How can I increase CUDA load on a Tensorflow deep learning task after reducing batch size to fit the GPU?












0












$begingroup$


I am running this TensorFlow task for Swahili-to-English on an NVidia GeForce 1060 GPU with 6GB of VRAM:
https://colab.research.google.com/github/tensorflow/tensorflow/blob/master/tensorflow/contrib/eager/python/examples/nmt_with_attention/nmt_with_attention.ipynb



I have to reduce the batch size to get the example code not to run out of memory on the GPU. With trial and error, I find a batch size (minimum is 1) where the code runs and uses the most memory. I notice that, as I reduce the batch size, the CUDA core load as reported by Windows Task Manager GPU view goes down.



The application described in the link above creates a complex TensorFlow network. I don't know whether Tensorflow creates one copy of the network or multiple copies to load the GPU.



If it can create multiple copies, is there a TensorFlow switch for that? I don't think memory speed should be a bottleneck in feeding the GPU. That is, I should be able to optimize between batch size or number of jobs resident in the GPU, and number of compute networks in the GPU.



Is there an easy way in TensorFlow to assess the size in CUDA cores of a compute network?



What are the factors I can use to optimize CUDA load in a small GPU with a large deep learning task?










share|improve this question









$endgroup$

















    0












    $begingroup$


    I am running this TensorFlow task for Swahili-to-English on an NVidia GeForce 1060 GPU with 6GB of VRAM:
    https://colab.research.google.com/github/tensorflow/tensorflow/blob/master/tensorflow/contrib/eager/python/examples/nmt_with_attention/nmt_with_attention.ipynb



    I have to reduce the batch size to get the example code not to run out of memory on the GPU. With trial and error, I find a batch size (minimum is 1) where the code runs and uses the most memory. I notice that, as I reduce the batch size, the CUDA core load as reported by Windows Task Manager GPU view goes down.



    The application described in the link above creates a complex TensorFlow network. I don't know whether Tensorflow creates one copy of the network or multiple copies to load the GPU.



    If it can create multiple copies, is there a TensorFlow switch for that? I don't think memory speed should be a bottleneck in feeding the GPU. That is, I should be able to optimize between batch size or number of jobs resident in the GPU, and number of compute networks in the GPU.



    Is there an easy way in TensorFlow to assess the size in CUDA cores of a compute network?



    What are the factors I can use to optimize CUDA load in a small GPU with a large deep learning task?










    share|improve this question









    $endgroup$















      0












      0








      0





      $begingroup$


      I am running this TensorFlow task for Swahili-to-English on an NVidia GeForce 1060 GPU with 6GB of VRAM:
      https://colab.research.google.com/github/tensorflow/tensorflow/blob/master/tensorflow/contrib/eager/python/examples/nmt_with_attention/nmt_with_attention.ipynb



      I have to reduce the batch size to get the example code not to run out of memory on the GPU. With trial and error, I find a batch size (minimum is 1) where the code runs and uses the most memory. I notice that, as I reduce the batch size, the CUDA core load as reported by Windows Task Manager GPU view goes down.



      The application described in the link above creates a complex TensorFlow network. I don't know whether Tensorflow creates one copy of the network or multiple copies to load the GPU.



      If it can create multiple copies, is there a TensorFlow switch for that? I don't think memory speed should be a bottleneck in feeding the GPU. That is, I should be able to optimize between batch size or number of jobs resident in the GPU, and number of compute networks in the GPU.



      Is there an easy way in TensorFlow to assess the size in CUDA cores of a compute network?



      What are the factors I can use to optimize CUDA load in a small GPU with a large deep learning task?










      share|improve this question









      $endgroup$




      I am running this TensorFlow task for Swahili-to-English on an NVidia GeForce 1060 GPU with 6GB of VRAM:
      https://colab.research.google.com/github/tensorflow/tensorflow/blob/master/tensorflow/contrib/eager/python/examples/nmt_with_attention/nmt_with_attention.ipynb



      I have to reduce the batch size to get the example code not to run out of memory on the GPU. With trial and error, I find a batch size (minimum is 1) where the code runs and uses the most memory. I notice that, as I reduce the batch size, the CUDA core load as reported by Windows Task Manager GPU view goes down.



      The application described in the link above creates a complex TensorFlow network. I don't know whether Tensorflow creates one copy of the network or multiple copies to load the GPU.



      If it can create multiple copies, is there a TensorFlow switch for that? I don't think memory speed should be a bottleneck in feeding the GPU. That is, I should be able to optimize between batch size or number of jobs resident in the GPU, and number of compute networks in the GPU.



      Is there an easy way in TensorFlow to assess the size in CUDA cores of a compute network?



      What are the factors I can use to optimize CUDA load in a small GPU with a large deep learning task?







      machine-learning tensorflow machine-translation






      share|improve this question













      share|improve this question











      share|improve this question




      share|improve this question










      asked 16 mins ago









      Lars EricsonLars Ericson

      1334




      1334






















          0






          active

          oldest

          votes











          Your Answer





          StackExchange.ifUsing("editor", function () {
          return StackExchange.using("mathjaxEditing", function () {
          StackExchange.MarkdownEditor.creationCallbacks.add(function (editor, postfix) {
          StackExchange.mathjaxEditing.prepareWmdForMathJax(editor, postfix, [["$", "$"], ["\\(","\\)"]]);
          });
          });
          }, "mathjax-editing");

          StackExchange.ready(function() {
          var channelOptions = {
          tags: "".split(" "),
          id: "557"
          };
          initTagRenderer("".split(" "), "".split(" "), channelOptions);

          StackExchange.using("externalEditor", function() {
          // Have to fire editor after snippets, if snippets enabled
          if (StackExchange.settings.snippets.snippetsEnabled) {
          StackExchange.using("snippets", function() {
          createEditor();
          });
          }
          else {
          createEditor();
          }
          });

          function createEditor() {
          StackExchange.prepareEditor({
          heartbeatType: 'answer',
          autoActivateHeartbeat: false,
          convertImagesToLinks: false,
          noModals: true,
          showLowRepImageUploadWarning: true,
          reputationToPostImages: null,
          bindNavPrevention: true,
          postfix: "",
          imageUploader: {
          brandingHtml: "Powered by u003ca class="icon-imgur-white" href="https://imgur.com/"u003eu003c/au003e",
          contentPolicyHtml: "User contributions licensed under u003ca href="https://creativecommons.org/licenses/by-sa/3.0/"u003ecc by-sa 3.0 with attribution requiredu003c/au003e u003ca href="https://stackoverflow.com/legal/content-policy"u003e(content policy)u003c/au003e",
          allowUrls: true
          },
          onDemand: true,
          discardSelector: ".discard-answer"
          ,immediatelyShowMarkdownHelp:true
          });


          }
          });














          draft saved

          draft discarded


















          StackExchange.ready(
          function () {
          StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fdatascience.stackexchange.com%2fquestions%2f47918%2fhow-can-i-increase-cuda-load-on-a-tensorflow-deep-learning-task-after-reducing-b%23new-answer', 'question_page');
          }
          );

          Post as a guest















          Required, but never shown

























          0






          active

          oldest

          votes








          0






          active

          oldest

          votes









          active

          oldest

          votes






          active

          oldest

          votes
















          draft saved

          draft discarded




















































          Thanks for contributing an answer to Data Science Stack Exchange!


          • Please be sure to answer the question. Provide details and share your research!

          But avoid



          • Asking for help, clarification, or responding to other answers.

          • Making statements based on opinion; back them up with references or personal experience.


          Use MathJax to format equations. MathJax reference.


          To learn more, see our tips on writing great answers.




          draft saved


          draft discarded














          StackExchange.ready(
          function () {
          StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fdatascience.stackexchange.com%2fquestions%2f47918%2fhow-can-i-increase-cuda-load-on-a-tensorflow-deep-learning-task-after-reducing-b%23new-answer', 'question_page');
          }
          );

          Post as a guest















          Required, but never shown





















































          Required, but never shown














          Required, but never shown












          Required, but never shown







          Required, but never shown

































          Required, but never shown














          Required, but never shown












          Required, but never shown







          Required, but never shown







          Popular posts from this blog

          How to label and detect the document text images

          Vallis Paradisi

          Tabula Rosettana