{
    "info": {
        "author": "Bin Wang",
        "author_email": "wwwbbb8510@gmail.com",
        "bugtrack_url": null,
        "classifiers": [
            "License :: OSI Approved :: MIT License",
            "Operating System :: OS Independent",
            "Programming Language :: Python :: 3"
        ],
        "description": "# cudam\nCuda Mangement - multi-process, scheduled jobs, distributed processing\n\n## command to check all cuda server status\n\n```bash\ndate >> cuda_status.txt && echo 'cuda1' >> cuda_status.txt && ssh cuda1 'nvidia-smi' >> cuda_status.txt && echo 'cuda2' >> cuda_status.txt && ssh cuda2 'nvidia-smi' >> cuda_status.txt && echo 'cuda3' >> cuda_status.txt && ssh cuda3 'nvidia-smi' >> cuda_status.txt && echo 'cuda4' >> cuda_status.txt && ssh cuda4 'nvidia-smi' >> cuda_status.txt && echo 'cuda5' >> cuda_status.txt && ssh cuda5 'nvidia-smi' >> cuda_status.txt && echo 'cuda6' >> cuda_status.txt && ssh cuda6 'nvidia-smi' >> cuda_status.txt && echo 'cuda11' >> cuda_status.txt && ssh cuda11 'nvidia-smi' >> cuda_status.txt\n```\n\n## server-client mode to utilize multi-GPUs across Multi-Machines\n\n### server side - develop the code that runs on a single GPU\n\n```python\n# here is a dumb function to evaluate densenet\n# it should be replaced by the actual code of evaluation\ndef evaluate_densenet(model):\n    acc = 0.99\n    return acc\n```\n\n### client size - develop the code to send the models to server for evaluation\n\n* Add available GPU servers in the server list configuration file\n\n```text\n# configuration of server list\ncuda4,8000\ncuda4,8001\ncuda5,8000\ncuda5,8001\ncuda5,8002\n```\n\n* The client code that concurrently evaluates models\n\n```python\nfrom cudam.cudam_socket.client import GPUClientPool\nDEFAULT_RUN_CODE_WORK_DIRECTORY = \"/home/www/server\" # the folder where the server side code resides \nDEFAULT_RUN_CODE_PATH = \"server_file\" # the file name of the server side code\nSERVER_LIST_CONFIG = 'config/server_list.txt' # the configuration file of the server list\ndef pool_evaluate_densenet(model_list):\n    # generat the arguments which will passed to client pool\n    arr_args = []\n    for m in model_list:        \n        singe_args = {'model': m}\n        arr_args.append({\n            'path': DEFAULT_RUN_CODE_PATH,\n            'entry': \"evaluate_densenet\",\n            'work_directory': DEFAULT_RUN_CODE_WORK_DIRECTORY,\n            'args': singe_args,\n            'use_cuda': True # whether to use GPU or not\n        })\n    # init client pool\n    server_list = GPUClientPool.load_server_list_from_file(SERVER_LIST_CONFIG)\n    pool = GPUClientPool(server_list)\n    # perform evaluation\n    eval_result = pool.run_code_batch(arr_args)\n    return eval_result\n# main entrance\nif __name__ == '__main__':\n    model_list =[] # dumb model list which needs to be replaced by real models\n    pool_evaluate_densenet(model_list)\n```\n\n### start the server \n\n* After installation of this package, `cudam_server.py` should be automatically copied to the bin path; if not, please manually copy this file to the root folder of the project. The server can be started by running the following command:  \n\n```bash\nnohup python cudam_server.py -s 1 -i cuda1 -p 8000 -g 0 >& log/nohup_cuda_1_8000_0.log &\n```\n\n### run the client side python code to evaluate a batch of models\n\n## task manager\n\n### task template\n\n```bash\n#!/usr/bin/env bash\n\nwhile getopts g: option;do\n    case \"${option}\" in\n    g) GPU_ID=${OPTARG};;\n    esac\ndone\n\nprint_help(){\n    printf \"Parameter g(GPU ID) is mandatory\\n\"\n    printf \"g values - GPU ID\"\n    exit 1\n}\n\nif [ -z \"${GPU_ID}\" ];then\n    print_help\nfi\n\necho \"start task on GPU: $GPU_ID\"\n\n# the root directory of your python script\ncd ~/code/psocnn/\n# the main python script accepting the gpu ID in -g argument\npython3 main.py -g ${GPU_ID}\n```\n\n### task folder structure\n\n![task folder structure](https://github.com/wwwbbb8510/cudam/blob/master/sh_task_structure.PNG \"Task folder structure\")\n\n### task manager \n\n```bash\n# start task manager\nnohup cudam_task_manager.py -n 2 -s 2 -i 60 -f 300 &\n# snap gpu\ncudam_snap_gpu.py -s 2 -l 60 -g 1\n```\n\n### install cumdam for a specific user and can not add the local path into executable PATH\n\n* Switch to the root folder of your project\n\n* Install cudam package\n```bash\npip install --user cudam\n``` \n\n* Create a soft link of the executable file\n```bash\nln -s /home/{YOURUSER}/.local/bin/cudam_task_manager.py cudam_task_manager.py\nln -s /home/{YOURUSER}/.local/bin/cudam_snap_gpu.py cudam_snap_gpu.py\n```\n\n* Run the task manager\n```bash\n# run interactively\npython cudam_task_manager.py -n 2 -s 2 -i 60 -f 300\n# run in background\nnohup python cudam_task_manager.py -n 2 -s 2 -i 60 -f 300 &\n```\n\n\n\n",
        "description_content_type": "text/markdown",
        "docs_url": null,
        "download_url": "",
        "downloads": {
            "last_day": -1,
            "last_month": -1,
            "last_week": -1
        },
        "home_page": "https://github.com/wwwbbb8510/cudam.git",
        "keywords": "",
        "license": "",
        "maintainer": "",
        "maintainer_email": "",
        "name": "cudam",
        "package_url": "https://pypi.org/project/cudam/",
        "platform": "",
        "project_url": "https://pypi.org/project/cudam/",
        "project_urls": {
            "Homepage": "https://github.com/wwwbbb8510/cudam.git"
        },
        "release_url": "https://pypi.org/project/cudam/0.0.6/",
        "requires_dist": null,
        "requires_python": "",
        "summary": "Cuda Mangement - multi-process, scheduled jobs, distributed processing",
        "version": "0.0.6"
    },
    "last_serial": 4812915,
    "releases": {
        "0.0.3": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "2dc4d8ffb2794287c4a445f1827e75da",
                    "sha256": "d2e0a0060d8a9e2ede6adab64250e3aa43285deda2855cb3a28f4c4e72434a96"
                },
                "downloads": -1,
                "filename": "cudam-0.0.3-py3-none-any.whl",
                "has_sig": false,
                "md5_digest": "2dc4d8ffb2794287c4a445f1827e75da",
                "packagetype": "bdist_wheel",
                "python_version": "py3",
                "requires_python": null,
                "size": 16440,
                "upload_time": "2018-08-13T02:51:28",
                "url": "https://files.pythonhosted.org/packages/a1/55/cf75c2386ce5b2026e50a0ea260bffd066c83d38a70c781b775a1d5b5856/cudam-0.0.3-py3-none-any.whl"
            },
            {
                "comment_text": "",
                "digests": {
                    "md5": "a51dfab5159291c5becbe40a9bfde846",
                    "sha256": "85b9d65b5418816c1c82d688650ec0fa11ac46322b17a8357a9929b67bf352a5"
                },
                "downloads": -1,
                "filename": "cudam-0.0.3.tar.gz",
                "has_sig": false,
                "md5_digest": "a51dfab5159291c5becbe40a9bfde846",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": null,
                "size": 7334,
                "upload_time": "2018-08-13T02:51:29",
                "url": "https://files.pythonhosted.org/packages/b2/62/dbb10014590217bf327bf9e0a26e50e5e46682d60428f1a0f2ed0fe62b29/cudam-0.0.3.tar.gz"
            }
        ],
        "0.0.4": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "e02344dccb3a4f85c19f9f3c08e950d3",
                    "sha256": "d50680be5e3d3926e6339db707fabc1a6d5397e48f2256403b0ce5002d0245c8"
                },
                "downloads": -1,
                "filename": "cudam-0.0.4-py3-none-any.whl",
                "has_sig": false,
                "md5_digest": "e02344dccb3a4f85c19f9f3c08e950d3",
                "packagetype": "bdist_wheel",
                "python_version": "py3",
                "requires_python": null,
                "size": 16697,
                "upload_time": "2018-08-13T03:44:36",
                "url": "https://files.pythonhosted.org/packages/ce/ea/cc7eb569908683330d5d636b4bc7010c7a9281b8aa235c3e758f15d403cd/cudam-0.0.4-py3-none-any.whl"
            },
            {
                "comment_text": "",
                "digests": {
                    "md5": "a47f88d1d8c98a4c51eb32c5ba20fcc8",
                    "sha256": "d7f1ec3f74b98860b7c34e4eb76a5e4270a6cb33e029ef5da7f15fe7d042a0d6"
                },
                "downloads": -1,
                "filename": "cudam-0.0.4.tar.gz",
                "has_sig": false,
                "md5_digest": "a47f88d1d8c98a4c51eb32c5ba20fcc8",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": null,
                "size": 7543,
                "upload_time": "2018-08-13T03:44:37",
                "url": "https://files.pythonhosted.org/packages/bb/73/3f6af1e4062fcc4e716609c964b96bc03037380179d94107ac59eb385434/cudam-0.0.4.tar.gz"
            }
        ],
        "0.0.5": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "d2260302eb674e427853dd583050130d",
                    "sha256": "4e6f9b871e64bdcb1b9f1f44f507a9d27fd49be894a4d1fdbdab3dc639c0d94f"
                },
                "downloads": -1,
                "filename": "cudam-0.0.5-py3-none-any.whl",
                "has_sig": false,
                "md5_digest": "d2260302eb674e427853dd583050130d",
                "packagetype": "bdist_wheel",
                "python_version": "py3",
                "requires_python": null,
                "size": 53774,
                "upload_time": "2019-02-12T21:23:21",
                "url": "https://files.pythonhosted.org/packages/ac/56/f9eb6e106b5edc55707fcd095273199f9725eab7ae62e0c722fa87becb6c/cudam-0.0.5-py3-none-any.whl"
            },
            {
                "comment_text": "",
                "digests": {
                    "md5": "67fa8f181ff3162c934d9d3cd68bab15",
                    "sha256": "c6ae788a566ed49456640dbef6a2545660d47b125bd6da1cce57e8ce1d5297a1"
                },
                "downloads": -1,
                "filename": "cudam-0.0.5.tar.gz",
                "has_sig": false,
                "md5_digest": "67fa8f181ff3162c934d9d3cd68bab15",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": null,
                "size": 27161,
                "upload_time": "2019-02-12T21:23:23",
                "url": "https://files.pythonhosted.org/packages/f6/2b/93b7e5f91f8bbfca8d1cde224240f75da3ec3c85b5aa658cb0187c76e204/cudam-0.0.5.tar.gz"
            }
        ],
        "0.0.6": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "cfa6915ec31e997f874b4e7e6e5ba1dd",
                    "sha256": "1358914da93359e1c01727855d18ad68e438713a8f7ffd40ef64d16149949961"
                },
                "downloads": -1,
                "filename": "cudam-0.0.6-py3-none-any.whl",
                "has_sig": false,
                "md5_digest": "cfa6915ec31e997f874b4e7e6e5ba1dd",
                "packagetype": "bdist_wheel",
                "python_version": "py3",
                "requires_python": null,
                "size": 53768,
                "upload_time": "2019-02-12T21:29:33",
                "url": "https://files.pythonhosted.org/packages/78/34/e13d6e69ef1a1fab8f5c4344e01cde055c09e5ee8af511fbbc240d8d8711/cudam-0.0.6-py3-none-any.whl"
            },
            {
                "comment_text": "",
                "digests": {
                    "md5": "2a7ee4388b7d9b4734bf54db9a1ceaf7",
                    "sha256": "e87b137d1cf70e817a9781f1ca194865bfe6e37bee4323e6d93f883d7bbfda07"
                },
                "downloads": -1,
                "filename": "cudam-0.0.6.tar.gz",
                "has_sig": false,
                "md5_digest": "2a7ee4388b7d9b4734bf54db9a1ceaf7",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": null,
                "size": 27156,
                "upload_time": "2019-02-12T21:29:35",
                "url": "https://files.pythonhosted.org/packages/ce/cb/e3c028d59eeab9853ba73fb12c35b10e114917b5bdb57b1c6209546b39d2/cudam-0.0.6.tar.gz"
            }
        ]
    },
    "urls": [
        {
            "comment_text": "",
            "digests": {
                "md5": "cfa6915ec31e997f874b4e7e6e5ba1dd",
                "sha256": "1358914da93359e1c01727855d18ad68e438713a8f7ffd40ef64d16149949961"
            },
            "downloads": -1,
            "filename": "cudam-0.0.6-py3-none-any.whl",
            "has_sig": false,
            "md5_digest": "cfa6915ec31e997f874b4e7e6e5ba1dd",
            "packagetype": "bdist_wheel",
            "python_version": "py3",
            "requires_python": null,
            "size": 53768,
            "upload_time": "2019-02-12T21:29:33",
            "url": "https://files.pythonhosted.org/packages/78/34/e13d6e69ef1a1fab8f5c4344e01cde055c09e5ee8af511fbbc240d8d8711/cudam-0.0.6-py3-none-any.whl"
        },
        {
            "comment_text": "",
            "digests": {
                "md5": "2a7ee4388b7d9b4734bf54db9a1ceaf7",
                "sha256": "e87b137d1cf70e817a9781f1ca194865bfe6e37bee4323e6d93f883d7bbfda07"
            },
            "downloads": -1,
            "filename": "cudam-0.0.6.tar.gz",
            "has_sig": false,
            "md5_digest": "2a7ee4388b7d9b4734bf54db9a1ceaf7",
            "packagetype": "sdist",
            "python_version": "source",
            "requires_python": null,
            "size": 27156,
            "upload_time": "2019-02-12T21:29:35",
            "url": "https://files.pythonhosted.org/packages/ce/cb/e3c028d59eeab9853ba73fb12c35b10e114917b5bdb57b1c6209546b39d2/cudam-0.0.6.tar.gz"
        }
    ]
}