[ Skip to the content ]

Institute of Formal and Applied Linguistics Wiki

[ Back to the navigation ]

You are here: start » gpu

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
Next revision Both sides next revision
gpu [2017/01/05 12:18]
kocmanek 		gpu [2017/07/15 18:03]
kocmanek [Performance tests]
Line 5: Line 5:
 ===== Servers with GPU units ===== ===== Servers with GPU units =====
  
-| machine | GPU; [[https://en.wikipedia.org/wiki/CUDA#Supported_GPUs|Capability]] [cc]  | cores | GPU RAM | Comment | +| machine                    | GPU; [[https://en.wikipedia.org/wiki/CUDA#Supported_GPUs|Capability]] [cc]  | cores | GPU RAM | Comment | 
-| titan-gpu  | GeForce GTX Titan Z; cc3.5 | 2 | 6 GB each core |        | +| titan                      | GeForce GTX 1080 Ti; cc6.1 | 1  | 11 GB           |  | 
-| twister1   | Tesla K40c; cc3.5          | 1 | 12 GB          |        | +| titan-gpu                  | GeForce GTX Titan Z; cc3.5 | 2  | 6 GB each core  |  | 
-| twister2   | Tesla K40c; cc3.5          | 1 | 12 GB          |        | +| twister1; twister2; kronos | Tesla K40c; cc3.5          | 1  | 12 GB           |  | 
-| kronos-dev | Tesla K40c; cc3.5          | 1 | 12 GB          |        | +| iridium                    | Quadro K2000; cc3.0        | 1  | 2 GB            |  | 
-| iridium    | Quadro K2000; cc3.0        | 1 | 2 GB           |        | +| victoria; arc              | GeForce GT 630; cc3.0      | 1  | 2 GB            | desktop machine | 
-| victoria   | GeForce GT 630; cc3.0      | 1 | 2 GB           | Ondrej Bojar's desktop machine | +| athena                     | GeForce GTX 1080; cc6.1    | 1  | 8 GB            | Tom's desktop machine | 
-| arc        | GeForce GT 630; cc3.0      | 1 | 2 GB           | Ales's desktop machine | +| dll1; dll2                 | GeForce GTX 1080; cc6.1    | 8  | 8 GB each core  |  | 
-| athena     | GeForce GTX 1080; cc6.1    | 1 | 8 GB           | Tom's desktop machine | +| dll3; dll4; dll5           | GeForce GTX 1080 Ti; cc6.1 | 10 | 11 GB each core |  |
-| dll1     | GeForce GTX 1080; cc6.1    | 2 | 8 GB each core |  | +
-| dll2     | GeForce GTX 1080; cc6.1    | 2 | 8 GB each core |  |+
  
 not used at the moment: GeForce GTX 570 (from twister2) not used at the moment: GeForce GTX 570 (from twister2)
Line 30: Line 28:
 === Individual acquisitions: NVIDIA Academic Hardware Grants == === Individual acquisitions: NVIDIA Academic Hardware Grants ==
  
-There is an easy way to get one high-end GPU: [[https://developer.nvidia.com/academic_gpu_seeding|ask NVIDIA for an Academic Hardware Grant]]. All it takes is writing a short grant application (at most ~2 hrs of work from scratch; if you have a GAUK, ~15 minutes of copy-pasting). Due to the GPU housing issues (mainly rack space and cooling), Milan F. said we should request the Tesla-line cards. If you want to have a look at an application, feel free to ask at hajicj@ufal.mff.cuni.cz :)+There is an easy way to get one high-end GPU: [[https://developer.nvidia.com/academic_gpu_seeding|ask NVIDIA for an Academic Hardware Grant]]. All it takes is writing a short grant application (at most ~2 hrs of work from scratch; if you have a GAUK, ~15 minutes of copy-pasting). Due to the GPU housing issues (mainly rack space and cooling), Milan F. said we should request the Tesla-line cards (2017 check with Milan about this issue). If you want to have a look at an application, feel free to ask at hajicj@ufal.mff.cuni.cz :)
  
 Take care, however, to coordinate the grant applications a little, so that not too many arrive from UFAL within a short time: these grants are explicitly //not// intended to build GPU clusters, they are "seeding" grants meant for researchers to try out GPUs (and fall in love with them, and buy a cluster later). If you are planning to submit the hardware grant, have submitted one, or have already been awarded one, please add yourself here. Take care, however, to coordinate the grant applications a little, so that not too many arrive from UFAL within a short time: these grants are explicitly //not// intended to build GPU clusters, they are "seeding" grants meant for researchers to try out GPUs (and fall in love with them, and buy a cluster later). If you are planning to submit the hardware grant, have submitted one, or have already been awarded one, please add yourself here.
Line 38: Line 36:
   * Ondřej Plátek - granted (2015)   * Ondřej Plátek - granted (2015)
   * Jan Hajič jr. - granted (early 2016)   * Jan Hajič jr. - granted (early 2016)
-  * Jindra Helcl - planning to apply (fall 2016) 
  
  
      
  
-===== Performance tests =====+===== How to use cluster =====
  
-* [[http://www.trustedreviews.com/nvidia-geforce-gtx-1080-review-performance-benchmarks-and-conclusion-page-2| 980 vs 1080 vs Titan X (not the Titan Z we have)]]+In this section will be explained how to use cluster properly.  
 +==== TensorFlow Environment ====
  
-In the following table is the experiment conducted by Tom Kocmi. You can replicate experiment: /a/merkur3/kocmanek/GPUBenchmark (you will need to prepare environment of TensorFlow11 or use my ANACONDA)+Majority people at UFAL use TensorFlow. To start using it you need to create python virtual environment (virtualenv or use Anaconda for it). Into the environment you must place TensorFlow. The TF is either in CPU or GPU version.
  
-I am preparing department-wide benchmark, but meanwhile the results for different experiment: +  pip install tensorflow 
- * Athena (GTX 1080) - 2 hodiny 32 minut +  pip install tensorflow-gpu 
- * Twister (Tesla K40c) - 6 hodin 46 minut+   
 +You can use prepared environment by adding into your ~/.bashrc
  
-| machine | Setup; CPU/GPU; [[https://en.wikipedia.org/wiki/CUDA#Supported_GPUs|Capability]] [cc] | Walltime | Note | +  export PATH=/a/merkur3/kocmanek/ANACONDA/bin:$PATH
-| athena     | GeForce GTX 1080; cc6.1            |  9:55:58 | Tom's desktop  | +
-| dll2       | (2 GPU) GeForce GTX 1080; cc6.1    |  10:19:40 | with CUDA_VISIBILITY=0 | +
-| dll1       | (2 GPU) GeForce GTX 1080; cc6.1    |  12:34:34 | Probably only one GPU was used | +
-| dll2       | (2 GPU) GeForce GTX 1080; cc6.1    |  13:01:05 | Only one GPU was used | +
-| titan-gpu  | (2 GPU) GeForce GTX Titan Z; cc3.5 |  16:05:24 | Probably only one GPU was used | +
-| kronos-dev | Tesla K40c; cc3.5                  |  22:41:01 |  | +
-| twister2   | Tesla K40c; cc3.5                  |  22:43:10 |  | +
-| twister1   | Tesla K40c; cc3.5                  |  24:19:45 |  | +
-| helena1    | 16x CPU                            |  46:33:14 |  | +
-| belzebub   | 16x CPU                            |  52:36:56 |  | +
-| iridium    | Quadro K2000; cc3.0                |  59:47:58 |  | +
-| helena7     | 8x CPU                            |  61:10:00 | approximated (still running) | +
-| arc        | GeForce GT 630; cc3.0              |  103:42:30 | (approximated after 66 hours) | +
-| lucifer4   | 8x CPU                             |  134:41:22 |  | +
-| victoria   | GeForce GT 630; cc3.0              |  --- | never run, same GPU as Arc |+
  
 +And then you can activate your environment:
  
-A comparison with Ondrej's small data set: +  source activate tf1 
-  * dll2 (2xGPU) takes 13m for one reporting period +  source activate tf1cpu
-  * achilles2 (4xCPU with 8 CPUs reserved) takes 24m for one reporting period+
  
 +This environment have TensorFlow 1.0 and all necessary requirements for NeuralMonkey.
  
-===== Installed toolkits =====+==== Using cluster ====
  
-//This should mention where each interesting toolkit lives (on a particular machine).//+Rule number one, always use the GPU queue (never log in machine by ssh). Always use qsub or qsubmit with proper arguments.
  
-==== TensorFlow ====+For testing and using the cluster interactively you can use qrsh (this should not be used for long running experiments since the console is not closed on the end of the experiment). Following command will assign you a GPU and creates interactive console.
  
-[[https://redmine.ms.mff.cuni.cz/projects/mmmt/repository/revisions/6a064187fc6959db9b77cf2d5350c5f4918a8067/entry/prepare_env.sh|This script]] installs TensorFlow 0.7.1 (and all other dependencies we need for Multimodal Translation) into `tf' and `tf-gpu' virtual environments. The GPU environment can be loaded by calling <code>source tf-gpu/bin/activate-gpu</code>+  qrsh -q gpu.q -l gpu=1,gpu_ram=2G -pty yes bash 
 +   
 +For running experiments you must use qsub command:
  
-OP: I created [[https://gist.github.com/oplatek/323b63b8f116cd3d78c0f492f78cc289|script]] which install Tensorflow 0.8 and test it if it uses GPU. TF is installed into `user` or `global` installation either for `python3.4` or `python2.7`+  qsub -q gpu.q -l gpu=1,gpu_cc_min3.5=1,gpu_ram=2G WHAT_SHOULD_BE_RUN 
 +   
 +Cleaner way to use cluster is with /home/bojar/tools/shell/qsubmit
  
-=== Select GPU device === +  qsubmit --gpumem=2G --queue="gpu.q" WHAT_SHOULD_BE_RUN 
- +   
-Use variable CUDA_VISIBLE_DEVICES to constrain tensorflow to compute only on the selected one. For the use of first GPU use: +It is recommended to use priority -100 if you are not rushing for the results and don't need to leap over your colleagues jobs. 
-<code>export CUDA_VISIBLE_DEVICES=0</code> +==== Basic commands ====
- +
-To list available devices, use: +
-<code>/opt/cuda/samples/1_Utilities/deviceQuery/deviceQuery | grep ^Device</code> +
- +
-===== Basic commands =====+
  
   lspci   lspci
Line 111: Line 95:
   /usr/local/cuda/samples/1_Utilities/deviceQuery/deviceQuery   /usr/local/cuda/samples/1_Utilities/deviceQuery/deviceQuery
     # shows CUDA capability etc.     # shows CUDA capability etc.
 +    
 +=== Select GPU device ===
 +
 +Use variable CUDA_VISIBLE_DEVICES to constrain tensorflow to compute only on the selected one. For the use of first GPU use (GPU queue do this for you):
 +  export CUDA_VISIBLE_DEVICES=0
 +
 +To list available devices, use:
 +  /opt/cuda/samples/1_Utilities/deviceQuery/deviceQuery | grep ^Device
 +
 +===== Performance tests =====
 +
 +* [[http://www.trustedreviews.com/nvidia-geforce-gtx-1080-review-performance-benchmarks-and-conclusion-page-2| 980 vs 1080 vs Titan X (not the Titan Z we have)]]
 +
 +In the following table is the experiment conducted by Tom Kocmi. You can replicate experiment: /a/merkur3/kocmanek/Projects/GPUBenchmark (you will need to prepare environment of TensorFlow11 or use my ANACONDA). The benchmark uses 2GB model.
 +
 +| machine | Setup; CPU/GPU; [[https://en.wikipedia.org/wiki/CUDA#Supported_GPUs|Capability]] [cc] | Walltime | Note |
 +| athena     | GeForce GTX 1080; cc6.1            |    9:55:58 | Tom's desktop  |
 +| dll2       | (2 GPU) GeForce GTX 1080; cc6.1    |   10:19:40 | with CUDA_VISIBLE_DEVICES=0 |
 +| titan      | GeForce GTX 1080 Ti                |   10:45:11 | (new result with correct CUDA version) |
 +| dll1       | (2 GPU) GeForce GTX 1080; cc6.1    |   12:34:34 | Probably only one GPU was used |
 +| dll2       | (2 GPU) GeForce GTX 1080; cc6.1    |   13:01:05 | Only one GPU was used |
 +| titan-gpu  | (2 GPU) GeForce GTX Titan Z; cc3.5 |   16:05:24 | Probably only one GPU was used |
 +| kronos-dev | Tesla K40c; cc3.5                  |   22:41:01 |  |
 +| twister2   | Tesla K40c; cc3.5                  |   22:43:10 |  |
 +| twister1   | Tesla K40c; cc3.5                  |   24:19:45 |  |
 +| helena1    | 16x cores CPU                      |   46:33:14 |  |
 +| belzebub   | 16x cores CPU                      |   52:36:56 |  |
 +| iridium    | Quadro K2000; cc3.0                |   59:47:58 |  |
 +| helena7    | 8x cores CPU                       |   60:39:17 |  |
 +| arc        | GeForce GT 630; cc3.0              |  103:42:30 | (approximated after 66 hours) |
 +| lucifer4   | 8x cores CPU                       |  134:41:22 |  |
 +| victoria   | GeForce GT 630; cc3.0              |        --- | never run, same GPU as Arc |
 +
 +
 +=== Better Benchmark ===
 +
 +The previous benchmark only compares the speed of processing units within the GPUs and do not take into account the size of memory. Therefore I have conducted another benchmark, this time for each graphic card increased the batch_size as much as possible so the model still could fit into the GPU. This way the results should be more representative of the power for each GPU.
 +
 +| GPU                        | GPU RAM | Walltime | Batch size | Machine |
 +| Tesla K40c; cc3.5          | 12 GB   |          |            | |
 +| GeForce GTX 1080 Ti; cc6.1 | 11 GB   |          |            | |
 +| GeForce GTX 1080; cc6.1    | 8 GB    |          |            | |
 +|GeForce GTX 1080; cc6.1    | 8 GB    |          |            | Athena (without virtualization) |
 +| GeForce GTX Titan Z; cc3.5 | 6 GB    |          |            | |
  
 ===== Links ===== ===== Links =====

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