OOD-apps.server.jupyter/submit.yml.erb at master · WFU-HPC/OOD-apps.server.jupyter · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
<%-
  cores_lookup = {
      "any" => "64",
      "any-gpu" => "64",
      "64cores_zen4_higmem" => "64",
      "64cores_zen4" => "64",
      "48cores_cascade" => "48",
      "44cores_cascade" => "44",
      "44cores_skylake" => "44",
      "a100_80gb_zen4" => "64",
      "a100_40gb_zen3" => "64",
      "v100_32gb_zen3" => "64",
      "v100_32gb_skylake" => "64",
      "v100_32gb_cascade" => "48",
      "lowmangrp_lynx" => "32",
      "lowmangrp_lascar" => "32",
      "curtisgrp_crux" => "32",
      "yanggrp_lovelace" => "64",
      "devarakondagrp_hopper" => "128",
  }

  nodes = bc_num_slots.blank? ? 1 : bc_num_slots.to_i
  mem = memory.blank? ? 4 : memory.to_i
  max_cores = cores_lookup[node_type]
  ppn = num_cores.blank? ? max_cores : num_cores.to_i
  gpus = num_gpus.blank? ? 1: num_gpus.to_i

  case node_type
  when "any"
    slurm_args = [ "--nodes", "#{nodes}", "--ntasks-per-node", "#{ppn}", "--mem", "#{mem}G", "--constraint", "cpu"                                                      ]
  when "any-gpu"
    slurm_args = [ "--nodes", "#{nodes}", "--ntasks-per-node", "#{ppn}", "--mem", "#{mem}G", "--constraint", "gpu",                     "--gpus", "#{gpus}"             ]
  when "64cores_zen4_higmem"
    slurm_args = [ "--nodes", "#{nodes}", "--ntasks-per-node", "#{ppn}", "--mem", "#{mem}G", "--constraint", "cpu&64cores&zen4&highmem"                                 ]
  when "64cores_zen4"
    slurm_args = [ "--nodes", "#{nodes}", "--ntasks-per-node", "#{ppn}", "--mem", "#{mem}G", "--constraint", "cpu&64cores&zen4"                                         ]
  when "48cores_cascade"
    slurm_args = [ "--nodes", "#{nodes}", "--ntasks-per-node", "#{ppn}", "--mem", "#{mem}G", "--constraint", "cpu&48cores&cascade"                                      ]
  when "44cores_cascade"
    slurm_args = [ "--nodes", "#{nodes}", "--ntasks-per-node", "#{ppn}", "--mem", "#{mem}G", "--constraint", "cpu&44cores&cascade"                                      ]
  when "44cores_skylake"
    slurm_args = [ "--nodes", "#{nodes}", "--ntasks-per-node", "#{ppn}", "--mem", "#{mem}G", "--constraint", "cpu&44cores&skylake"                                      ]
  when "a100_80gb_zen4"
    slurm_args = [ "--nodes", "#{nodes}", "--ntasks-per-node", "#{ppn}", "--mem", "#{mem}G", "--constraint", "gpu&a100_80&zen4",        "--gres", "gpu:A100_80:#{gpus}" ]
  when "a100_40gb_zen3"
    slurm_args = [ "--nodes", "#{nodes}", "--ntasks-per-node", "#{ppn}", "--mem", "#{mem}G", "--constraint", "gpu&a100_40&zen3",        "--gres", "gpu:A100_40:#{gpus}" ]
  when "v100_32gb_zen3"
    slurm_args = [ "--nodes", "#{nodes}", "--ntasks-per-node", "#{ppn}", "--mem", "#{mem}G", "--constraint", "gpu&v100_32&zen3",        "--gres", "gpu:V100_32:#{gpus}" ]
  when "lowmangrp_lynx"
    slurm_args = [ "--nodes", "#{nodes}", "--ntasks-per-node", "#{ppn}", "--mem", "#{mem}G", "--nodelist", "lynx"                                                       ]
  when "lowmangrp_lascar"
    slurm_args = [ "--nodes", "#{nodes}", "--ntasks-per-node", "#{ppn}", "--mem", "#{mem}G", "--nodelist", "lascar"                                                     ]
  when "curtisgrp_crux"
    slurm_args = [ "--nodes", "#{nodes}", "--ntasks-per-node", "#{ppn}", "--mem", "#{mem}G", "--nodelist", "crux"                                                       ]
  when "yanggrp_lovelace"
    slurm_args = [ "--nodes", "#{nodes}", "--ntasks-per-node", "#{ppn}", "--mem", "#{mem}G", "--nodelist", "lovelace",                  "--gres", "gpu:L40S:#{gpus}"    ]
  when "devarakondagrp_hopper"
    slurm_args = [ "--nodes", "#{nodes}", "--ntasks-per-node", "#{ppn}", "--mem", "#{mem}G", "--nodelist", "hopper"                                                     ]
  else
    slurm_args = [ "--nodes", "#{nodes}", "--ntasks-per-node", "#{ppn}", "--mem", "#{mem}G"                                                                             ]
  end
%>
---
batch_connect:
  template: basic
script:
  accounting_id: "<%= accounts %>"
  queue_name: "<%= queues %>"
  native:
  <%- slurm_args.each do |arg| %>
    - "<%= arg %>"
  <%- end %>