{"id":705,"date":"2023-04-14T21:01:04","date_gmt":"2023-04-14T13:01:04","guid":{"rendered":"https:\/\/www.luckyshi.cn\/?p=705"},"modified":"2023-04-14T21:47:03","modified_gmt":"2023-04-14T13:47:03","slug":"%e8%b6%85%e7%ae%97%e4%bb%bb%e5%8a%a1%e6%8f%90%e4%ba%a4%e7%b3%bb%e7%bb%9f","status":"publish","type":"post","link":"https:\/\/www.luckyshi.cn\/?p=705","title":{"rendered":"\u8d85\u7b97\u4efb\u52a1\u63d0\u4ea4\u7cfb\u7edf"},"content":{"rendered":"<p><em>Abstract<\/em>: \u4ecb\u7ecd\u4e86Slurm\u7cfb\u7edf\u7684\u57fa\u7840\u7528\u6cd5\uff0c\u5217\u4e3e\u4e86\u4e00\u4e9b\u6837\u4f8b\u3002\u4ecb\u7ecd\u4e86<code>sbatch<\/code>,<code>srun<\/code>, <code>salloc<\/code>, <code>squeue<\/code>\u5e38\u89c1\u547d\u4ee4\u7684\u7528\u6cd5<\/p>\n<p><!-- more --><br \/>\n<!-- TOC --><\/p>\n<ul>\n<li><a class=\"wp-editor-md-post-content-link\" href=\"#1-Slurm\u7b80\u4ecb\u4e0e\u57fa\u7840\u547d\u4ee4\">Slurm\u7b80\u4ecb\u4e0e\u57fa\u7840\u547d\u4ee4<\/a>\n<ul>\n<li><a class=\"wp-editor-md-post-content-link\" href=\"#11-\u63d0\u4ea4\u4efb\u52a1\">\u63d0\u4ea4\u4efb\u52a1<\/a>\n<ul>\n<li><a class=\"wp-editor-md-post-content-link\" href=\"#111-sbatch\">sbatch<\/a><\/li>\n<li><a class=\"wp-editor-md-post-content-link\" href=\"#112-srun\">srun<\/a><\/li>\n<li><a class=\"wp-editor-md-post-content-link\" href=\"#113-\u5e38\u7528options\u4e0e\u91ca\u4e49\">\u5e38\u7528options\u4e0e\u91ca\u4e49<\/a><\/li>\n<\/ul>\n<\/li>\n<li><a class=\"wp-editor-md-post-content-link\" href=\"#12-\u67e5\u770b\u4efb\u52a1\">\u67e5\u770b\u4efb\u52a1<\/a><\/li>\n<li><a class=\"wp-editor-md-post-content-link\" href=\"#13-\u5176\u4ed6\">\u5176\u4ed6<\/a><\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<p><!-- \/TOC --><\/p>\n<p><a id=\"toc_anchor\" name=\"#1-Slurm\u7b80\u4ecb\u4e0e\u57fa\u7840\u547d\u4ee4\"><\/a><\/p>\n<h1>1. Slurm\u7b80\u4ecb\u4e0e\u57fa\u7840\u547d\u4ee4<\/h1>\n<p>\u73b0\u5728\u7684\u79d1\u7814\u4efb\u52a1\u5927\u90e8\u5206\u662f\u8fd0\u884c\u5728\u8d85\u7b97\u4e0a\u9762\u7684\uff0c\u672c\u5730\u7684\u5c0f\u673a\u5668\u5927\u90e8\u5206\u65f6\u95f4\u53ea\u662f\u8d77\u5230\u4e86\u4e00\u4e2a\u6587\u672c\u7f16\u8f91\u5668\u7684\u4f5c\u7528\u3002\u4e00\u822c\u8d85\u7b97\u4e0a\u4f1a\u6709\u5f88\u591a\u4e2a\u8282\u70b9\uff0c\u4e5f\u4f1a\u6709\u5f88\u591a\u4e2a\u7528\u6237\u63d0\u4ea4\u7684\u5f88\u591a\u4efb\u52a1\u3002 \u8fd9\u65f6\u5019\u6211\u4eec\u63d0\u4ea4\u4efb\u52a1\u5c31\u4e0d\u518d\u662f\u8ddf\u672c\u5730\u4e00\u6837<code>run<\/code>\u6216\u8005<code>F5<\/code>\u4e00\u4e0b\u5c31\u5f00\u59cb\u8fd0\u884c\u4e86\u7684\uff0c\u6211\u4eec\u9700\u8981\u5148\u7533\u8bf7\u8d44\u6e90\uff0c\u7b49\u5f85\u8d85\u7b97\u7684\u6279\u51c6\uff0c\u7136\u540e\u8d85\u7b97\u624d\u4f1a\u5c06\u6211\u4eec\u7684\u4efb\u52a1\u5206\u914d\u5230\u6307\u5b9a\u7684\u8282\u70b9\u4e0a\u8fd0\u884c\uff0c\u8fd9\u4e2a\u7cfb\u7edf\u5c31\u662f<code>SLURM<\/code>\u3002Slurm\u7cfb\u7edf\u5305\u542b\u5e38\u89c1\u7684\u4e24\u4e2a\u63d0\u4ea4\u4efb\u52a1\u7684\u547d\u4ee4<code>sbatch<\/code>, <code>srun<\/code>\uff1b\u67e5\u770b\u4efb\u52a1\u4fe1\u606f\u7684\u547d\u4ee4\uff0c<code>squeue<\/code>,<code>sinfo<\/code>; \u4ee5\u53ca\u5f00\u8f9f\u8d44\u6e90\u7684<code>salloc<\/code>\u548c\u67e5\u770b\u5386\u53f2\u7684<code>sacct<\/code>\u3002<\/p>\n<p><a id=\"toc_anchor\" name=\"#11-\u63d0\u4ea4\u4efb\u52a1\"><\/a><\/p>\n<h2>1.1. \u63d0\u4ea4\u4efb\u52a1<\/h2>\n<p>\u5411slurm\u7cfb\u7edf\u63d0\u4ea4\u4efb\u52a1\uff0c\u9700\u8981\u5305\u542b\u4e24\u4e2a\u4fe1\u606f\uff1a\u4e00\u4e2a\u662f\u9700\u8981\u7684\u8ba1\u7b97\u8d44\u6e90\uff0c\u5305\u62ec\u8ba1\u7b97\u65f6\u95f4\uff0c\u8282\u70b9\u6570\u7b49\uff1b\u7b2c\u4e8c\u4e2a\u662f\u5728\u8fd9\u4e2a\u8ba1\u7b97\u8d44\u6e90\u4e0b\u9700\u8981\u8fd0\u884c\u7684\u6307\u4ee4\uff0c\u5305\u62ec\u8f7d\u5165\u73af\u5883\u4e0e\u8fd0\u884c\u7a0b\u5e8f\u4e24\u90e8\u5206\u3002sbatch\u4e0esrun\u4e24\u4e2a\u6307\u4ee4\u5728\u8fd9\u4e24\u4e2a\u65b9\u9762\u7528\u6cd5\u4e0d\u540c\uff0c\u4f46\u662f\u5bf9\u5e94\u8d44\u6e90\u7684\u5173\u952e\u8bcd\u90fd\u662f\u4e00\u81f4\u7684\uff0c\u5927\u90e8\u5206\u60c5\u51b5\u4e0b\u53ea\u9700\u8981\u719f\u6089\u4e00\u4e2a\u6307\u4ee4\u5c31\u53ef\u4ee5\u4e86\u3002<\/p>\n<p><a id=\"toc_anchor\" name=\"#111-sbatch\"><\/a><\/p>\n<h3>1.1.1. sbatch<\/h3>\n<p>sbatch\u7684\u7528\u6cd5\u662f<code>sbatch xx.txt<\/code>\u3002\u5176\u4e2d\u4f60\u7684\u4efb\u52a1\u4fe1\u606f\u5c31\u5305\u542b\u5728<code>xx.txt<\/code>\u8fd9\u4e2a\u4ee5ASCII\u7801\u7f16\u7801\u7684\u6587\u4ef6\u4e2d\u3002\u6587\u4ef6\u57fa\u672c\u5f62\u5f0f\u662f\uff1a<\/p>\n<pre><code class=\"language-bash line-numbers\">#sbatch -N 1\n#sbatch --time 01:00:00\n#sbatch --mem 80GB\n#sbatch --gres=gpu:1\n#sbatch -J job_name\n#SBATCH -o .\/OUTPUT\/crack_detection.%J.out\n#SBATCH -e .\/OUTPUT\/crack_detection.%J.err\n\n.\/hellow_world\n<\/code><\/pre>\n<p>\u5176\u4e2d<code>#sbatch<\/code>\u540e\u9762\u8ddf\u7684\u4fe1\u606f\u5c31\u662f\u8ba1\u7b97\u8d44\u6e90\u3001\u4efb\u52a1\u547d\u540d\u5206\u533a\u4ee5\u53ca\u8f93\u5165\u8f93\u51fa\u6240\u5bf9\u5e94\u7684\u6587\u4ef6\u3002\u8fd9\u91cc\u9700\u8981\u6ce8\u610f\u7684\u662fsbatch\u63d0\u4ea4\u4efb\u52a1\u4e4b\u540e\u662f\u4e0d\u4f1a\u5c06\u7a0b\u5e8f\u7684\u8f93\u51fa\u4e0e\u62a5\u9519\u4fe1\u606f\u76f4\u63a5print\u51fa\u6765\uff0c\u800c\u662f\u63a5\u7ba1\u7a0b\u5e8f\u7684\u7cfb\u7edf\u8f93\u51fa\u4e0e\u62a5\u9519\u5230\u6307\u5b9a\u6587\u4ef6\u4e2d\u3002\u4f46\u662f\u5bf9\u6587\u4ef6\u7684\u8bfb\u5199\u5185\u5bb9\u662f\u6b63\u5e38\u8fdb\u884c\u7684\uff0c\u56e0\u4e3a\u8d85\u7b97\u7cfb\u7edf\u7684\u6587\u4ef6\u7cfb\u7edf\u4e00\u822c\u662f\u4e00\u4e2a\u6574\u4f53\uff0c\u5982\u679c\u6709\u9ad8IO\u7684\u5185\u5bb9\u5c31\u9700\u8981\u5c06\u6587\u4ef6\u5b58\u653e\u5728\u76f8\u5e94\u7684\u533a\u57df\u5185\u3002sbatch\u63d0\u4ea4\u4efb\u52a1\u4e4b\u540e\uff0c\u8d85\u7b97\u7684\u64cd\u4f5c\u7cfb\u7edf\u8bb0\u5f55\u8fd9\u4e00\u4e2a\u4efb\u52a1\uff0c\u4f1a\u6839\u636e\u4f18\u5148\u7ea7\u7ed9\u4efb\u52a1\u6392\u961f\uff0c\u5728\u5408\u9002\u7684\u65f6\u95f4\u8fd0\u884c\u3002\u7528\u6237\u6b64\u65f6\u53ef\u4ee5\u65ad\u5f00\u7ec8\u7aef\u6216\u8fd0\u884c\u5176\u4ed6\u4efb\u52a1\u3002<\/p>\n<p><a id=\"toc_anchor\" name=\"#112-srun\"><\/a><\/p>\n<h3>1.1.2. srun<\/h3>\n<p>srun\u662f\u76f4\u63a5\u4ee5\u547d\u4ee4\u800c\u975e\u811a\u672c\u7684\u5f62\u5f0f\u7533\u8bf7\u8d44\u6e90\u5e76\u8fd0\u884c\u4efb\u52a1\uff0c\u57fa\u672c\u7528\u6cd5\u662f <code>srun -o1 xx1 --option2 xx2 .\/hello_world<\/code>\u3002srun\u4f1a\u5b9e\u65f6\u8f93\u51fa\u4f60\u7684\u7a0b\u5e8f\u7684\u8fd0\u884c\u7ed3\u679c\uff0c\u7528\u6237\u53ef\u4ee5\u4e2d\u9014\u76f4\u63a5\u7ec8\u6b62\u7a0b\u5e8f\u3002\u4f46\u662f\u4e00\u65e6\u8fdc\u7a0b\u8fde\u63a5\u7aef\u53e3\u7ec8\u7aef\uff0c\u4efb\u52a1\u5c31\u4f1a\u505c\u6b62\u3002sbatch\u4e0esrun\u63d0\u4ea4\u4efb\u52a1\u7ed9\u8d85\u7b97\u7684\u64cd\u4f5c\u7cfb\u7edf\uff0c\u4e0d\u5b58\u5728\u54ea\u4e00\u4e2a\u6307\u4ee4\u53ef\u4ee5\u63d2\u961f\u7684\u60c5\u51b5\u3002<\/p>\n<p><a id=\"toc_anchor\" name=\"#113-\u5e38\u7528options\u4e0e\u91ca\u4e49\"><\/a><\/p>\n<h3>1.1.3. \u5e38\u7528options\u4e0e\u91ca\u4e49<\/h3>\n<table>\n<thead>\n<tr>\n<th>options<\/th>\n<th>\u91ca\u4e49<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>&#8211;batch=?<\/td>\n<td>\u6307\u5b9a\u4efb\u52a1\u6240\u5728\u5206\u533a\uff0c\u6709\u7684\u8d85\u7b97\u4f1a\u8bbe\u7f6ebatch\u4e0edebug\u4e24\u4e2a\u5206\u533a\uff0cdebug\u5206\u533a\u57fa\u672c\u4e0d\u7528\u6392\u961f\uff0c\u4f46\u662f\u4efb\u52a1\u8fd0\u884c\u65f6\u95f4\u6709\u9650\u5236<\/td>\n<\/tr>\n<tr>\n<td>-N, &#8211;nodes=?<\/td>\n<td>\u7533\u8bf7\u7684\u8282\u70b9\u4e2a\u6570\u3002\u4e00\u822c\u7684\u8d85\u7b97\u7cfb\u7edf\u4e0a\u6bcf\u4e2a\u8282\u70b9\u4f1a\u67091\u4e2acpu\uff1a\u6bcf\u4e2aCPU\u7684\u5178\u578b\u914d\u7f6e\u662f32\u4e2a\u6838\uff08\u6216\u8005\u7406\u89e3\u4e3a\u7269\u7406\u7ebf\u7a0b\uff0c\u4e0d\u540c\u67b6\u6784\u7684\u8d85\u7b97\u4f1a\u6709\u5dee\u5f02\uff09\uff0c\u5176\u4e2d\u6bcf16\u4e2a\u6838\u662f\u5728\u540c\u4e00\u4e2apie\u4e0a\u7684\uff0c\u4e0d\u540cpie\u4e0a\u7684\u6838\u7684\u901a\u8baf\u65f6\u95f4\u4f1a\u66f4\u5927\uff0c\u4f46\u662f\u8981\u8fdc\u8fdc\u5c0f\u4e8e\u4e0d\u540c\u8282\u70b9\u4e4b\u95f4\u7684\u901a\u8baf\u65f6\u95f4<\/td>\n<\/tr>\n<tr>\n<td>-n, &#8211;ntasks=?<\/td>\n<td>\u5c06\u6838\u5fc3\u7a0b\u5e8f\u590d\u5236\u591a\u5c11\u4efd\uff0c\u76f8\u5f53\u4e8e\u5e76\u884c\u8ba1\u7b97\u4e2d\u6700\u9876\u5c42\u7684\u5e76\u884c\u6570\u3002\u5982\u679c\u4efb\u52a1\u662fmpi\u7684\u65b9\u5f0f\u8fd0\u884c\uff0c\u8fd9\u4e2a\u53c2\u6570\u5c06\u8f6c\u5316\u4e3ampi\u91cc\u9762\u7684ntasks; \u5982\u679c\u662f\u5176\u4ed6\u7684\u65b9\u5f0f\u5e76\u884c\uff0c\u7a0b\u5e8f\u5185\u90e8\u53ef\u4ee5\u8c03\u51fa\u4efb\u52a1\u53f7\uff0c\u4ece\u800c\u6267\u884c\u4e0d\u540c\u7684\u4efb\u52a1\uff0c\u6bd4\u5982<code>printenv SLURM_PROCID<\/code>\u5c31\u53ef\u4ee5\u8c03\u51fa\u76f8\u5e94\u7684\u4efb\u52a1id<\/td>\n<\/tr>\n<tr>\n<td>-c, &#8211;cpus-per-task=?<\/td>\n<td>\u4e0e\u4e0a\u9762\u7684-n\u914d\u5408\u6267\u884c\uff0c\u6307\u5b9a\u6bcf\u4e2a\u4efb\u52a1\u6709\u591a\u5c11\u4e2a\u6838\u6765\u8fd0\u884c\uff08\u5e76\u4e0d\u662f\u4e00\u4e2a\u5927\u7684cpu, \u800c\u662f\u6838\u6570\uff09\uff0c\u5b9e\u9645\u5bf9\u5e94\u7740mpi\u91cc\u9762\u7684<code>OMP_NUM_THREADS<\/code>\uff0c\u6b21\u7ea7\u5e76\u884c\u53c2\u6570<\/td>\n<\/tr>\n<tr>\n<td>&#8211;mem=?<\/td>\n<td>\u6bcf\u4e2a\u8282\u70b9\u9884\u5148\u5206\u914d\u7684\u5185\u5b58\u5927\u5c0f\uff0c\u53ef\u4ee5\u4f7f\u7528GB,MB, KB\u4e09\u4e2a\u5355\u4f4d<\/td>\n<\/tr>\n<tr>\n<td>&#8211;gres=<em>name:type:num<\/em><\/td>\n<td>\u53ef\u4ee5\u6839\u636e\u8d85\u7b97\u914d\u7f6e\u7533\u8bf7\u76f8\u5e94\u7684\u8d44\u6e90\uff0c\u9700\u8981\u8d85\u7b97\u652f\u6301\u3002\u6bd4\u5982\u7533\u8bf7\u4e24\u4e2a\u578b\u53f7\u4e3akepler\u7684GPU\u4e0e\u4e24\u4e2a\u578b\u53f7\u4e3av100\u7684GPU,\u5c31\u662f<code>--gres=gpu:kepler:2,gpu:v100:2<\/code><\/td>\n<\/tr>\n<tr>\n<td>-t, &#8211;time=xx:xx:xx<\/td>\n<td>\u7a0b\u5e8f\u8fd0\u884c\u7684\u6700\u5927\u65f6\u95f4\uff0c\u8d85\u8fc7\u8fd9\u4e2a\u65f6\u95f4\u8d85\u7b97\u4f1a\u968f\u65f6\u505c\u6b62\u4f60\u7684\u4efb\u52a1\uff0c\u5bf9\u5e94\u7684\u5355\u4f4d\u662f  <code>\u65f6:\u5206:\u79d2<\/code><\/td>\n<\/tr>\n<tr>\n<td>-e, &#8211;error=<em>file_name<\/em><\/td>\n<td>file_name\u53ef\u4ee5\u4f7f\u7528\u6a21\u677f\u3002 \u6bd4\u5982\u5e38\u89c1\u7684\u5355\u4efb\u52a1\u811a\u672c\u8f93\u51fa\u8bbe\u7f6e\u4e3a<code>xxx_%J_xxx.txt<\/code>\uff0c\u5176\u4e2d<code>%J<\/code>\u4f1a\u88ab\u4efb\u52a1\u5e8f\u53f7\u4ee3\u66ff<\/td>\n<\/tr>\n<tr>\n<td>-o, &#8211;output=<em>file_name<\/em><\/td>\n<td>\u540c\u4e0a<\/td>\n<\/tr>\n<tr>\n<td>&#8211;exclusive<\/td>\n<td>\u662f\u5426\u5141\u8bb8\u540c\u4e00\u4e2a\u8282\u70b9\u4e0a\u5141\u8bb8\u5176\u4ed6\u4eba\u7684\u7a0b\u5e8f\u8fd0\u884c\uff0c\u9632\u6b62\u8d44\u6e90\u5171\u4eab\u540e\u8fd0\u884c\u901f\u5ea6\u6162\u7684\u95ee\u9898\uff0c\u4f46\u662f\u4e00\u822c\u4e0d\u88ab\u63a8\u8350\u4f7f\u7528<\/td>\n<\/tr>\n<tr>\n<td>-C, &#8211;constraint=&#8221;?&#8221;<\/td>\n<td>\u5bf9\u7533\u8bf7\u7684\u8282\u70b9\u4fe1\u606f\u7684\u9650\u5236\uff0c\u4e00\u822c\u662f\u7528\u6765\u9650\u5236cpu\u7684\u578b\u53f7\uff0c\u9632\u6b62\u7f16\u8bd1\u597d\u7684\u7a0b\u5e8f\u88ab\u7528\u5728\u9519\u8bef\u7684\u67b6\u6784\u4e0a\uff0c\u6bd4\u5982<code>--constraint=\"AMD\"<\/code> \u5c31\u9650\u5b9a\u53ea\u80fd\u8fd0\u884c\u5728AMD cpu\u7684\u8282\u70b9\u4e0a\uff0c\u4e0d\u8981\u5206\u914dIntel\u7684cpu(<code>AMD, YES!<\/code>)<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p>\u66f4\u591a\u7684\u7528\u6cd5\u8bf7\u53c2\u8003<a class=\"wp-editor-md-post-content-link\" href=\"https:\/\/slurm.schedmd.com\/sbatch.html\">slurm\u7684\u8bf4\u660e\u6587\u6863<\/a>\uff0c\u6216\u8005\u4f7f\u7528 <code>sbatch --help<\/code>\u8c03\u51fa\u6765\u3002\u4e5f\u53ef\u4ee5\u4f7f\u7528\u4e00\u4e2a\u811a\u672c\u751f\u6210\u5668\uff0c <a class=\"wp-editor-md-post-content-link\" href=\"https:\/\/www.hpc.kaust.edu.sa\/ibex\/job\">IBEX Jobscript generator<\/a>\u3002\u4e0b\u9762\u662f\u6211\u7684\u4e00\u4e2a\u7a0b\u5e8f\u811a\u672c:<\/p>\n<pre><code class=\"language-bash line-numbers\">#!\/bin\/bash\n#SBATCH -N 1\n#SBATCH --partition=batch\n#SBATCH -J crack_detection\n#SBATCH -o .\/OUTPUT\/crack_detection.%J.out\n#SBATCH -e .\/OUTPUT\/crack_detection.%J.err\n#SBATCH --mail-user=example@example.com\n#SBATCH --mail-type=ALL\n#SBATCH --time=02:00:00\n#SBATCH --mem=80GB\n#SBATCH --gres=gpu:1\n\n# load envs\nmodule load tensorflow\/1.13.1-cuda10.0-cudnn7.6-py3.6\nmodule load keras\/2.2.4-cuda10.0-cudnn7.6-py3.6\nmodule load anaconda3\/4.4.0\n\n#run the application:\npython trainCNN.py\n<\/code><\/pre>\n<p><a id=\"toc_anchor\" name=\"#12-\u67e5\u770b\u4efb\u52a1\"><\/a><\/p>\n<h2>1.2. \u67e5\u770b\u4efb\u52a1<\/h2>\n<p>\u67e5\u770b\u4efb\u52a1\u7684\u8fd0\u884c\u72b6\u6001\uff0c\u53ef\u4ee5\u4f7f\u7528<code>squeue<\/code>\u547d\u4ee4\u8fdb\u884c\u3002\u67e5\u770b\u7279\u5b9a\u7528\u6237\u7684\u53ef\u4ee5\u4f7f\u7528<code>-u your_name<\/code>\uff0c\u4e5f\u53ef\u4ee5\u4e0e<code>grep<\/code>\u914d\u5408\u4f7f\u7528\uff0c\u68c0\u7d22\u7279\u5b9a\u7684\u5173\u952e\u8bcd\u3002\u5982\u679c\u5411\u5b9e\u65f6\u66f4\u65b0\uff0c\u53ef\u4ee5\u4e0ewatch\u547d\u4ee4\u914d\u5408\uff0c<code>watch -n ?? squeue -u your_name<\/code>\uff0c\u5c31\u4f1a\u8fdb\u5165\u4e00\u4e2a\u5b9e\u65f6\u66f4\u65b0\u753b\u9762\uff0c\u9891\u7387\u662f<code>??<\/code>s.<\/p>\n<pre><code class=\"language-bash line-numbers\">-bash-4.2<span class=\"katex math inline\">squeue -u intra5\nJOBID       PARTITION     NAME   USER    ST    TIME    NODES NODELIST(REASON)\n10471799_47     batch   intra5  soldanm  R    3:22:44      1 gpu211-14\n10471799_46     batch   intra5  soldanm  R    3:29:06      1 gpu208-06\n-bash-4.2<\/span> squeue -J 10471799_47\nJOBID       PARTITION     NAME   USER    ST    TIME    NODES NODELIST(REASON)\n10471799_47     batch   intra5  soldanm  R    3:22:44      1 gpu211-14\n-bash-4.2<span class=\"katex math inline\">squeue |grep soldanm\nJOBID       PARTITION     NAME   USER    ST    TIME    NODES NODELIST(REASON)\n10471799_47     batch   intra5  soldanm  R    3:22:44      1 gpu211-14\n10471799_46     batch   intra5  soldanm  R    3:29:06      1 gpu208-06\n-bash-4.2<\/span> watch -n 10 squeue -u intra5\nEvery 10.0s: squeue -u intra5                  Wed May  6 09:27:25 2020\nJOBID       PARTITION     NAME   USER    ST    TIME    NODES NODELIST(REASON)\n10471799_47     batch   intra5  soldanm  R    3:22:44      1 gpu211-14\n10471799_46     batch   intra5  soldanm  R    3:29:06      1 gpu208-06\n\n<\/code><\/pre>\n<p>\u5982\u679c\u60f3\u67e5\u770b\u8d85\u7b97\u4e0a\u8d44\u6e90\u4f7f\u7528\u95ee\u9898\uff0c\u53ef\u4ee5\u4f7f\u7528<code>sinfo<\/code>\u547d\u4ee4\u67e5\u770b\uff0c\u67e5\u770b\u7a7a\u95f2\u8d44\u6e90\uff0c\u53ef\u4ee5\u4f7f\u7528<code>sinfo | grep idle<\/code>\u67e5\u770b\u7a7a\u95f2\u8282\u70b9\u3002<\/p>\n<p><a id=\"toc_anchor\" name=\"#13-\u5176\u4ed6\"><\/a><\/p>\n<h2>1.3. \u5176\u4ed6<\/h2>\n<p>\u5982\u679c\u60f3\u7533\u8bf7\u4e00\u4e2a\u5b9e\u65f6\u7684\u4ea4\u4e92\u8d44\u6e90\uff0c\u53ef\u4ee5\u4f7f\u7528<code>salloc<\/code>\u547d\u4ee4\uff0c\u5176\u5173\u952e\u8bcd\u662f\u4e0e<code>srun<\/code>\u7684\u9009\u9879\u662f\u76f8\u540c\u7684\uff0c\u53ea\u662f\u4e0d\u9700\u8981\u6307\u5b9a\u8fd0\u884c\u4ec0\u4e48\u7a0b\u5e8f\u3002\u6bd4\u5982\u4e0b\u9762\u7684\u6307\u4ee4\uff0c\u4f60\u5c31\u5df2\u7ecf\u7533\u8bf7\u5230\u4e86\u4e00\u4e2a\u8282\u70b9\uff0c\u8fd0\u884c\u65f6\u95f4\u662f10\u5206\u949f\uff0c\u5728\u8fd9\u4e2a\u5e73\u53f0\u4e0a\uff0c\u4f60\u53ef\u4ee5\u5b9e\u65f6\u8fd0\u884c\u6307\u4ee4\uff0c\u4e0e\u672c\u5730\u8fd0\u884c\u76f8\u540c\u3002\u4f46\u662f\u76f8\u6bd4\u4e8e<code>sbatch<\/code>\uff0c<code>srun<\/code>\u547d\u4ee4,\u53ef\u80fd\u5b58\u5728\u4e00\u4e9b\u6027\u80fd\u635f\u5931.<\/p>\n<pre><code class=\"language-bash line-numbers\">bash-4.2<span class=\"katex math inline\">echo CORES_NUM\n12\n-bash-4.2<\/span> salloc --time 00:10:00 -N 1\nsalloc: Pending job allocation 10516365\nsalloc: job 10516365 queued and waiting for resources\nsalloc: job 10516365 has been allocated resources\nsalloc: Granted job allocation 10516365\nsalloc: Waiting for resource configuration\nsalloc: Nodes cn509-21-r are ready for job\nbash-4.2<span class=\"katex math inline\">echo NODE_NAME\ncn509-21-r\nbash-4.2<\/span> echo CORES_NUM\n32\nbash-4.2$ .\/hello_world\n<\/code><\/pre>\n","protected":false},"excerpt":{"rendered":"<p>Abstract: \u4ecb\u7ecd\u4e86Slurm\u7cfb\u7edf\u7684\u57fa\u7840\u7528\u6cd5\uff0c\u5217\u4e3e\u4e86\u4e00\u4e9b\u6837\u4f8b\u3002\u4ecb\u7ecd\u4e86sbatch,srun, salloc, squeue\u5e38\u89c1\u547d\u4ee4\u7684\u7528\u6cd5 Slurm\u7b80\u4ecb\u4e0e\u57fa\u7840\u547d\u4ee4 \u63d0\u4ea4\u4efb\u52a1 sbatch srun \u5e38\u7528options\u4e0e\u91ca\u4e49 \u67e5\u770b\u4efb\u52a1 \u5176\u4ed6 1. Slurm\u7b80\u4ecb\u4e0e\u57fa\u7840\u547d\u4ee4 \u73b0\u5728\u7684\u79d1\u7814\u4efb\u52a1\u5927\u90e8\u5206\u662f\u8fd0\u884c\u5728\u8d85\u7b97\u4e0a\u9762\u7684\uff0c\u672c\u5730\u7684\u5c0f\u673a\u5668\u5927\u90e8\u5206\u65f6\u95f4\u53ea\u662f\u8d77\u5230\u4e86\u4e00\u4e2a\u6587\u672c\u7f16\u8f91\u5668\u7684\u4f5c\u7528\u3002\u4e00\u822c\u8d85\u7b97\u4e0a\u4f1a\u6709\u5f88\u591a\u4e2a\u8282\u70b9\uff0c\u4e5f\u4f1a\u6709\u5f88\u591a\u4e2a\u7528\u6237\u63d0\u4ea4\u7684\u5f88\u591a\u4efb\u52a1\u3002 \u8fd9\u65f6\u5019\u6211\u4eec\u63d0\u4ea4\u4efb\u52a1\u5c31\u4e0d\u518d\u662f\u8ddf\u672c\u5730\u4e00\u6837run\u6216\u8005F5\u4e00\u4e0b\u5c31\u5f00\u59cb\u8fd0\u884c\u4e86\u7684\uff0c\u6211\u4eec\u9700\u8981\u5148\u7533\u8bf7\u8d44\u6e90\uff0c\u7b49\u5f85\u8d85\u7b97\u7684\u6279\u51c6\uff0c\u7136\u540e\u8d85\u7b97\u624d\u4f1a\u5c06\u6211\u4eec\u7684\u4efb\u52a1\u5206\u914d\u5230\u6307\u5b9a\u7684\u8282\u70b9\u4e0a\u8fd0\u884c\uff0c\u8fd9\u4e2a\u7cfb\u7edf\u5c31\u662fSLURM\u3002Slurm\u7cfb\u7edf\u5305\u542b\u5e38\u89c1\u7684\u4e24\u4e2a\u63d0\u4ea4\u4efb\u52a1\u7684\u547d\u4ee4sbatch, srun\uff1b\u67e5\u770b\u4efb\u52a1\u4fe1\u606f\u7684\u547d\u4ee4\uff0csqueue,sinfo; \u4ee5\u53ca\u5f00\u8f9f\u8d44\u6e90\u7684salloc\u548c\u67e5\u770b\u5386\u53f2\u7684sacct\u3002 1.1. \u63d0\u4ea4\u4efb\u52a1 \u5411slurm\u7cfb\u7edf\u63d0\u4ea4\u4efb\u52a1\uff0c\u9700\u8981\u5305\u542b\u4e24\u4e2a\u4fe1\u606f\uff1a\u4e00\u4e2a\u662f\u9700\u8981\u7684\u8ba1\u7b97\u8d44\u6e90\uff0c\u5305\u62ec\u8ba1\u7b97\u65f6\u95f4\uff0c\u8282\u70b9\u6570\u7b49\uff1b\u7b2c\u4e8c\u4e2a\u662f\u5728\u8fd9\u4e2a\u8ba1\u7b97\u8d44\u6e90\u4e0b\u9700\u8981\u8fd0\u884c\u7684\u6307\u4ee4\uff0c\u5305\u62ec\u8f7d\u5165\u73af\u5883\u4e0e\u8fd0\u884c\u7a0b\u5e8f\u4e24\u90e8\u5206\u3002sbatch\u4e0esrun\u4e24\u4e2a\u6307\u4ee4\u5728\u8fd9\u4e24\u4e2a\u65b9\u9762\u7528\u6cd5\u4e0d\u540c\uff0c\u4f46\u662f\u5bf9\u5e94\u8d44\u6e90\u7684\u5173\u952e\u8bcd\u90fd\u662f\u4e00\u81f4\u7684\uff0c\u5927\u90e8\u5206\u60c5\u51b5\u4e0b\u53ea\u9700\u8981\u719f\u6089\u4e00\u4e2a\u6307\u4ee4\u5c31\u53ef\u4ee5\u4e86\u3002 1.1.1. sbatch sbatch\u7684\u7528\u6cd5\u662fsbatch xx.txt\u3002\u5176\u4e2d\u4f60\u7684\u4efb\u52a1\u4fe1\u606f\u5c31\u5305\u542b\u5728xx.txt\u8fd9\u4e2a\u4ee5ASCII\u7801\u7f16\u7801\u7684\u6587\u4ef6\u4e2d\u3002\u6587\u4ef6\u57fa\u672c\u5f62\u5f0f\u662f\uff1a #sbatch -N 1 #sbatch &#8211;time 01:00:00 #sbatch &#8211;mem 80GB #sbatch &#8211;gres=gpu:1 #sbatch -J job_name #SBATCH -o .\/OUTPUT\/crack_detection.%J.out #SBATCH -e .\/OUTPUT\/crack_detection.%J.err .\/hellow_world \u5176\u4e2d#sbatch\u540e\u9762\u8ddf\u7684\u4fe1\u606f\u5c31\u662f\u8ba1\u7b97\u8d44\u6e90\u3001\u4efb\u52a1\u547d\u540d\u5206\u533a\u4ee5\u53ca\u8f93\u5165\u8f93\u51fa\u6240\u5bf9\u5e94\u7684\u6587\u4ef6\u3002\u8fd9\u91cc\u9700\u8981\u6ce8\u610f\u7684\u662fsbatch\u63d0\u4ea4\u4efb\u52a1\u4e4b\u540e\u662f\u4e0d\u4f1a\u5c06\u7a0b\u5e8f\u7684\u8f93\u51fa\u4e0e\u62a5\u9519\u4fe1\u606f\u76f4\u63a5print\u51fa\u6765\uff0c\u800c\u662f\u63a5\u7ba1\u7a0b\u5e8f\u7684\u7cfb\u7edf\u8f93\u51fa\u4e0e\u62a5\u9519\u5230\u6307\u5b9a\u6587\u4ef6\u4e2d\u3002\u4f46\u662f\u5bf9\u6587\u4ef6\u7684\u8bfb\u5199\u5185\u5bb9\u662f\u6b63\u5e38\u8fdb\u884c\u7684\uff0c\u56e0\u4e3a\u8d85\u7b97\u7cfb\u7edf\u7684\u6587\u4ef6\u7cfb\u7edf\u4e00\u822c\u662f\u4e00\u4e2a\u6574\u4f53\uff0c\u5982\u679c\u6709\u9ad8IO\u7684\u5185\u5bb9\u5c31\u9700\u8981\u5c06\u6587\u4ef6\u5b58\u653e\u5728\u76f8\u5e94\u7684\u533a\u57df\u5185\u3002sbatch\u63d0\u4ea4\u4efb\u52a1\u4e4b\u540e\uff0c\u8d85\u7b97\u7684\u64cd\u4f5c\u7cfb\u7edf\u8bb0\u5f55\u8fd9\u4e00\u4e2a\u4efb\u52a1\uff0c\u4f1a\u6839\u636e\u4f18\u5148\u7ea7\u7ed9\u4efb\u52a1\u6392\u961f\uff0c\u5728\u5408\u9002\u7684\u65f6\u95f4\u8fd0\u884c\u3002\u7528\u6237\u6b64\u65f6\u53ef\u4ee5\u65ad\u5f00\u7ec8\u7aef\u6216\u8fd0\u884c\u5176\u4ed6\u4efb\u52a1\u3002 1.1.2. srun srun\u662f\u76f4\u63a5\u4ee5\u547d\u4ee4\u800c\u975e\u811a\u672c\u7684\u5f62\u5f0f\u7533\u8bf7\u8d44\u6e90\u5e76\u8fd0\u884c\u4efb\u52a1\uff0c\u57fa\u672c\u7528\u6cd5\u662f srun -o1 xx1 &#8211;option2 xx2 .\/hello_world\u3002srun\u4f1a\u5b9e\u65f6\u8f93\u51fa\u4f60\u7684\u7a0b\u5e8f\u7684\u8fd0\u884c\u7ed3\u679c\uff0c\u7528\u6237\u53ef\u4ee5\u4e2d\u9014\u76f4\u63a5\u7ec8\u6b62\u7a0b\u5e8f\u3002\u4f46\u662f\u4e00\u65e6\u8fdc\u7a0b\u8fde\u63a5\u7aef\u53e3\u7ec8\u7aef\uff0c\u4efb\u52a1\u5c31\u4f1a\u505c\u6b62\u3002sbatch\u4e0esrun\u63d0\u4ea4\u4efb\u52a1\u7ed9\u8d85\u7b97\u7684\u64cd\u4f5c\u7cfb\u7edf\uff0c\u4e0d\u5b58\u5728\u54ea\u4e00\u4e2a\u6307\u4ee4\u53ef\u4ee5\u63d2\u961f\u7684\u60c5\u51b5\u3002 1.1.3. \u5e38\u7528options\u4e0e\u91ca\u4e49 options \u91ca\u4e49 &#8211;batch=? \u6307\u5b9a\u4efb\u52a1\u6240\u5728\u5206\u533a\uff0c\u6709\u7684\u8d85\u7b97\u4f1a\u8bbe\u7f6ebatch\u4e0edebug\u4e24\u4e2a\u5206\u533a\uff0cdebug\u5206\u533a\u57fa\u672c\u4e0d\u7528\u6392\u961f\uff0c\u4f46\u662f\u4efb\u52a1\u8fd0\u884c\u65f6\u95f4\u6709\u9650\u5236 -N, &#8211;nodes=? \u7533\u8bf7\u7684\u8282\u70b9\u4e2a\u6570\u3002\u4e00\u822c\u7684\u8d85\u7b97\u7cfb\u7edf\u4e0a\u6bcf\u4e2a\u8282\u70b9\u4f1a\u67091\u4e2acpu\uff1a\u6bcf\u4e2aCPU\u7684\u5178\u578b\u914d\u7f6e\u662f32\u4e2a\u6838\uff08\u6216\u8005\u7406\u89e3\u4e3a\u7269\u7406\u7ebf\u7a0b\uff0c\u4e0d\u540c\u67b6\u6784\u7684\u8d85\u7b97\u4f1a\u6709\u5dee\u5f02\uff09\uff0c\u5176\u4e2d\u6bcf16\u4e2a\u6838\u662f\u5728\u540c\u4e00\u4e2apie\u4e0a\u7684\uff0c\u4e0d\u540cpie\u4e0a\u7684\u6838\u7684\u901a\u8baf\u65f6\u95f4\u4f1a\u66f4\u5927\uff0c\u4f46\u662f\u8981\u8fdc\u8fdc\u5c0f\u4e8e\u4e0d\u540c\u8282\u70b9\u4e4b\u95f4\u7684\u901a\u8baf\u65f6\u95f4 -n, &#8211;ntasks=? \u5c06\u6838\u5fc3\u7a0b\u5e8f\u590d\u5236\u591a\u5c11\u4efd\uff0c\u76f8\u5f53\u4e8e\u5e76\u884c\u8ba1\u7b97\u4e2d\u6700\u9876\u5c42\u7684\u5e76\u884c\u6570\u3002\u5982\u679c\u4efb\u52a1\u662fmpi\u7684\u65b9\u5f0f\u8fd0\u884c\uff0c\u8fd9\u4e2a\u53c2\u6570\u5c06\u8f6c\u5316\u4e3ampi\u91cc\u9762\u7684ntasks; \u5982\u679c\u662f\u5176\u4ed6\u7684\u65b9\u5f0f\u5e76\u884c\uff0c\u7a0b\u5e8f\u5185\u90e8\u53ef\u4ee5\u8c03\u51fa\u4efb\u52a1\u53f7\uff0c\u4ece\u800c\u6267\u884c\u4e0d\u540c\u7684\u4efb\u52a1\uff0c\u6bd4\u5982printenv SLURM_PROCID\u5c31\u53ef\u4ee5\u8c03\u51fa\u76f8\u5e94\u7684\u4efb\u52a1id -c, &#8211;cpus-per-task=? \u4e0e\u4e0a\u9762\u7684-n\u914d\u5408\u6267\u884c\uff0c\u6307\u5b9a\u6bcf\u4e2a\u4efb\u52a1\u6709\u591a\u5c11\u4e2a\u6838\u6765\u8fd0\u884c\uff08\u5e76\u4e0d\u662f\u4e00\u4e2a\u5927\u7684cpu, \u800c\u662f\u6838\u6570\uff09\uff0c\u5b9e\u9645\u5bf9\u5e94\u7740mpi\u91cc\u9762\u7684OMP_NUM_THREADS\uff0c\u6b21\u7ea7\u5e76\u884c\u53c2\u6570 &#8211;mem=? \u6bcf\u4e2a\u8282\u70b9\u9884\u5148\u5206\u914d\u7684\u5185\u5b58\u5927\u5c0f\uff0c\u53ef\u4ee5\u4f7f\u7528GB,MB, KB\u4e09\u4e2a\u5355\u4f4d &#8211;gres=name:type:num \u53ef\u4ee5\u6839\u636e\u8d85\u7b97\u914d\u7f6e\u7533\u8bf7\u76f8\u5e94\u7684\u8d44\u6e90\uff0c\u9700\u8981\u8d85\u7b97\u652f\u6301\u3002\u6bd4\u5982\u7533\u8bf7\u4e24\u4e2a\u578b\u53f7\u4e3akepler\u7684GPU\u4e0e\u4e24\u4e2a\u578b\u53f7\u4e3av100\u7684GPU,\u5c31\u662f&#8211;gres=gpu:kepler:2,gpu:v100:2 -t, &#8211;time=xx:xx:xx \u7a0b\u5e8f\u8fd0\u884c\u7684\u6700\u5927\u65f6\u95f4\uff0c\u8d85\u8fc7\u8fd9\u4e2a\u65f6\u95f4\u8d85\u7b97\u4f1a\u968f\u65f6\u505c\u6b62\u4f60\u7684\u4efb\u52a1\uff0c\u5bf9\u5e94\u7684\u5355\u4f4d\u662f \u65f6:\u5206:\u79d2 -e, &#8211;error=file_name file_name\u53ef&#8230;<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[6],"tags":[20],"class_list":["post-705","post","type-post","status-publish","format-standard","hentry","category-softwares","tag-sbatch"],"jetpack_featured_media_url":"","_links":{"self":[{"href":"https:\/\/www.luckyshi.cn\/index.php?rest_route=\/wp\/v2\/posts\/705","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.luckyshi.cn\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.luckyshi.cn\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.luckyshi.cn\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.luckyshi.cn\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=705"}],"version-history":[{"count":2,"href":"https:\/\/www.luckyshi.cn\/index.php?rest_route=\/wp\/v2\/posts\/705\/revisions"}],"predecessor-version":[{"id":721,"href":"https:\/\/www.luckyshi.cn\/index.php?rest_route=\/wp\/v2\/posts\/705\/revisions\/721"}],"wp:attachment":[{"href":"https:\/\/www.luckyshi.cn\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=705"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.luckyshi.cn\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=705"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.luckyshi.cn\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=705"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}