Differences

This shows you the differences between two versions of the page.

--- slurm [2023/01/11 16:13]
krubinski [gpu-ms]
+++ slurm [2024/01/09 19:54] (current)
popel
@@ Line 2: / Line 2: @@
 LRC (Linguistic Research Cluster) is the name of ÚFAL's computational grid/cluster. The cluster is built on top of [[https://slurm.schedmd.com/|SLURM]] and is using [[https://www.lustre.org/|Lustre]] for [[internal:linux-network#directory-structure|data storage]].
+See Milan Straka's intro to Slurm (and Spark and possibly also the [[https://ufal.mff.cuni.cz/courses/npfl118#assignments|NPFL118 assingments]] if you want). Use the username=ufal and small linguistic password:
+  * https://lectures.ms.mff.cuni.cz/video/rec/npfl118/2324/npfl118-2324-winter-slurm.mp4
+  * https://lectures.ms.mff.cuni.cz/video/rec/npfl118/2324/npfl118-2324-winter-spark.mp4
+  * https://lectures.ms.mff.cuni.cz/video/rec/npfl118/2324/npfl118-2324-winter-assignments.mp4
 Currently there are following partitions (queues) available for computing:
@@ Line 12: / Line 18: @@
 | Node name | Thread count | Socket:Core:Thread | RAM (MB) |
-| achilles1 | 32 | 2:8:2 | 128810 |
+| achilles[1-8] | 32 | 2:8:2 | 128810 |
-| achilles2 | 32 | 2:8:2 | 128810 |
+| hector[1-8] | 32 | 2:8:2 | 128810 |
-| achilles3 | 32 | 2:8:2 | 128810 |
+| helena[1-8] | 32 | 2:8:2 | 128811 |
-| achilles4 | 32 | 2:8:2 | 128810 |
+| paris[1-8] | 32 | 2:8:2 | 128810 |
-| achilles5 | 32 | 2:8:2 | 128810 |
+| hyperion[2-8] | 64 | 2:16:2 | 257667 |
-| achilles6 | 32 | 2:8:2 | 128810 |
-| achilles7 | 32 | 2:8:2 | 128810 |
-| achilles8 | 32 | 2:8:2 | 128810 |
-| hector1 | 32 | 2:8:2 | 128810 |
-| hector2 | 32 | 2:8:2 | 128810 |
-| hector3 | 32 | 2:8:2 | 128810 |
-| hector4 | 32 | 2:8:2 | 128810 |
-| hector5 | 32 | 2:8:2 | 128810 |
-| hector6 | 32 | 2:8:2 | 128810 |
-| hector7 | 32 | 2:8:2 | 128810 |
-| hector8 | 32 | 2:8:2 | 128810 |
-| helena1 | 32 | 2:8:2 | 128811 |
-| helena2 | 32 | 2:8:2 | 128811 |
-| helena3 | 32 | 2:8:2 | 128811 |
-| helena4 | 32 | 2:8:2 | 128811 |
-| helena5 | 32 | 2:8:2 | 128810 |
-| helena6 | 32 | 2:8:2 | 128811 |
-| helena7 | 32 | 2:8:2 | 128810 |
-| helena8 | 32 | 2:8:2 | 128811 |
-| paris1 | 32 | 2:8:2 | 128810 |
-| paris2 | 32 | 2:8:2 | 128810 |
-| paris3 | 32 | 2:8:2 | 128810 |
-| paris4 | 32 | 2:8:2 | 128810 |
-| paris5 | 32 | 2:8:2 | 128810 |
-| paris6 | 32 | 2:8:2 | 128810 |
-| paris7 | 32 | 2:8:2 | 128810 |
-| paris8 | 32 | 2:8:2 | 128810 |
-| hyperion2 | 64 | 2:16:2 | 257667 |
-| hyperion3 | 64 | 2:16:2 | 257667 |
-| hyperion4 | 64 | 2:16:2 | 257667 |
-| hyperion5 | 64 | 2:16:2 | 257667 |
-| hyperion6 | 64 | 2:16:2 | 257667 |
-| hyperion7 | 64 | 2:16:2 | 257667 |
-| hyperion8 | 64 | 2:16:2 | 257667 |
 ==== cpu-ms ====
 | Node name | Thread count | Socket:Core:Thread | RAM (MB) |
 | iridium | 16 | 2:4:2 | 515977 |
-| orion1 | 40 | 2:10:2 | 128799 |
+| orion[1-8] | 40 | 2:10:2 | 128799 |
-| orion2 | 40 | 2:10:2 | 128799 |
-| orion3 | 40 | 2:10:2 | 128799 |
-| orion4 | 40 | 2:10:2 | 128799 |
-| orion5 | 40 | 2:10:2 | 128799 |
-| orion6 | 40 | 2:10:2 | 128799 |
-| orion7 | 40 | 2:10:2 | 128799 |
-| orion8 | 40 | 2:10:2 | 128799 |
 ==== gpu-troja ====
 | Node name | Thread count | Socket:Core:Thread | RAM (MB) | Features | GPU type |
-| tdll-3gpu1 | 64 | 2:16:2 | 128642 | gpuram48G gpu_cc8.6 | NVIDIA A40 |
+| tdll-3gpu[1-4] | 64 | 2:16:2 | 128642 | gpuram48G gpu_cc8.6 | NVIDIA A40 |
-| tdll-3gpu2 | 64 | 2:16:2 | 128642 | gpuram48G gpu_cc8.6 | NVIDIA A40 |
+| tdll-8gpu[1,2] | 64 | 2:16:2 | 257666 | gpuram40G gpu_cc8.0 | NVIDIA A100 |
-| tdll-3gpu3 | 64 | 2:16:2 | 128642 | gpuram48G gpu_cc8.6 | NVIDIA A40 |
+| tdll-8gpu[3-7] | 32 | 2:8:2 | 253725 | gpuram16G gpu_cc7.5 | NVIDIA Quadro P5000 |
-| tdll-3gpu4 | 64 | 2:16:2 | 128642 | gpuram48G gpu_cc8.6 | NVIDIA A40 |
-| tdll-8gpu1 | 64 | 2:16:2 | 257666 | gpuram40G gpu_cc8.0 | NVIDIA A100 |
-| tdll-8gpu2 | 64 | 2:16:2 | 257666 | gpuram40G gpu_cc8.0 | NVIDIA A100 |
-| tdll-8gpu3 | 32 | 2:8:2 | 253725 | gpuram16G gpu_cc7.5 | NVIDIA Quadro P5000 |
-| tdll-8gpu4 | 32 | 2:8:2 | 253725 | gpuram16G gpu_cc7.5 | NVIDIA Quadro P5000 |
-| tdll-8gpu5 | 32 | 2:8:2 | 253725 | gpuram16G gpu_cc7.5 | NVIDIA Quadro P5000 |
-| tdll-8gpu6 | 32 | 2:8:2 | 253725 | gpuram16G gpu_cc7.5 | NVIDIA Quadro P5000 |
-| tdll-8gpu7 | 32 | 2:8:2 | 253725 | gpuram16G gpu_cc7.5 | NVIDIA Quadro P5000 |
 ==== gpu-ms ====
 | Node name | Thread count | Socket:Core:Thread | RAM (MB) | Features | GPU type |
-| dll-3gpu1 | 64 | 2:16:2 | 128642 | gpuram48G gpu_cc8.6 | NVIDIA A40 |
+| dll-3gpu[1-5] | 64 | 2:16:2 | 128642 | gpuram48G gpu_cc8.6 | NVIDIA A40 |
-| dll-3gpu2 | 64 | 2:16:2 | 128642 | gpuram48G gpu_cc8.6 | NVIDIA A40 |
+| dll-4gpu[1,2] | 40 | 2:10:2 | 187978 | gpuram24G gpu_cc8.6 | NVIDIA RTX 3090 |
-| dll-3gpu3 | 64 | 2:16:2 | 128642 | gpuram48G gpu_cc8.6 | NVIDIA A40 |
+| dll-4gpu3 | 62 | 1:32:2 | 515652 | gpuram48G gpu_cc8.9 | NVIDIA L40 |
-| dll-3gpu4 | 64 | 2:16:2 | 128642 | gpuram48G gpu_cc8.6 | NVIDIA A40 |
+| dll-4gpu4 | 30 | 1:16:2 | 257616 | gpuram48G gpu_cc8.6 | NVIDIA A40 |
-| dll-3gpu5 | 64 | 2:16:2 | 128642 | gpuram48G gpu_cc8.6 | NVIDIA A40 |
+| dll-8gpu[1,2] | 64 | 2:16:2 | 515838 | gpuram24G gpu_cc8.0 | NVIDIA A30 |
-| dll-4gpu1 | 40 | 2:10:2 | 187978 | gpuram24G gpu_cc8.6 | NVIDIA RTX 3090 |
+| dll-8gpu[3,4] | 32 | 2:8:2 | 257830 | gpuram16G gpu_cc8.6 | NVIDIA RTX A4000 |
-| dll-4gpu2 | 40 | 2:10:2 | 187978 | gpuram24G gpu_cc8.6 | NVIDIA RTX 3090 |
+| dll-8gpu[5,6] | 40 | 2:10:2 | 385595 | gpuram16G gpu_cc7.5 | NVIDIA Quadro RTX 5000 |
-| dll-8gpu1 | 64 | 2:16:2 | 515838 | gpuram24G gpu_cc8.0 | NVIDIA A30 |
-| dll-8gpu2 | 64 | 2:16:2 | 515838 | gpuram24G gpu_cc8.0 | NVIDIA A30 |
-| dll-8gpu3 | 32 | 2:8:2 | 257830 | gpuram16G gpu_cc8.6 | NVIDIA RTX A4000 |
-| dll-8gpu4 | 32 | 2:8:2 | 253721 | gpuram16G gpu_cc8.6 | NVIDIA RTX A4000 |
-| dll-8gpu5 | 40 | 2:10:2 | 385595 | gpuram16G gpu_cc7.5 | NVIDIA Quadro RTX 5000 |
-| dll-8gpu6 | 40 | 2:10:2 | 385595 | gpuram16G gpu_cc7.5 | NVIDIA Quadro RTX 5000 |
 | dll-10gpu1 | 32 | 2:8:2 | 257830 | gpuram16G gpu_cc8.6 | NVIDIA RTX A4000 |
-| dll-10gpu2 | 32 | 2:8:2 | 257830 | gpuram11G gpu_cc6.1 | NVIDIA GeForce GTX 1080 Ti |
+| dll-10gpu[2,3] | 32 | 2:8:2 | 257830 | gpuram11G gpu_cc6.1 | NVIDIA GeForce GTX 1080 Ti |
-| dll-10gpu3 | 32 | 2:8:2 | 257830 | gpuram11G gpu_cc6.1 | NVIDIA GeForce GTX 1080 Ti |
@@ Line 123: / Line 73: @@
 #!/bin/bash
 #SBATCH -J helloWorld					  # name of job
-#SBATCH -p cpu-troja					  # name of partition or queue (if not specified default partition is used)
+#SBATCH -p cpu-troja					  # name of partition or queue (default=cpu-troja)
 #SBATCH -o helloWorld.out				  # name of output file for this submission script
 #SBATCH -e helloWorld.err				  # name of error file for this submission script
@@ Line 138: / Line 88: @@
 #SBATCH -D /some/path/                        # change directory before executing the job
 #SBATCH -N 2                                  # number of nodes (default 1)
-#SBATCH --nodelist=node1,node2...             # required node, or comma separated list of required nodes
+#SBATCH --nodelist=node1,node2...             # execute on *all* the specified nodes (and possibly more)
 #SBATCH --cpus-per-task=4                     # number of cores/threads per task (default 1)
 #SBATCH --gres=gpu:1                          # number of GPUs to request (default 0)
@@ Line 159: / Line 109: @@
 </code>
-==== Running jobs ====
+=== Rudolf's template ===
+The main point is for log files to have the job name and job id in them automatically.
+<code>
+#SBATCH -J RuRjob
+#SBATCH -o %x.%j.out
+#SBATCH -e %x.%j.err
+#SBATCH -p gpu-troja
+#SBATCH --gres=gpu:1
+#SBATCH --mem=16G
+#SBATCH --constraint="gpuram16G|gpuram24G"
+# Print each command to STDERR before executing (expanded), prefixed by "+ "
+set -o xtrace
+</code>
+==== Inspecting jobs ====
 In order to inspect all running jobs on the cluster use:
@@ Line 240: / Line 207: @@
 In the example above you can see comments at all lines relevant to CPU allocation.
+=== Priority ====
+When running srun or sbatch, you can pass ''-q high/normal/low/preempt-low''. These represent priorities 300/200/100/100, with ''normal'' (200) being the default. Furthermore, the ''preempt-low'' QOS is actually preemptible -- if there is a job with normal or high QOS, they can interrupt your ''preempt-low'' job.
+The preemption has probably not been used by anyone yet; some documentation about it is on https://slurm.schedmd.com/preempt.html, we use the REQUEUE regime (so your job is killed, very likely with some signal, so you could monitor it and for example save a checkpoint; but currently I do not know any details), and then started again when there are resources.
 ==== Interactive mode ====
@@ Line 256: / Line 226: @@
   * ''-p cpu-troja'' explicitly requires partition ''cpu-troja''. If not specified slurm will use default partition.
-  * ''--mem=64G'' requires 64G of memory for the job
+  * ''-''''-mem=64G'' requires 64G of memory for the job
 **To get interactive job with a single GPU of any kind:**
 <code>srun -p gpu-troja,gpu-ms --gres=gpu:1 --pty bash</code>
   * ''-p gpu-troja,gpu-ms'' require only nodes from these two partitions
-  * ''--gres=gpu:1'' requires 1 GPUs
+  * ''-''''-gres=gpu:1'' requires 1 GPUs
 <code>srun -p gpu-troja,gpu-ms --nodelist=tdll-3gpu1 --mem=64G --gres=gpu:2 --pty bash</code>
   * ''-p gpu-troja,gpu-ms'' require only nodes from these two partitions
-  * ''--nodelist=tdll-3gpu1'' explicitly requires one specific node
+  * ''-''''-nodelist=tdll-3gpu1'' explicitly requires one specific node
-  * ''--gres=gpu:2'' requires 2 GPUs
+  * Note that e.g. ''-''''-nodelist=tdll-3gpu[1-4]'' would execute 4 jobs on **all** the four machines ''tdll-3gpu[1-4]''. The documentation says "The job will contain all of these hosts and possibly additional hosts as needed to satisfy resource requirements." I am not aware of any [[https://stackoverflow.com/a/37555321/3310232|simple way]] how to specify that **any** of the listed nodes can be used, i.e. an equivalent of SGE ''-q '*@hector[14]'''.
+  * ''-''''-gres=gpu:2'' requires 2 GPUs
 <code>srun -p gpu-troja --constraint="gpuram48G|gpuram40G" --mem=64G --gres=gpu:2 --pty bash</code>
-  * ''--constraint="gpuram48G|gpuram40G"'' only consider nodes that have either ''gpuram48G'' or ''gpuram40G'' feature defined
+  * ''-''''-constraint="gpuram48G|gpuram40G"'' only consider nodes that have either ''gpuram48G'' or ''gpuram40G'' feature defined
+\\
+**Unexpected Behavior of ''srun -c1''**
+When you execute a command using ''srun'' and pass ''-c1'' like
+<code>srun -c1 date</code>
+then the command is actually executed **twice in parallel**. To avoid it, you have to either **remove the ''-c1''** or also **add explicit ''-n1''.**
 ==== Delete Job ====
 <code>scancel <job_id> </code>
+<code>scancel -n <job_name> </code>
 To see all the available options type:
-<code>man srun</code>
+<code>man scancel</code>
+==== Basic commands on cluster machines ====
+  lspci
+    # is any such hardware there?
+  nvidia-smi
+    # more details, incl. running processes on the GPU
+    # nvidia-* are typically located in /usr/bin
+  watch nvidia-smi
+    # For monitoring GPU activity in a separate terminal (thanks to Jindrich Libovicky for this!)
+    # You can also use nvidia-smi -l TIME
+  nvcc --version
+    # this should tell CUDA version
+    # nvcc is typically installed in /usr/local/cuda/bin/
+  theano-test
+    # dela to vubec neco uzitecneho? :-)
+    # theano-* are typically located in /usr/local/bin/
+  /usr/local/cuda/samples/1_Utilities/deviceQuery/deviceQuery
+    # shows CUDA capability etc.
+  ssh dll1; ~popel/bin/gpu_allocations
+    # who occupies which card on a given machine
 ===== See also =====

[ Back to the navigation ] [ Back to the content ]

Institute of Formal and Applied Linguistics Wiki

Differences