GeForce RTX 2070 SUPER major: 7 minor: 5 memoryClockRate(GHz): 1.815\r\npciBusID: 0000:01:00.0\r\ntotalMemory: 7.79GiB freeMemory: 7.39GiB\r\n2020-09-15 03:00:42.662025: I tensorflow\/core\/common_runtime\/gpu\/gpu_device.cc:1433] Found device 1 with properties: \r\nname: GeForce RTX 2070 SUPER major: 7 minor: 5 memoryClockRate(GHz): 1.815\r\npciBusID: 0000:02:00.0\r\ntotalMemory: 7.79GiB freeMemory: 7.69GiB\r\n2020-09-15 03:00:42.662047: I tensorflow\/core\/common_runtime\/gpu\/gpu_device.cc:1512] Adding visible gpu devices: 0, 1\r\n2020-09-15 03:00:43.136398: I tensorflow\/core\/common_runtime\/gpu\/gpu_device.cc:984] Device interconnect StreamExecutor with strength 1 edge matrix:\r\n2020-09-15 03:00:43.136431: I tensorflow\/core\/common_runtime\/gpu\/gpu_device.cc:990]      0 1 \r\n2020-09-15 03:00:43.136437: I tensorflow\/core\/common_runtime\/gpu\/gpu_device.cc:1003] 0:   N Y \r\n2020-09-15 03:00:43.136441: I tensorflow\/core\/common_runtime\/gpu\/gpu_device.cc:1003] 1:   Y N \r\n2020-09-15 03:00:43.136529: I tensorflow\/core\/common_runtime\/gpu\/gpu_device.cc:1115] Created TensorFlow device (\/job:localhost\/replica:0\/task:0\/device:GPU:0 with 7097 MB memory) -> physical GPU (device: 0, name: GeForce RTX 2070 SUPER, pci bus id: 0000:01:00.0, compute capability: 7.5)\r\n2020-09-15 03:00:43.136958: I tensorflow\/core\/common_runtime\/gpu\/gpu_device.cc:1115] Created TensorFlow device (\/job:localhost\/replica:0\/task:0\/device:GPU:1 with 7390 MB memory) -> physical GPU (device: 1, name: GeForce RTX 2070 SUPER, pci bus id: 0000:02:00.0, compute capability: 7.5)\r\nWARNING:tensorflow:From \/usr\/local\/lib\/python3.5\/dist-packages\/tensorflow\/python\/training\/saver.py:1266: checkpoint_exists (from tensorflow.python.training.checkpoint_management) is deprecated and will be removed in a future version.\r\nInstructions for updating:\r\nUse standard file APIs to check for files with this prefix.\r\nWARNING:tensorflow:From \/usr\/local\/lib\/python3.5\/dist-packages\/tensorflow\/python\/training\/saver.py:1070: get_checkpoint_mtimes (from tensorflow.python.training.checkpoint_management) is deprecated and will be removed in a future version.\r\nInstructions for updating:\r\nUse standard file utilities to get mtimes.\r\nProcessing file: \/projects\/1-billion-word-language-modeling-benchmark-r13output\/training-monolingual.tokenized.shuffled\/news.en-00084-of-00100\r\nFinished processing!\r\n2020-09-15 03:01:01.977579: I tensorflow\/stream_executor\/dso_loader.cc:152] successfully opened CUDA library libcublas.so.10.0 locally\r\nIteration 1679, time = 13.12s, wps = 390, train loss = 118.9416\r\nIteration 1680, time = 9.67s, wps = 529, train loss = 84.3850\r\nIteration 1681, time = 0.13s, wps = 39567, train loss = 107.0070\r\nIteration 1682, time = 0.12s, wps = 43874, train loss = 46.1559\r\nIteration 1683, time = 0.12s, wps = 43468, train loss = 131.2872\r\nIteration 1684, time = 0.12s, wps = 41286, train loss = 24.5033\r\nIteration 1685, time = 0.13s, wps = 39113, train loss = 10.7746\r\nIteration 1686, time = 0.13s, wps = 40528, train loss = 8.6051\r\nIteration 1687, time = 0.13s, wps = 40463, train loss = 9.6032\r\nIteration 1698, time = 1.29s, wps = 43670, train loss = 6.2653\r\nIteration 1718, time = 2.38s, wps = 43072, train loss = 5.7573\r\nIteration 1738, time = 2.40s, wps = 42656, train loss = 5.4796\r\nIteration 1758, time = 2.38s, wps = 42996, train loss = 5.5244\r\nIteration 1778, time = 2.39s, wps = 42835, train loss = 5.2760\r\nIteration 1798, time = 2.37s, wps = 43159, train loss = 5.3952\r\nIteration 1818, time = 2.36s, wps = 43388, train loss = 5.3673\r\nIteration 1838, time = 2.37s, wps = 43276, train loss = 5.1904\r\nIteration 1858, time = 2.36s, wps = 43450, train loss = 5.2283\r\nIteration 1878, time = 2.37s, wps = 43183, train loss = 5.2852\r\nIteration 1898, time = 2.37s, wps = 43123, train loss = 5.1537\r\nIteration 1918, time = 2.39s, wps = 42910, train loss = 5.1917\r\nIteration 1938, time = 2.37s, wps = 43116, train loss = 5.1037\r\nIteration 1958, time = 2.40s, wps = 42714, train loss = 5.1471\r\nIteration 1978, time = 2.34s, wps = 43729, train loss = 5.1707\r\nIteration 1998, time = 2.37s, wps = 43210, train loss = 5.1987\r\nIteration 2018, time = 2.36s, wps = 43455, train loss = 5.2258\r\nIteration 2038, time = 2.40s, wps = 42578, train loss = 5.1705\r\nIteration 2058, time = 2.38s, wps = 42976, train loss = 5.1520\r\nIteration 2078, time = 2.37s, wps = 43233, train loss = 5.1081\r\nIteration 2098, time = 2.36s, wps = 43329, train loss = 5.0912\r\nIteration 2118, time = 2.39s, wps = 42915, train loss = 5.0336\r\nIteration 2138, time = 2.42s, wps = 42392, train loss = 5.1834\r\nIteration 2158, time = 2.41s, wps = 42498, train loss = 5.1764\r\nIteration 2178, time = 2.43s, wps = 42054, train loss = 5.0807\r\nIteration 2198, time = 2.39s, wps = 42920, train loss = 5.0449\r\nIteration 2218, time = 2.38s, wps = 42971, train loss = 5.1210\r\nIteration 2238, time = 2.39s, wps = 42810, train loss = 5.0886\r\nIteration 2258, time = 2.37s, wps = 43293, train loss = 5.0831\r\nIteration 2278, time = 2.36s, wps = 43421, train loss = 5.0909\r\nIteration 2298, time = 2.37s, wps = 43118, train loss = 5.0486\r\nIteration 2318, time = 2.35s, wps = 43585, train loss = 5.0188\r\nIteration 2338, time = 2.44s, wps = 42022, train loss = 5.0590\r\nIteration 2358, time = 2.38s, wps = 42990, train loss = 4.9272\r\nIteration 2378, time = 2.37s, wps = 43289, train loss = 5.0500\r\nIteration 2398, time = 2.45s, wps = 41724, train loss = 4.9856\r\nIteration 2418, time = 2.36s, wps = 43446, train loss = 5.0758\r\nIteration 2438, time = 2.34s, wps = 43741, train loss = 4.9630\r\nIteration 2458, time = 2.38s, wps = 42944, train loss = 5.0655\r\nIteration 2478, time = 2.45s, wps = 41783, train loss = 4.9605\r\nIteration 2498, time = 2.35s, wps = 43632, train loss = 5.0073\r\nIteration 2518, time = 2.39s, wps = 42790, train loss = 4.9711\r\nIteration 2538, time = 2.41s, wps = 42567, train loss = 5.0031\r\nIteration 2558, time = 2.42s, wps = 42400, train loss = 5.0181\r\nIteration 2578, time = 2.41s, wps = 42504, train loss = 4.9823\r\nIteration 2598, time = 2.40s, wps = 42630, train loss = 4.9870\r\nIteration 2618, time = 2.42s, wps = 42401, train loss = 4.9919\r\nIteration 2638, time = 2.44s, wps = 41888, train loss = 4.8977\r\nIteration 2658, time = 2.36s, wps = 43455, train loss = 4.9557\r\nIteration 2678, time = 2.39s, wps = 42842, train loss = 4.9760\r\nIteration 2698, time = 2.40s, wps = 42700, train loss = 4.9979\r\nIteration 2718, time = 2.40s, wps = 42586, train loss = 4.9647\r\nIteration 2738, time = 2.42s, wps = 42345, train loss = 4.9623\r\nIteration 2758, time = 2.43s, wps = 42118, train loss = 4.9696\r\nIteration 2778, time = 2.38s, wps = 43028, train loss = 4.9197\r\nIteration 2798, time = 2.40s, wps = 42663, train loss = 4.9882\r\nIteration 2818, time = 2.36s, wps = 43447, train loss = 5.0041\r\nIteration 2838, time = 2.39s, wps = 42926, train loss = 4.9814\r\nIteration 2858, time = 2.42s, wps = 42396, train loss = 4.8906\r\nIteration 2878, time = 2.38s, wps = 43072, train loss = 4.9684\r\nIteration 2898, time = 2.39s, wps = 42780, train loss = 4.8839\r\nIteration 2918, time = 2.38s, wps = 43013, train loss = 4.9438\r\nIteration 2938, time = 2.42s, wps = 42382, train loss = 4.9158\r\nIteration 2958, time = 2.39s, wps = 42872, train loss = 4.8627\r\nIteration 2978, time = 2.46s, wps = 41696, train loss = 4.9900\r\nIteration 2998, time = 2.41s, wps = 42408, train loss = 4.9702\r\nIteration 3018, time = 2.38s, wps = 42959, train loss = 4.9032\r\nIteration 3038, time = 2.40s, wps = 42754, train loss = 4.9259\r\nIteration 3058, time = 2.43s, wps = 42223, train loss = 4.8369\r\nIteration 3078, time = 2.41s, wps = 42545, train loss = 4.8835\r\nIteration 3098, time = 2.39s, wps = 42882, train loss = 4.9040\r\nIteration 3118, time = 2.38s, wps = 42957, train loss = 4.8566\r\nIteration 3138, time = 2.36s, wps = 43320, train loss = 4.8393\r\nIteration 3158, time = 2.40s, wps = 42588, train loss = 4.8855\r\nIteration 3178, time = 2.40s, wps = 42647, train loss = 4.8611\r\nIteration 3198, time = 2.36s, wps = 43448, train loss = 4.9116\r\nIteration 3218, time = 2.39s, wps = 42789, train loss = 4.8257\r\nIteration 3238, time = 2.41s, wps = 42556, train loss = 4.8717\r\nIteration 3258, time = 2.38s, wps = 42967, train loss = 4.8063\r\nProcessing file: \/projects\/1-billion-word-language-modeling-benchmark-r13output\/training-monolingual.tokenized.shuffled\/news.en-00096-of-00100\r\nFinished processing!\r\nIteration 3278, time = 4.03s, wps = 25391, train loss = 4.8513\r\nIteration 3298, time = 2.41s, wps = 42489, train loss = 4.7685\r\nIteration 3318, time = 2.40s, wps = 42735, train loss = 4.8255\r\nIteration 3338, time = 2.38s, wps = 43092, train loss = 4.8143\r\nIteration 3358, time = 2.37s, wps = 43154, train loss = 4.8304\r\nIteration 3378, time = 2.42s, wps = 42307, train loss = 4.8739\r\nIteration 3398, time = 2.40s, wps = 42619, train loss = 4.7210\r\nIteration 3418, time = 2.41s, wps = 42414, train loss = 4.8176\r\n\/usr\/local\/lib\/python3.5\/dist-packages\/tensorflow\/python\/summary\/writer\/writer.py:386: UserWarning: Attempting to use a closed FileWriter. The operation will be a noop unless the FileWriter is explicitly reopened.\r\n  warnings.warn(\"Attempting to use a closed FileWriter. \"<\/code><\/pre>\n<p>\u4f55\u3092\u4f55\u3068\u6bd4\u3079\u308b\u3079\u304d\u304b\u826f\u304f\u308f\u304b\u3089\u306a\u3044\u3002\u4e0a\u624b\u304f\u3044\u3063\u3066\u3044\u308b\u306e\u304b\uff1f\u6e29\u5ea6\u7b49\u3092\u30e2\u30cb\u30bf\u30fc\u3057\u305f\u3082\u306e\u304c\u4e0b\u8a18\u306e\u30b0\u30e9\u30d5\u3067\u3059\u3002<\/p>\n<p><a href=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-15-12-19-09.png\"><img loading=\"lazy\" decoding=\"async\" src=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-15-12-19-09-300x225.png\" alt=\"\" width=\"300\" height=\"225\" class=\"aligncenter size-medium wp-image-9477\" srcset=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-15-12-19-09-300x225.png 300w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-15-12-19-09-768x577.png 768w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-15-12-19-09-1024x770.png 1024w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-15-12-19-09-600x451.png 600w, http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/wp-content\/uploads\/2020\/07\/Screenshot-from-2020-09-15-12-19-09.png 1071w\" sizes=\"auto, (max-width: 300px) 100vw, 300px\" \/><\/a><\/p>\n<p>\u6700\u3082\u65e9\u304f\u7acb\u3061\u4e0a\u304c\u3063\u3066\u3044\u3066\u7acb\u3061\u4e0b\u304c\u3063\u3066\u3044\u308b\u9ed2\u3068\u9ec4\u306e\u7dda\u304c\u30b0\u30e9\u30d5\u30a3\u30c3\u30af\u30ab\u30fc\u30c9\u306e\u30e1\u30e2\u30ea\u30fc\u4f7f\u7528\u7387\u3067\u3059\u3002\u3053\u306e\u7acb\u3061\u4e0b\u304c\u3063\u3066\u3044\u308b\u3042\u305f\u308a\u304c\u8a08\u7b97\u7d42\u4e86\u6642\u523b\u3060\u3068\u601d\u3044\u307e\u3059\u3002\u6025\u6fc0\u306b\u7acb\u3061\u4e0a\u304c\u308a\u3060\u3089\u3060\u3089\u3068\u4e0b\u304c\u3063\u3066\u3044\u308b\u9752\u3068\u767d\u306e\u7dda\u304c\u6e29\u5ea6\u3002\u6e29\u5ea6\u304b\u3089\u9045\u308c\u3066\u6025\u6fc0\u306b\u7acb\u3061\u4e0a\u304c\u3063\u3066\u3001\u30b9\u30c8\u30f3\u3068\u843d\u3061\u3066\u3044\u308b\u8336\u3068\u9ed2\u306e\u7dda\u304c\u30d5\u30a1\u30f3\u306e\u56de\u8ee2\u6570\u3067\u3059\u3002\u8d64\u306e\u7dda\u306f\u30b1\u30fc\u30b9\u5185\u306e\u3069\u3053\u304b\u306e\u6e29\u5ea6\u3067\u3059\u3002<\/p>\n<p>\u4e00\u5fdc\u3053\u3053\u3067\u4e00\u533a\u5207\u308a\u3068\u3057\u3066\u3001\u3057\u3070\u3089\u304f tensorflow \u306e\u672c\u8aad\u307f\u3067\u3082\u3057\u3088\u3046\u304b\u3068\u601d\u3044\u307e\u3059\u3002<\/p>\n","protected":false},"excerpt":{"rendered":"<p>NVLink\u3092\u5229\u7528\u3057\uff0c\u30d3\u30c7\u30aa\u30ab\u30fc\u30c9\u30922\u679a\u523a\u3057\u3066\u8a08\u7b97\u3057\u3066\u307f\u305f\u3044\u3002\u4ee5\u4e0b\uff0c\u4e3b\u306a\u30d1\u30fc\u30c4\u3092 &#8230; <a class=\"more-link\" href=\"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/?p=9195\">Read More &raquo;<\/a><\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[4],"tags":[],"class_list":["post-9195","post","type-post","status-publish","format-standard","hentry","category-column"],"_links":{"self":[{"href":"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/9195","targetHints":{"allow":["GET"]}}],"collection":[{"href":"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=9195"}],"version-history":[{"count":195,"href":"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/9195\/revisions"}],"predecessor-version":[{"id":9491,"href":"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/9195\/revisions\/9491"}],"wp:attachment":[{"href":"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=9195"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=9195"},{"taxonomy":"post_tag","embeddable":true,"href":"http:\/\/zairyo.susi.oita-u.ac.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=9195"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}