Huawei_Technology
/
mindspore
Not watched
1
0
Fork 0
Code
Releases 13 Wiki evaluate Activity
Issues 0 Pull Requests 0 Datasets Model Cloudbrain HPC
You can not select more than 25 topics Topics must start with a chinese character,a letter or number, can include dashes ('-') and can be up to 35 characters long.
Tree: 2a695cfe24
Branches Tags
${ item.name }
Create branch ${ searchTerm }
from '2a695cfe24'
${ noResults }
mindspore/config/hccl_single_machine_multi_r...

111 lines
3.7 kB
Raw Blame History 

  1. {

  2.     "board_id": "0x0000",

  3.     "chip_info": "910",

  4.     "deploy_mode": "lab",

  5.     "group_count": "1",

  6.     "group_list": [{

  7.         "device_num": "8",

  8.         "server_num": "1",

  9.         "group_name": "",

  10.         "instance_count": "8",

  11.         "instance_list": [{

  12.                 "devices": [{

  13.                     "device_id": "0",

  14.                     "device_ip": "[device_ip_0]"

  15.                 }],

  16.                 "rank_id": "0",

  17.                 "server_id": "[server_id]"

  18.             },

  19.             {

  20.                 "devices": [{

  21.                     "device_id": "1",

  22.                     "device_ip": "[device_ip_1]"

  23.                 }],

  24.                 "rank_id": "1",

  25.                 "server_id": "[server_id]"

  26.             },

  27.             {

  28.                 "devices": [{

  29.                     "device_id": "2",

  30.                     "device_ip": "[device_ip_2]"

  31.                 }],

  32.                 "rank_id": "2",

  33.                 "server_id": "[server_id]"

  34.             },

  35.             {

  36.                 "devices": [{

  37.                     "device_id": "3",

  38.                     "device_ip": "[device_ip_3]"

  39.                 }],

  40.                 "rank_id": "3",

  41.                 "server_id": "[server_id]"

  42.             },

  43.             {

  44.                 "devices": [{

  45.                     "device_id": "4",

  46.                     "device_ip": "[device_ip_4]"

  47.                 }],

  48.                 "rank_id": "4",

  49.                 "server_id": "[server_id]"

  50.             },

  51.             {

  52.                 "devices": [{

  53.                     "device_id": "5",

  54.                     "device_ip": "[device_ip_5]"

  55.                 }],

  56.                 "rank_id": "5",

  57.                 "server_id": "[server_id]"

  58.             },

  59.             {

  60.                 "devices": [{

  61.                     "device_id": "6",

  62.                     "device_ip": "[device_ip_6]"

  63.                 }],

  64.                 "rank_id": "6",

  65.                 "server_id": "[server_id]"

  66.             },

  67.             {

  68.                 "devices": [{

  69.                     "device_id": "7",

  70.                     "device_ip": "[device_ip_7]"

  71.                 }],

  72.                 "rank_id": "7",

  73.                 "server_id": "[server_id]"

  74.             }

  75.         ]

  76.     }],

  77.     "para_plane_nic_location": "device",

  78.     "para_plane_nic_name": [

  79.         "eth0",

  80.         "eth1",

  81.         "eth2",

  82.         "eth3",

  83.         "eth4",

  84.         "eth5",

  85.         "eth6",

  86.         "eth7"

  87.     ],

  88.     "para_plane_nic_num": "8",

  89.     "status": "completed",

  90. 

  91.     "hccl_config_json_spec": {

  92.         "board_id": "board id, current support x0000 or 0x3000",

  93.         "chip_info": "chip info, current is 910",

  94.         "deploy_mode": "current use lab",

  95.         "group_count": "number of groups used",

  96.         "group_list": "detailed group information",

  97.         "device_num": "number of devices used, the value is the nth power of 2",

  98.         "server_num": "single machine is 1",

  99.         "group_name": "default is hccl_world_group or specified",

  100.         "instance_count": "number of instance used, generally equal to device_num",

  101.         "instance_list": "detailed instance information",

  102.         "device_id": "designated davinic device id to use, values start from 0, but no more than single machine total device num",

  103.         "device_ip": "ip corresponding to device_id",

  104.         "rank_id": "the first device must be 0 and then increase in order",

  105.         "server_id": "can be specified as the machine's ip address",

  106.         "para_plane_nic_location": "current use device",

  107.         "para_plane_nic_name": "network card corresponding to device ip",

  108.         "para_plane_nic_num": "number of network cards used",

  109.         "status": "current use completed"

  110.     }

  111. }