"megatron/legacy/data/orqa_wiki_dataset.py" did not exist on "fabad46192f78dfac87da5d0198ec3fac972841d"
__init__.py 2.24 KB
Newer Older
wangsen's avatar
wangsen committed
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
# Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved.

import os
import pathlib
import subprocess

from torch.utils import cpp_extension

# Setting this param to a list has a problem of generating different
# compilation commands (with diferent order of architectures) and
# leading to recompilation of fused kernels. Set it to empty string
# to avoid recompilation and assign arch flags explicity in
# extra_cuda_cflags below
os.environ["TORCH_CUDA_ARCH_LIST"] = ""


def load(args):

    # Check if cuda 11 is installed for compute capability 8.0
    cc_flag = []
    _, bare_metal_major, bare_metal_minor = _get_cuda_bare_metal_version(
        cpp_extension.CUDA_HOME
    )
    if int(bare_metal_major) >= 11:
        cc_flag.append('-gencode')
        cc_flag.append('arch=compute_80,code=sm_80')
        if int(bare_metal_minor) >= 8:
            cc_flag.append('-gencode')
            cc_flag.append('arch=compute_90,code=sm_90')

    # Build path
    srcpath = pathlib.Path(__file__).parent.absolute()
    buildpath = srcpath / "build"
    _create_build_dir(buildpath)

    # Helper function to build the kernels.
    def _cpp_extention_load_helper(name, sources, extra_cuda_flags):
        return cpp_extension.load(
            name=name,
            sources=sources,
            build_directory=buildpath,
            extra_cflags=[
                "-O3",
            ],
            extra_cuda_cflags=[
                "-O3",
                "-gencode",
                "arch=compute_70,code=sm_70",
                "--use_fast_math",
            ]
            + extra_cuda_flags
            + cc_flag,
            verbose=(args.rank == 0),
        )


def _get_cuda_bare_metal_version(cuda_dir):
    raw_output = subprocess.check_output(
        [cuda_dir + "/bin/nvcc", "-V"], universal_newlines=True
    )
    output = raw_output.split()
    release_idx = output.index("release") + 1
    release = output[release_idx].split(".")
    bare_metal_major = release[0]
    bare_metal_minor = release[1][0]

    return raw_output, bare_metal_major, bare_metal_minor


def _create_build_dir(buildpath):
    try:
        os.mkdir(buildpath)
    except OSError:
        if not os.path.isdir(buildpath):
            print(f"Creation of the build directory {buildpath} failed")