[ Skip to the content ]

Institute of Formal and Applied Linguistics Wiki


[ Back to the navigation ]

Differences

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

Link to this comparison view

Both sides previous revision Previous revision
Next revision
Previous revision
gpu [2018/06/21 10:04]
vodrazka [Servers with GPU units]
gpu [2019/04/17 15:13]
ufal [Servers with GPU units]
Line 5: Line 5:
 ===== Servers with GPU units ===== ===== Servers with GPU units =====
 GPU cluster ''​gpu-ms.q''​ at Malá Strana: GPU cluster ''​gpu-ms.q''​ at Malá Strana:
-| machine | GPU type | GPU driver version | [[https://​en.wikipedia.org/​wiki/​CUDA#​GPUs_supported|cc]] | GPU cnt | GPU RAM (GB) | machine ​ + 
-RAM (GB)| +| machine | GPU type | GPU driver version | [[https://​en.wikipedia.org/​wiki/​CUDA#​GPUs_supported|cc]] | GPU cnt | GPU RAM (GB) | machine RAM (GB)| Comment ​
-| dll1 |  GeForce GTX 1080 |  396.24 |  6.1 |  8 |  8 |  249 | +| dll1 |  GeForce GTX 1080 |  396.24 |  6.1 |  8 |  8 |  249 |  moved to AIC|                                                                      
-| dll2 |  GeForce GTX 1080 |  396.24 |  6.1 |  8 |  8 |  249 | +| dll2 |  GeForce GTX 1080 |  396.24 |  6.1 |  8 |  8 |  249 |                                                                         ​ 
-| dll3 |  GeForce GTX 1080 Ti |  396.24 |  6.1 |  ​|  11 |  249 | +| dll3 |  GeForce GTX 1080 Ti |  396.24 |  6.1 |  ​10 |  11 |  249 |                                                                     
-| dll4 |  GeForce GTX 1080 Ti |  396.24 |  6.1 |  10 |  11 |  249 | +| dll4 |  GeForce GTX 1080 Ti |  396.24 |  6.1 |  10 |  11 |  249 |                                                                     
-| dll5 |  GeForce GTX 1080 Ti |  396.24 |  6.1 |  10 |  11 |  249 | +| dll5 |  GeForce GTX 1080 Ti |  396.24 |  6.1 |  10 |  11 |  249 |                                                                     
-| dll6 |  GeForce GTX 1080 Ti |  396.24 |  6.1 |  9 |  11 |  123 | +| dll6 |  GeForce GTX 1080 Ti |  396.24 |  6.1 |  9 |  11 |  123 |                                                                     ​ 
-kronos ​|  GeForce GTX 1080 Ti |  396.24 |  6.1 |  1 |  11 |  123 | +dll7 |  GeForce GTX 1080 Ti |  396.24 |  6.1 |  1 |  11 |  123 |                                                                     ​ 
-titan1 ​|  ​GeForce GTX 1080 |  ​396.24 |  ​6.1 |  1 |  8 |  ​30 +kronos ​|  ​Tesla K40c |  ​418.39 |  ​3.5 |  ​1 |  ​11 |  123 |                                                                             
-titan2 ​|  ​Tesla K40c |  ​396.24 |  ​3.|  ​|  ​11 |  ​30 + 
-twister1 ​|  ​Tesla K40c |  ​396.24 |  ​3.|  ​|  ​11 |  ​45 +GPU cluster ''​gpu-troja.q''​ at Troja: 
-twister2 ​|  ​Tesla K40c |  ​396.24 |  ​3.|  1 |  ​11 |  ​45 |+ 
 +| machine | GPU type | GPU driver version | [[https://​en.wikipedia.org/​wiki/​CUDA#​GPUs_supported|cc]] | GPU cnt | GPU RAM (GB) | machine RAM (GB)| 
 +| tdll1 |  Quadro P5000 |  410.48 |  6.1 |  8 |  ​17  245 |                                                                            
 +tdll2 |  ​Quadro P5000 |  ​410.48 |  ​6.|  ​|  ​17 |  ​249                                                                           ​ 
 +tdll3 |  ​Quadro P5000 |  ​410.48 |  ​6.|  ​|  ​17 |  ​249                                                                           ​ 
 +tdll4 |  ​Quadro P5000 |  ​410.48 |  ​6.|  ​8 |  17 |  245 |                                                                            
 +| tdll5 |  Quadro P5000 |  410.48 |  6.1 |  ​|  ​17 |  249                                                                                                                                                     ​
  
 Desktop machines: Desktop machines:
Line 24: Line 30:
 | athena ​                    | GeForce GTX 1080 | cc6.1 |  1 |  8 GB | Tom's desktop machine | | athena ​                    | GeForce GTX 1080 | cc6.1 |  1 |  8 GB | Tom's desktop machine |
  
-Not used at the moment: GeForce GTX 570 (from twister2) 
 Multiple versions of CUDA library are accessible on each machine together with cudnn. Theano and TensorFlow is supported. Multiple versions of CUDA library are accessible on each machine together with cudnn. Theano and TensorFlow is supported.
  
-TODO - update link: +[[http://​ufaladm2.ufal.hide.ms.mff.cuni.cz/​munin/​ufal.hide.ms.mff.cuni.cz/​lrc-master.ufal.hide.ms.mff.cuni.cz/​index.html#​dll|GPU usage rolling graphs]]
-[[https://​ufaladm2.ufal.hide.ms.mff.cuni.cz/​munin/​ufal.hide.ms.mff.cuni.cz/​lrc-headnode.ufal.hide.ms.mff.cuni.cz/​index.html#​dll|GPU usage rolling graphs]]+
  
  
 ===== Rules ===== ===== Rules =====
   * First, read [[internal:​Linux network]] and [[:Grid]].   * First, read [[internal:​Linux network]] and [[:Grid]].
-  * All the rules from [[:Grid]] apply, even more strictly than for CPU because there are too many GPU users and not as many GPUs available. So as a reminder: always use GPUs via ''​qsub''​ (or ''​qrsh''​),​ never via ''​ssh''​. You can ssh to any machine e.g. to run ''​nvidia-smi''​ or ''​htop'',​ but not to start computing on GPU. Don't forget to specify you RAM requirements with e.g. ''​-l mem_free=8G,​act_mem_free=8G,​h_vmem=12G''​. +  * All the rules from [[:Grid]] apply, even more strictly than for CPU because there are too many GPU users and not as many GPUs available. So as a reminder: always use GPUs via ''​qsub''​ (or ''​qrsh''​),​ never via ''​ssh''​. You can ssh to any machine e.g. to run ''​nvidia-smi''​ or ''​htop'',​ but not to start computing on GPU. Don't forget to specify you RAM requirements with e.g. ''​-l mem_free=8G,​act_mem_free=8G,​h_data=12G''​. 
-  * Always specify the number of GPU cards (e.g. ''​gpu=1''​),​ the minimal Cuda capability you need (e.g. ''​gpu_cc_min3.5=1''​) and your GPU memory requirements (e.g. ''​gpu_ram=2G''​). Thus e.g. <​code>​qsub -q gpu-ms.q -l gpu=1,​gpu_cc_min3.5=1,​gpu_ram=2G</​code>​ +    * **Note that you need to use ''​h_data''​ instead of ''​h_vmem''​ for GPU jobs.** CUDA driver allocates a lot of "​unused"​ virtual memory (tens of GB per card), which is counted in ''​h_vmem'',​ but not in ''​h_data''​. All usual allocations (''​malloc'',​ ''​new'',​ Python allocations) seem to be included in ''​h_data''​. 
-  * If you need more than one GPU card (on a single machine), always require as many CPU cores (''​-pe smp X''​) as many GPU cards you need. E.g. <​code>​qsub -q gpu-ms.q -l gpu=4,​gpu_cc_min3.5=1,​gpu_ram=7G -pe smp 4</​code> ​**Warning**:​ currently, this does not work, so you can omit the ''​-pe smp X''​ part. Milan Fučík is working on a fix. +  * Always specify the number of GPU cards (e.g. ''​gpu=1''​),​ the minimal Cuda capability you need (e.g. ''​gpu_cc_min3.5=1''​) and your GPU memory requirements (e.g. ''​gpu_ram=2G''​). Thus e.g. <​code>​qsub -q 'gpu*' ​-l gpu=1,​gpu_cc_min3.5=1,​gpu_ram=2G</​code>​ 
-  * For interactive jobs, you can use ''​qrsh'',​ but make sure to end your job as soon as you don't need the GPU (so don't use qrsh for long training). **Warning: ''​-pty yes bash''​ is necessary**,​ otherwise the variable ''​$CUDA_VISIBLE_DEVICES''​ will not be set correctly. E.g. <​code>​qrsh -q gpu-ms.q -l gpu=1,​gpu_ram=2G -pty yes bash</​code>​In general: don't reserve a GPU (as described above) without actually using it for longer time. (E.g. try separating steps which need GPU and steps which do not and execute those separately on our GPU resp. CPU cluster.) Ondřej Bojar has a script /​home/​bojar/​tools/​servers/​watch_gpus for watching reserved but unused GPU on most machines which will e-mail you, but don't rely on in only.+  * If you need more than one GPU card (on a single machine), always require as many CPU cores (''​-pe smp X''​) as many GPU cards you need. E.g. <​code>​qsub -q 'gpu*' ​-l gpu=4,​gpu_cc_min3.5=1,​gpu_ram=7G -pe smp 4</​code>​ 
 +  * For interactive jobs, you can use ''​qrsh'',​ but make sure to end your job as soon as you don't need the GPU (so don't use qrsh for long training). **Warning: ''​-pty yes bash -l''​ is necessary**,​ otherwise the variable ''​$CUDA_VISIBLE_DEVICES''​ will not be set correctly. E.g. <​code>​qrsh -q 'gpu*' ​-l gpu=1,​gpu_ram=2G -pty yes bash -l</​code>​In general: don't reserve a GPU (as described above) without actually using it for longer time. (E.g. try separating steps which need GPU and steps which do not and execute those separately on our GPU resp. CPU cluster.) Ondřej Bojar has a script /​home/​bojar/​tools/​servers/​watch_gpus for watching reserved but unused GPU on most machines which will e-mail you, but don't rely on it only.
   * Note that the dll machines have typically 10 cards, but "​just"​ 250 GB RAM (DLL6 has only 128 GB). So the expected (maximal) ''​mem_free''​ requirement for jobs with 1 GPU is 25GB. If your 1-GPU job takes e.g. 80 GB and you submit three such jobs on the same machine, you have effectively blocked the whole machine and seven GPUs remain unused. If you really need to submit more high-memory jobs, send each on a different machine.   * Note that the dll machines have typically 10 cards, but "​just"​ 250 GB RAM (DLL6 has only 128 GB). So the expected (maximal) ''​mem_free''​ requirement for jobs with 1 GPU is 25GB. If your 1-GPU job takes e.g. 80 GB and you submit three such jobs on the same machine, you have effectively blocked the whole machine and seven GPUs remain unused. If you really need to submit more high-memory jobs, send each on a different machine.
  
Line 44: Line 49:
  
 Multiple versions of ''​cuda''​ can be accessed in ''/​opt/​cuda''​. ​ Multiple versions of ''​cuda''​ can be accessed in ''/​opt/​cuda''​. ​
 +
 You need to set library path from your ''​~/​.bashrc'':​ You need to set library path from your ''​~/​.bashrc'':​
  
Line 94: Line 100:
   qsubmit --gpumem=2G --queue="​gpu-ms.q"​ WHAT_SHOULD_BE_RUN   qsubmit --gpumem=2G --queue="​gpu-ms.q"​ WHAT_SHOULD_BE_RUN
   ​   ​
-It is recommended to use priority lower than the default -100 if you are not rushing for the results and don't need to leap over your colleagues jobs.+It is recommended to use priority lower than the default -100 if you are not rushing for the results and don't need to leap over your colleagues jobs. Please, do not use priority between -99 to 0 for jobs taking longer than a few hours, unless it is absolutely necessary for your work.
 ==== Basic commands ==== ==== Basic commands ====
  

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