{
  "id": 21961,
  "title": "Google Compute Engine and xgboost",
  "url": "/competitions/facebook-v-predicting-check-ins/discussion/21961",
  "author_name": "",
  "post_date": "2016-06-29T18:14:53.180Z",
  "votes": 1,
  "comment_count": 2,
  "views": 471,
  "content": "<p>I am trying to use an instance with 8vCPUs, but looking at the CPU utilization and considering the running time, the training doesn't seem to use all the cores. I have tried nthread = 8 and nthread = -1.</p>\n\n<p>Anyone has experience with this?</p>",
  "messages": [
    {
      "id": "125466",
      "postDate": "06/29/2016 18:14:53",
      "content": "<p>I am trying to use an instance with 8vCPUs, but looking at the CPU utilization and considering the running time, the training doesn't seem to use all the cores. I have tried nthread = 8 and nthread = -1.</p>\n\n<p>Anyone has experience with this?</p>",
      "rawMarkdown": "I am trying to use an instance with 8vCPUs, but looking at the CPU utilization and considering the running time, the training doesn't seem to use all the cores. I have tried nthread = 8 and nthread = -1.\r\n\r\nAnyone has experience with this?",
      "votes": null
    },
    {
      "id": "125468",
      "postDate": "06/29/2016 18:47:39",
      "content": "<p>[quote=Radu Stoicescu;125466]</p>\n\n<p>I am trying to use an instance with 8vCPUs, but looking at the CPU utilization and considering the running time, the training doesn't seem to use all the cores. I have tried nthread = 8 and nthread = -1.</p>\n\n<p>Anyone has experience with this?</p>\n\n<p>[/quote]\nFor the grid method, you can do embarrassingly parallel by putting different blocks on different threads</p>",
      "rawMarkdown": "[quote=Radu Stoicescu;125466]\r\n\r\nI am trying to use an instance with 8vCPUs, but looking at the CPU utilization and considering the running time, the training doesn't seem to use all the cores. I have tried nthread = 8 and nthread = -1.\r\n\r\nAnyone has experience with this?\r\n\r\n[/quote]\r\nFor the grid method, you can do embarrassingly parallel by putting different blocks on different threads",
      "votes": null
    },
    {
      "id": "125581",
      "postDate": "06/30/2016 15:08:06",
      "content": "<p>[quote=Radu Stoicescu;125466]</p>\n\n<p>I am trying to use an instance with 8vCPUs, but looking at the CPU utilization and considering the running time, the training doesn't seem to use all the cores. I have tried nthread = 8 and nthread = -1.</p>\n\n<p>Anyone has experience with this?</p>\n\n<p>[/quote]</p>\n\n<p>here is an example of multithreaded script which processes grid in parallel, you can use any classifier you like with this script: <a href=\"https://www.kaggle.com/valeriur/facebook-v-predicting-check-ins/process-grid-in-parallel/discussion\">https://www.kaggle.com/valeriur/facebook-v-predicting-check-ins/process-grid-in-parallel/discussion</a></p>",
      "rawMarkdown": "[quote=Radu Stoicescu;125466]\r\n\r\nI am trying to use an instance with 8vCPUs, but looking at the CPU utilization and considering the running time, the training doesn't seem to use all the cores. I have tried nthread = 8 and nthread = -1.\r\n\r\nAnyone has experience with this?\r\n\r\n[/quote]\r\n\r\nhere is an example of multithreaded script which processes grid in parallel, you can use any classifier you like with this script: https://www.kaggle.com/valeriur/facebook-v-predicting-check-ins/process-grid-in-parallel/discussion",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 125468,
      "author_name": "weizhezhang",
      "author_url": "",
      "post_date": "06/29/2016 18:47:39",
      "content": "<p>[quote=Radu Stoicescu;125466]</p>\n\n<p>I am trying to use an instance with 8vCPUs, but looking at the CPU utilization and considering the running time, the training doesn't seem to use all the cores. I have tried nthread = 8 and nthread = -1.</p>\n\n<p>Anyone has experience with this?</p>\n\n<p>[/quote]\nFor the grid method, you can do embarrassingly parallel by putting different blocks on different threads</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 125581,
      "author_name": "valeriur",
      "author_url": "",
      "post_date": "06/30/2016 15:08:06",
      "content": "<p>[quote=Radu Stoicescu;125466]</p>\n\n<p>I am trying to use an instance with 8vCPUs, but looking at the CPU utilization and considering the running time, the training doesn't seem to use all the cores. I have tried nthread = 8 and nthread = -1.</p>\n\n<p>Anyone has experience with this?</p>\n\n<p>[/quote]</p>\n\n<p>here is an example of multithreaded script which processes grid in parallel, you can use any classifier you like with this script: <a href=\"https://www.kaggle.com/valeriur/facebook-v-predicting-check-ins/process-grid-in-parallel/discussion\">https://www.kaggle.com/valeriur/facebook-v-predicting-check-ins/process-grid-in-parallel/discussion</a></p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "125466": "I am trying to use an instance with 8vCPUs, but looking at the CPU utilization and considering the running time, the training doesn't seem to use all the cores. I have tried nthread = 8 and nthread = -1.\r\n\r\nAnyone has experience with this?",
    "125468": "[quote=Radu Stoicescu;125466]\r\n\r\nI am trying to use an instance with 8vCPUs, but looking at the CPU utilization and considering the running time, the training doesn't seem to use all the cores. I have tried nthread = 8 and nthread = -1.\r\n\r\nAnyone has experience with this?\r\n\r\n[/quote]\r\nFor the grid method, you can do embarrassingly parallel by putting different blocks on different threads",
    "125581": "[quote=Radu Stoicescu;125466]\r\n\r\nI am trying to use an instance with 8vCPUs, but looking at the CPU utilization and considering the running time, the training doesn't seem to use all the cores. I have tried nthread = 8 and nthread = -1.\r\n\r\nAnyone has experience with this?\r\n\r\n[/quote]\r\n\r\nhere is an example of multithreaded script which processes grid in parallel, you can use any classifier you like with this script: https://www.kaggle.com/valeriur/facebook-v-predicting-check-ins/process-grid-in-parallel/discussion"
  },
  "source": "meta"
}