TransformerEngine-FL/.github/configs/cuda.yml at 2bc1987aeebef82389f39a42b564e59d9b4b6905 · Caozhou1995/TransformerEngine-FL · 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
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
# CUDA Hardware Configuration for TransformerEngine-FL
# Refactored for A100 Nodes
# This file defines environment variables, volumes, and test filters for TE tests.

hardware_name: cuda
display_name: 'NVIDIA CUDA (A100)'
checkout_submodules: recursive

# CI image for online env
ci_image: harbor.baai.ac.cn/flagscale/cuda12.8.1-torch2.7.1-python3.10-te2.9:20260209

# Runner labels for self-hosted A100 node
runner_labels:
  - self-hosted
  - Linux
  - X64
  - nvidia
  - gpu-8

# Runner labels for online env
runner_labels:
  - nv-8g-cicd-te

# Container volumes
container_volumes:
  - /home/flagscale_cicd/flask/static:/workspace/report

# Container options
container_options: >-
  --privileged
  --gpus all
  --shm-size=500g
  --ipc=host
  --ulimit memlock=-1
  --ulimit stack=67108864
  --user root

# Platform-specific environment setup script
setup_script: .github/scripts/setup_cuda.sh

# Device types to run tests on
device_types:
  - a100

coverage:
  enabled: true
  required: false
  python: /opt/miniconda3/envs/flagscale-train/bin/python3
  sources:
    - transformer_engine
  include:
    - transformer_engine/pytorch/*
    - transformer_engine/debug/*
    - transformer_engine/plugin/*
  omit:
    - '*/setup.py'
    - '*/transformer_engine/plugin/core/_build_config.py'

unit_test_matrix:
  - name: pytorch_debug
    runner: script
    path: qa/L0_pytorch_debug_unittest/test.sh
  - name: pytorch_unittest
    runner: script
    path: qa/L0_pytorch_unittest/test.sh
  - name: pytorch_distributed_unittest
    runner: script
    path: qa/L1_pytorch_distributed_unittest/test.sh
  - name: pytorch_onnx_unittest
    runner: script
    path: qa/L1_pytorch_onnx_unittest/test.sh

integration_test_matrix:
  - name: pytorch_mcore_integration
    path: qa/L1_pytorch_mcore_integration/test.sh

# Test matrix configuration
test_matrix:
  l0_pytorch:
    path: 'qa/L0_pytorch_unittest/test.sh'
    ignored_tests:
      - test_sanity_layernorm_mlp
      - test_sanity_gpt
      - test_sanity_bert
      - test_sanity_T5
      - test_sanity_amp_and_nvfuser
      - test_sanity_drop_path
      - test_layernorm_mlp_accuracy
      - test_grouped_linear_accuracy
      - test_gpt_accuracy
      - test_basic_linear
      - test_layer_norm