Skip to content

Refine setup.py #10577

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 3 commits into from
May 15, 2025
Merged

Refine setup.py #10577

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
10ee376
fix
risemeup1 May 9, 2025
f4bb9e1
fix
risemeup1 May 15, 2025
049a630
fix
risemeup1 May 15, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Prev Previous commit
Next Next commit
fix
  • Loading branch information
@risemeup1
risemeup1 committed May 15, 2025
commit f4bb9e1c8eeadc0999a3720dbd07926aaf4fdcb3
103 changes: 12 additions & 91 deletions slm/model_zoo/gpt-3/external_ops/setup.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -15,12 +15,13 @@
import multiprocessing
import os

# def get_gencode_flags():
# import paddle

# prop = paddle.device.cuda.get_device_properties()
# cc = prop.major * 10 + prop.minor
# return ["-gencode", "arch=compute_{0},code=sm_{0}".format(cc)]
def get_gencode_flags():
import paddle

prop = paddle.device.cuda.get_device_properties()
cc = prop.major * 10 + prop.minor
return ["-gencode", "arch=compute_{0},code=sm_{0}".format(cc)]


def run(func):
Expand All @@ -42,7 +43,7 @@ def setup_fast_ln():
if is_compiled_with_rocm():
print("The 'fasl_ln' feature is temporarily not supported on the ROCm platform !!!")
else:
# gencode_flags = get_gencode_flags()
gencode_flags = get_gencode_flags()
change_pwd()
setup(
name="fast_ln",
Expand All @@ -66,7 +67,8 @@ def setup_fast_ln():
"--expt-relaxed-constexpr",
"--expt-extended-lambda",
"--use_fast_math",
],
]
+ gencode_flags,
},
),
)
Expand All @@ -76,7 +78,7 @@ def setup_fused_ln():
from paddle.device import is_compiled_with_rocm
from paddle.utils.cpp_extension import CUDAExtension, setup

# gencode_flags = get_gencode_flags()
gencode_flags = get_gencode_flags()
change_pwd()
if is_compiled_with_rocm():
setup(
Expand Down Expand Up @@ -122,93 +124,12 @@ def setup_fused_ln():
"--expt-extended-lambda",
"--use_fast_math",
"-maxrregcount=50",
],
]
+ gencode_flags,
},
),
)


def setup_fused_quant_ops():
"""setup_fused_fp8_ops"""
from paddle.utils.cpp_extension import CUDAExtension, setup

# gencode_flags = get_gencode_flags()
change_pwd()
setup(
name="FusedQuantOps",
ext_modules=CUDAExtension(
sources=[
"fused_quanted_ops/fused_swiglu_act_quant.cu",
"fused_quanted_ops/fused_act_quant.cu",
"fused_quanted_ops/fused_act_dequant.cu",
"fused_quanted_ops/fused_act_dequant_transpose_act_quant.cu",
],
extra_compile_args={
"cxx": ["-O3", "-w", "-Wno-abi", "-fPIC", "-std=c++17"],
"nvcc": [
"-O3",
"-U__CUDA_NO_HALF_OPERATORS__",
"-U__CUDA_NO_HALF_CONVERSIONS__",
"-U__CUDA_NO_BFLOAT16_OPERATORS__",
"-U__CUDA_NO_BFLOAT16_CONVERSIONS__",
"-U__CUDA_NO_BFLOAT162_OPERATORS__",
"-U__CUDA_NO_BFLOAT162_CONVERSIONS__",
"-DCUTE_ARCH_MMA_SM90A_ENABLE",
"--expt-relaxed-constexpr",
"--expt-extended-lambda",
"--use_fast_math",
"-lineinfo",
"-DCUTLASS_DEBUG_TRACE_LEVEL=0",
"-maxrregcount=50",
"-gencode=arch=compute_90a,code=sm_90a",
"-DNDEBUG",
],
},
),
)


def setup_token_dispatcher_utils():
from paddle.utils.cpp_extension import CUDAExtension, setup

change_pwd()
setup(
name="TokenDispatcherUtils",
ext_modules=CUDAExtension(
sources=[
"token_dispatcher_utils/topk_to_multihot.cu",
"token_dispatcher_utils/topk_to_multihot_grad.cu",
"token_dispatcher_utils/tokens_unzip_and_zip.cu",
"token_dispatcher_utils/tokens_stable_unzip.cu",
"token_dispatcher_utils/tokens_guided_unzip.cu",
"token_dispatcher_utils/regroup_tokens.cu",
],
extra_compile_args={
"cxx": ["-O3", "-w", "-Wno-abi", "-fPIC", "-std=c++17"],
"nvcc": [
"-O3",
"-U__CUDA_NO_HALF_OPERATORS__",
"-U__CUDA_NO_HALF_CONVERSIONS__",
"-U__CUDA_NO_BFLOAT16_OPERATORS__",
"-U__CUDA_NO_BFLOAT16_CONVERSIONS__",
"-U__CUDA_NO_BFLOAT162_OPERATORS__",
"-U__CUDA_NO_BFLOAT162_CONVERSIONS__",
"-DCUTE_ARCH_MMA_SM90A_ENABLE",
"--expt-relaxed-constexpr",
"--expt-extended-lambda",
"--use_fast_math",
"-maxrregcount=80",
"-lineinfo",
"-DCUTLASS_DEBUG_TRACE_LEVEL=0",
"-gencode=arch=compute_90a,code=sm_90a",
"-DNDEBUG",
],
},
),
)


run(setup_token_dispatcher_utils)
run(setup_fused_quant_ops)
run(setup_fast_ln)
run(setup_fused_ln)
122 changes: 122 additions & 0 deletions slm/model_zoo/gpt-3/external_ops/setup_fp8.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,122 @@
# Copyright (c) 2023 PaddlePaddle Authors. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

import multiprocessing
import os


def get_gencode_flags():
import paddle

prop = paddle.device.cuda.get_device_properties()
cc = prop.major * 10 + prop.minor
return ["-gencode", "arch=compute_{0},code=sm_{0}".format(cc)]


def run(func):
p = multiprocessing.Process(target=func)
p.start()
p.join()


def change_pwd():
path = os.path.dirname(__file__)
if path:
os.chdir(path)


def setup_fused_quant_ops():
"""setup_fused_fp8_ops"""
from paddle.utils.cpp_extension import CUDAExtension, setup

gencode_flags = get_gencode_flags()
change_pwd()
setup(
name="FusedQuantOps",
ext_modules=CUDAExtension(
sources=[
"fused_quanted_ops/fused_swiglu_act_quant.cu",
"fused_quanted_ops/fused_act_quant.cu",
"fused_quanted_ops/fused_act_dequant.cu",
"fused_quanted_ops/fused_act_dequant_transpose_act_quant.cu",
"fused_quanted_ops/fused_spaq.cu",
],
extra_compile_args={
"cxx": ["-O3", "-w", "-Wno-abi", "-fPIC", "-std=c++17"],
"nvcc": [
"-O3",
"-U__CUDA_NO_HALF_OPERATORS__",
"-U__CUDA_NO_HALF_CONVERSIONS__",
"-U__CUDA_NO_BFLOAT16_OPERATORS__",
"-U__CUDA_NO_BFLOAT16_CONVERSIONS__",
"-U__CUDA_NO_BFLOAT162_OPERATORS__",
"-U__CUDA_NO_BFLOAT162_CONVERSIONS__",
"-DCUTE_ARCH_MMA_SM90A_ENABLE",
"--expt-relaxed-constexpr",
"--expt-extended-lambda",
"--use_fast_math",
"-lineinfo",
"-DCUTLASS_DEBUG_TRACE_LEVEL=0",
"-maxrregcount=50",
"-gencode=arch=compute_90a,code=sm_90a",
"-DNDEBUG",
]
+ gencode_flags,
},
),
)


def setup_token_dispatcher_utils():
from paddle.utils.cpp_extension import CUDAExtension, setup

change_pwd()
setup(
name="TokenDispatcherUtils",
ext_modules=CUDAExtension(
sources=[
"token_dispatcher_utils/topk_to_multihot.cu",
"token_dispatcher_utils/topk_to_multihot_grad.cu",
"token_dispatcher_utils/tokens_unzip_and_zip.cu",
"token_dispatcher_utils/tokens_stable_unzip.cu",
"token_dispatcher_utils/tokens_guided_unzip.cu",
"token_dispatcher_utils/regroup_tokens.cu",
],
extra_compile_args={
"cxx": ["-O3", "-w", "-Wno-abi", "-fPIC", "-std=c++17"],
"nvcc": [
"-O3",
"-U__CUDA_NO_HALF_OPERATORS__",
"-U__CUDA_NO_HALF_CONVERSIONS__",
"-U__CUDA_NO_BFLOAT16_OPERATORS__",
"-U__CUDA_NO_BFLOAT16_CONVERSIONS__",
"-U__CUDA_NO_BFLOAT162_OPERATORS__",
"-U__CUDA_NO_BFLOAT162_CONVERSIONS__",
"-DCUTE_ARCH_MMA_SM90A_ENABLE",
"--expt-relaxed-constexpr",
"--expt-extended-lambda",
"--use_fast_math",
"-maxrregcount=80",
"-lineinfo",
"-DCUTLASS_DEBUG_TRACE_LEVEL=0",
"-gencode=arch=compute_90a,code=sm_90a",
"-DNDEBUG",
],
},
),
)


run(setup_token_dispatcher_utils)
run(setup_fused_quant_ops)