简体   繁体   English

使用 srun:如何设置节点数和核心数

[英]Using srun : how to set number of nodes and cores

I am new to using commands like srun.我不熟悉使用像 srun 这样的命令。 I want to launch a parallel job using 1 node 1cpu and 16 cores, but i dont know what i should indicate when using srun function. (1 node is composed of 2 CPUs with 18 cores on each CPU)我想使用 1 个节点 1cpu 和 16 个内核启动并行作业,但我不知道在使用 srun function 时应该指出什么。(1 个节点由 2 个 CPU 组成,每个 CPU 有 18 个内核)

Does anyone knows which parameters i should use?有谁知道我应该使用哪些参数?

To set minimums on the cpu resources lower than those available per node, you need to provide --cpus-per-task in your srun command (see https://slurm.schedmd.com/srun.html#OPT_cpus-per-task ).要将 cpu 资源的最小值设置为低于每个节点的可用资源,您需要在srun命令中提供--cpus-per-task (请参阅https://slurm.schedmd.com/srun.html#OPT_cpus-per-task ).

If you want to explicitly reserve the other 2 cores per cpu for system use, see https://slurm.schedmd.com/mc_support.html#srun_hints for usage of --cpu-bind or high level resource limits.如果您想明确保留每个 cpu 的其他 2 个内核供系统使用,请参阅https://slurm.schedmd.com/mc_support.html#srun_hints以了解--cpu-bind或高级资源限制的使用。

声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.

相关问题 如何在此 Fortran 代码中使用 OpenMP 将数字 lambdaeff 正确传递给内核? - How to correctly pass the number lambdaeff to the cores using OpenMP in this Fortran code? Caret in R:设置allowParallel的核心数量? - Caret in R: Set number of cores for allowParallel? doMPI 和节点、处理器和内核 - doMPI and nodes, processors and cores OpenMP如何确定系统中的内核数? - How does OpenMP determine the Number of cores in the system? 如果我保持核心总数一致,我应该如何选择执行器的数量和每个执行器的核心数? - If I keep the total number of cores consistent, how should I choose the number of executors and number of cores per executor? 是否可以对给定代码的 Cuda 编程中使用的内核数量设置限制? - Is it possible to set a limit on the number of cores to be used in Cuda programming for a given code? SLURM 节点、任务、核心和 CPU - SLURM nodes, tasks, cores, and cpus XGBoost 不使用 Windows 下可用的最大内核数? - XGBoost not using max number of cores available under Windows? 使用foreach和控制核心数并行运行多个R脚本 - Run multiple R scripts in parallel using foreach and controlling number of cores 如何使用一个或多个特定内核执行应用程序? - How to execute an application using a specific core or cores?
 
粤ICP备18138465号  © 2020-2024 STACKOOM.COM