Extremely high CPU usage when training a model on GPU











up vote
1
down vote

favorite












Recently I discovered something rather strange with the project that I worked on for quite a while already. The model I have is rather conventional: a convnet with a few fully connected layers. For data loading I use tf.data API, but the same thing happens with queue-based code that I had before porting to tf.data. After a few hours since the training of the model begins, the CPU usage rises to very high levels, 1500-2000% as reported by the htop util. And at the beginning of training everything is fine, the main process shows only about 200% CPU usage. Attached is the screenshot of the htop output, and another thing that's worrying is all the child processes that also have pretty high CPU load.



I am using tensorflow-gpu version 1.11, running it on NVIDIA Tesla V100. I am pretty sure that the model does run on the GPU and not on the CPU: nvidia-smi shows that GPU is occupied at an about 70% rate.



Obviously, I cannot ask for an exact cause of this, and it would be difficult to strip the problem down to a reproducible test case. However, may be you could point me at some debugging techniques that are applicable in such case.



htop output






python tensorflow







share|improve this question


























    up vote
    1
    down vote

    favorite












    Recently I discovered something rather strange with the project that I worked on for quite a while already. The model I have is rather conventional: a convnet with a few fully connected layers. For data loading I use tf.data API, but the same thing happens with queue-based code that I had before porting to tf.data. After a few hours since the training of the model begins, the CPU usage rises to very high levels, 1500-2000% as reported by the htop util. And at the beginning of training everything is fine, the main process shows only about 200% CPU usage. Attached is the screenshot of the htop output, and another thing that's worrying is all the child processes that also have pretty high CPU load.



    I am using tensorflow-gpu version 1.11, running it on NVIDIA Tesla V100. I am pretty sure that the model does run on the GPU and not on the CPU: nvidia-smi shows that GPU is occupied at an about 70% rate.



    Obviously, I cannot ask for an exact cause of this, and it would be difficult to strip the problem down to a reproducible test case. However, may be you could point me at some debugging techniques that are applicable in such case.



    htop output






    python tensorflow







    share|improve this question
























      up vote
      1
      down vote

      favorite









      up vote
      1
      down vote

      favorite











      Recently I discovered something rather strange with the project that I worked on for quite a while already. The model I have is rather conventional: a convnet with a few fully connected layers. For data loading I use tf.data API, but the same thing happens with queue-based code that I had before porting to tf.data. After a few hours since the training of the model begins, the CPU usage rises to very high levels, 1500-2000% as reported by the htop util. And at the beginning of training everything is fine, the main process shows only about 200% CPU usage. Attached is the screenshot of the htop output, and another thing that's worrying is all the child processes that also have pretty high CPU load.



      I am using tensorflow-gpu version 1.11, running it on NVIDIA Tesla V100. I am pretty sure that the model does run on the GPU and not on the CPU: nvidia-smi shows that GPU is occupied at an about 70% rate.



      Obviously, I cannot ask for an exact cause of this, and it would be difficult to strip the problem down to a reproducible test case. However, may be you could point me at some debugging techniques that are applicable in such case.



      htop output






      python tensorflow







      share|improve this question













      Recently I discovered something rather strange with the project that I worked on for quite a while already. The model I have is rather conventional: a convnet with a few fully connected layers. For data loading I use tf.data API, but the same thing happens with queue-based code that I had before porting to tf.data. After a few hours since the training of the model begins, the CPU usage rises to very high levels, 1500-2000% as reported by the htop util. And at the beginning of training everything is fine, the main process shows only about 200% CPU usage. Attached is the screenshot of the htop output, and another thing that's worrying is all the child processes that also have pretty high CPU load.



      I am using tensorflow-gpu version 1.11, running it on NVIDIA Tesla V100. I am pretty sure that the model does run on the GPU and not on the CPU: nvidia-smi shows that GPU is occupied at an about 70% rate.



      Obviously, I cannot ask for an exact cause of this, and it would be difficult to strip the problem down to a reproducible test case. However, may be you could point me at some debugging techniques that are applicable in such case.



      htop output






      python tensorflow




      python tensorflow






      share|improve this question













      share|improve this question











      share|improve this question




      share|improve this question










      asked Nov 21 at 11:26









      SimpleMan

      142315




      142315





























          active

          oldest

          votes











          Your Answer






          StackExchange.ifUsing("editor", function () {
          StackExchange.using("externalEditor", function () {
          StackExchange.using("snippets", function () {
          StackExchange.snippets.init();
          });
          });
          }, "code-snippets");

          StackExchange.ready(function() {
          var channelOptions = {
          tags: "".split(" "),
          id: "1"
          };
          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',
          convertImagesToLinks: true,
          noModals: true,
          showLowRepImageUploadWarning: true,
          reputationToPostImages: 10,
          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%2fstackoverflow.com%2fquestions%2f53411097%2fextremely-high-cpu-usage-when-training-a-model-on-gpu%23new-answer', 'question_page');
          }
          );

          Post as a guest















          Required, but never shown






























          active

          oldest

          votes













          active

          oldest

          votes









          active

          oldest

          votes






          active

          oldest

          votes
















           

          draft saved


          draft discarded



















































           


          draft saved


          draft discarded














          StackExchange.ready(
          function () {
          StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fstackoverflow.com%2fquestions%2f53411097%2fextremely-high-cpu-usage-when-training-a-model-on-gpu%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

          Contact image not getting when fetch all contact list from iPhone by CNContact

          Image
          up vote 0 down vote favorite I know this question already asked but not getting solution. From this code I will get all the information from the contact but image not found when open vcf files on mac os, also not getting when share this file. I use this stackoverflow link here but It's not help full. var contacts = [CNContact]() let keys = [CNContactVCardSerialization.descriptorForRequiredKeys() ] as [Any] let request = CNContactFetchRequest(keysToFetch: keys as! [CNKeyDescriptor]) do { try self.contactStore.enumerateContacts(with: request) { (contact, stop) in // Array containing all unified contacts from everywhere contacts.append(contact) } } catch { print("unable to fetch contacts") } do { let data = try
          Read more

          count number of partitions of a set with n elements into k subsets

          Image
          up vote 6 down vote favorite 1 This program is for count number of partitions of a set with n elements into k subsets I am confusing here return k*countP(n-1, k) + countP(n-1, k-1); can some one explain what is happening here? why we are multiplying with k? NOTE->I know this is not the best way to calculate number of partitions that would be DP // A C++ program to count number of partitions // of a set with n elements into k subsets #include<iostream> using namespace std; // Returns count of different partitions of n // elements in k subsets int countP(int n, int k) { // Base cases if (n == 0 || k == 0 || k > n) return 0; if (k == 1 || k == n) return 1; // S(n+1, k) = k*S(n, k) + S(n, k-1) return k*countP(n-1, k) + countP(n-1, k-1); } // Driver program int ma
          Read more

          A CLEAN and SIMPLE way to add appendices to Table of Contents and bookmarks

          Image
          0 I found a lot of questions abount appendices and ToC. Many users want appendices to be grouped in an Appendix part, however some problems arise with ToC, hyperref, PDF viewer bookmarks, and so on. There are different solutions which require extra packages, command patching and other extra code, however none of them satisfies me. I almost found an easy way to accomplish a good result, where appendices are added to bookmarks in the right way and hyperref links point to the right page. However, the number of the "Appendix" part page is wrong (it's the number of appendix A). Is there any EASY way to fix that? This is a MWE: documentclass{book} usepackage[nottoc,notlot,notlof]{tocbibind} usepackage{hyperref} begin{document} frontmatter tableofcontents mainmatter part{First} chapter{
          Read more