[Bug]:编译paddlenlp_ops报错

[Jakin-huang]

软件环境

paddle2onnx               1.2.3
paddlefsl                 1.1.0
paddlenlp                 3.0.0b1
paddleocr                 2.8.1
paddlepaddle              2.6.2
paddlepaddle-gpu          2.6.2

重复问题

• I have searched the existing issues

错误描述

self.initialize_options()
[2024-09-26 17:32:48,516] [    INFO] dist.py:985 - running bdist_egg
[2024-09-26 17:32:48,549] [    INFO] dist.py:985 - running egg_info
[2024-09-26 17:32:48,558] [    INFO] egg_info.py:668 - writing paddlenlp_ops.egg-info/PKG-INFO
[2024-09-26 17:32:48,559] [    INFO] egg_info.py:292 - writing dependency_links to paddlenlp_ops.egg-info/dependency_links.txt
[2024-09-26 17:32:48,559] [    INFO] egg_info.py:292 - writing top-level names to paddlenlp_ops.egg-info/top_level.txt
[2024-09-26 17:32:48,569] [    INFO] sdist.py:190 - reading manifest file 'paddlenlp_ops.egg-info/SOURCES.txt'
[2024-09-26 17:32:48,571] [    INFO] util.py:328 - writing manifest file 'paddlenlp_ops.egg-info/SOURCES.txt'
[2024-09-26 17:32:48,571] [    INFO] bdist_egg.py:161 - installing library code to build/paddlenlp_ops/bdist.linux-x86_64/egg
[2024-09-26 17:32:48,571] [    INFO] dist.py:985 - running install_lib
[2024-09-26 17:32:48,571] [    INFO] dist.py:985 - running build_ext
Compiling user custom op, it will cost a few seconds.....
[2024-09-26 17:32:48,733] [    INFO] build_ext.py:533 - building 'paddlenlp_ops' extension
[2024-09-26 17:32:48,734] [    INFO] spawn.py:38 - /usr/local/cuda/bin/nvcc -I/root/anaconda3/envs/paddle-nlp/lib/python3.10/site-packages/paddle/include -I/root/anaconda3/envs/paddle-nlp/lib/python3.10/site-packages/paddle/include/third_party -I/usr/local/cuda/include -I/root/anaconda3/envs/paddle-nlp/include/python3.10 -I/root/anaconda3/envs/paddle-nlp/include/python3.10 -c /data01/hyq/PaddleNLP/csrc/gpu/dequant_int8.cu -o /data01/hyq/PaddleNLP/csrc/build/paddlenlp_ops/lib.linux-x86_64-cpython-310/dequant_int8.cu.o -DPADDLE_WITH_CUDA -DEIGEN_USE_GPU -ccbin cc -Xcompiler -fPIC --expt-relaxed-constexpr -DNVCC -gencode arch=compute_80,code=sm_80 -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__ -Igpu -Igpu/cutlass_kernels -Igpu/fp8_gemm_with_cutlass -Igpu/cutlass_kernels/fp8_gemm_fused/autogen -Ithird_party/cutlass/include -Ithird_party/nlohmann_json/single_include -Igpu/sample_kernels -w -DPADDLE_WITH_CUSTOM_KERNEL -D_GLIBCXX_USE_CXX11_ABI=1 -std=c++14
[2024-09-26 17:33:09,976] [    INFO] spawn.py:38 - /usr/local/cuda/bin/nvcc -I/root/anaconda3/envs/paddle-nlp/lib/python3.10/site-packages/paddle/include -I/root/anaconda3/envs/paddle-nlp/lib/python3.10/site-packages/paddle/include/third_party -I/usr/local/cuda/include -I/root/anaconda3/envs/paddle-nlp/include/python3.10 -I/root/anaconda3/envs/paddle-nlp/include/python3.10 -c /data01/hyq/PaddleNLP/csrc/gpu/encode_rotary_qk.cu -o /data01/hyq/PaddleNLP/csrc/build/paddlenlp_ops/lib.linux-x86_64-cpython-310/encode_rotary_qk.cu.o -DPADDLE_WITH_CUDA -DEIGEN_USE_GPU -ccbin cc -Xcompiler -fPIC --expt-relaxed-constexpr -DNVCC -gencode arch=compute_80,code=sm_80 -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__ -Igpu -Igpu/cutlass_kernels -Igpu/fp8_gemm_with_cutlass -Igpu/cutlass_kernels/fp8_gemm_fused/autogen -Ithird_party/cutlass/include -Ithird_party/nlohmann_json/single_include -Igpu/sample_kernels -w -DPADDLE_WITH_CUSTOM_KERNEL -D_GLIBCXX_USE_CXX11_ABI=1 -std=c++14
[2024-09-26 17:33:31,497] [    INFO] spawn.py:38 - gcc -pthread -B /root/anaconda3/envs/paddle-nlp/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /root/anaconda3/envs/paddle-nlp/include -fPIC -O2 -isystem /root/anaconda3/envs/paddle-nlp/include -fPIC -I/root/anaconda3/envs/paddle-nlp/lib/python3.10/site-packages/paddle/include -I/root/anaconda3/envs/paddle-nlp/lib/python3.10/site-packages/paddle/include/third_party -I/usr/local/cuda/include -I/root/anaconda3/envs/paddle-nlp/include/python3.10 -I/root/anaconda3/envs/paddle-nlp/include/python3.10 -c /data01/hyq/PaddleNLP/csrc/gpu/flash_attn_bwd.cc -o /data01/hyq/PaddleNLP/csrc/build/paddlenlp_ops/lib.linux-x86_64-cpython-310/flash_attn_bwd.o -O3 -w -DPADDLE_WITH_CUSTOM_KERNEL -D_GLIBCXX_USE_CXX11_ABI=1 -DPADDLE_WITH_CUDA -std=c++14
[2024-09-26 17:33:39,460] [    INFO] spawn.py:38 - /usr/local/cuda/bin/nvcc -I/root/anaconda3/envs/paddle-nlp/lib/python3.10/site-packages/paddle/include -I/root/anaconda3/envs/paddle-nlp/lib/python3.10/site-packages/paddle/include/third_party -I/usr/local/cuda/include -I/root/anaconda3/envs/paddle-nlp/include/python3.10 -I/root/anaconda3/envs/paddle-nlp/include/python3.10 -c /data01/hyq/PaddleNLP/csrc/gpu/fused_get_rope.cu -o /data01/hyq/PaddleNLP/csrc/build/paddlenlp_ops/lib.linux-x86_64-cpython-310/fused_get_rope.cu.o -DPADDLE_WITH_CUDA -DEIGEN_USE_GPU -ccbin cc -Xcompiler -fPIC --expt-relaxed-constexpr -DNVCC -gencode arch=compute_80,code=sm_80 -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__ -Igpu -Igpu/cutlass_kernels -Igpu/fp8_gemm_with_cutlass -Igpu/cutlass_kernels/fp8_gemm_fused/autogen -Ithird_party/cutlass/include -Ithird_party/nlohmann_json/single_include -Igpu/sample_kernels -w -DPADDLE_WITH_CUSTOM_KERNEL -D_GLIBCXX_USE_CXX11_ABI=1 -std=c++14
[2024-09-26 17:33:58,618] [    INFO] spawn.py:38 - gcc -pthread -B /root/anaconda3/envs/paddle-nlp/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /root/anaconda3/envs/paddle-nlp/include -fPIC -O2 -isystem /root/anaconda3/envs/paddle-nlp/include -fPIC -I/root/anaconda3/envs/paddle-nlp/lib/python3.10/site-packages/paddle/include -I/root/anaconda3/envs/paddle-nlp/lib/python3.10/site-packages/paddle/include/third_party -I/usr/local/cuda/include -I/root/anaconda3/envs/paddle-nlp/include/python3.10 -I/root/anaconda3/envs/paddle-nlp/include/python3.10 -c /data01/hyq/PaddleNLP/csrc/gpu/get_output.cc -o /data01/hyq/PaddleNLP/csrc/build/paddlenlp_ops/lib.linux-x86_64-cpython-310/get_output.o -O3 -w -DPADDLE_WITH_CUSTOM_KERNEL -D_GLIBCXX_USE_CXX11_ABI=1 -DPADDLE_WITH_CUDA -std=c++14
[2024-09-26 17:34:05,474] [    INFO] spawn.py:38 - /usr/local/cuda/bin/nvcc -I/root/anaconda3/envs/paddle-nlp/lib/python3.10/site-packages/paddle/include -I/root/anaconda3/envs/paddle-nlp/lib/python3.10/site-packages/paddle/include/third_party -I/usr/local/cuda/include -I/root/anaconda3/envs/paddle-nlp/include/python3.10 -I/root/anaconda3/envs/paddle-nlp/include/python3.10 -c /data01/hyq/PaddleNLP/csrc/gpu/get_padding_offset.cu -o /data01/hyq/PaddleNLP/csrc/build/paddlenlp_ops/lib.linux-x86_64-cpython-310/get_padding_offset.cu.o -DPADDLE_WITH_CUDA -DEIGEN_USE_GPU -ccbin cc -Xcompiler -fPIC --expt-relaxed-constexpr -DNVCC -gencode arch=compute_80,code=sm_80 -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__ -Igpu -Igpu/cutlass_kernels -Igpu/fp8_gemm_with_cutlass -Igpu/cutlass_kernels/fp8_gemm_fused/autogen -Ithird_party/cutlass/include -Ithird_party/nlohmann_json/single_include -Igpu/sample_kernels -w -DPADDLE_WITH_CUSTOM_KERNEL -D_GLIBCXX_USE_CXX11_ABI=1 -std=c++14
[2024-09-26 17:34:25,606] [    INFO] spawn.py:38 - /usr/local/cuda/bin/nvcc -I/root/anaconda3/envs/paddle-nlp/lib/python3.10/site-packages/paddle/include -I/root/anaconda3/envs/paddle-nlp/lib/python3.10/site-packages/paddle/include/third_party -I/usr/local/cuda/include -I/root/anaconda3/envs/paddle-nlp/include/python3.10 -I/root/anaconda3/envs/paddle-nlp/include/python3.10 -c /data01/hyq/PaddleNLP/csrc/gpu/get_padding_offset_v2.cu -o /data01/hyq/PaddleNLP/csrc/build/paddlenlp_ops/lib.linux-x86_64-cpython-310/get_padding_offset_v2.cu.o -DPADDLE_WITH_CUDA -DEIGEN_USE_GPU -ccbin cc -Xcompiler -fPIC --expt-relaxed-constexpr -DNVCC -gencode arch=compute_80,code=sm_80 -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__ -Igpu -Igpu/cutlass_kernels -Igpu/fp8_gemm_with_cutlass -Igpu/cutlass_kernels/fp8_gemm_fused/autogen -Ithird_party/cutlass/include -Ithird_party/nlohmann_json/single_include -Igpu/sample_kernels -w -DPADDLE_WITH_CUSTOM_KERNEL -D_GLIBCXX_USE_CXX11_ABI=1 -std=c++14
[2024-09-26 17:34:39,265] [    INFO] spawn.py:38 - /usr/local/cuda/bin/nvcc -I/root/anaconda3/envs/paddle-nlp/lib/python3.10/site-packages/paddle/include -I/root/anaconda3/envs/paddle-nlp/lib/python3.10/site-packages/paddle/include/third_party -I/usr/local/cuda/include -I/root/anaconda3/envs/paddle-nlp/include/python3.10 -I/root/anaconda3/envs/paddle-nlp/include/python3.10 -c /data01/hyq/PaddleNLP/csrc/gpu/qkv_transpose_split.cu -o /data01/hyq/PaddleNLP/csrc/build/paddlenlp_ops/lib.linux-x86_64-cpython-310/qkv_transpose_split.cu.o -DPADDLE_WITH_CUDA -DEIGEN_USE_GPU -ccbin cc -Xcompiler -fPIC --expt-relaxed-constexpr -DNVCC -gencode arch=compute_80,code=sm_80 -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__ -Igpu -Igpu/cutlass_kernels -Igpu/fp8_gemm_with_cutlass -Igpu/cutlass_kernels/fp8_gemm_fused/autogen -Ithird_party/cutlass/include -Ithird_party/nlohmann_json/single_include -Igpu/sample_kernels -w -DPADDLE_WITH_CUSTOM_KERNEL -D_GLIBCXX_USE_CXX11_ABI=1 -std=c++14
[2024-09-26 17:35:00,601] [    INFO] spawn.py:38 - /usr/local/cuda/bin/nvcc -I/root/anaconda3/envs/paddle-nlp/lib/python3.10/site-packages/paddle/include -I/root/anaconda3/envs/paddle-nlp/lib/python3.10/site-packages/paddle/include/third_party -I/usr/local/cuda/include -I/root/anaconda3/envs/paddle-nlp/include/python3.10 -I/root/anaconda3/envs/paddle-nlp/include/python3.10 -c /data01/hyq/PaddleNLP/csrc/gpu/quant_int8.cu -o /data01/hyq/PaddleNLP/csrc/build/paddlenlp_ops/lib.linux-x86_64-cpython-310/quant_int8.cu.o -DPADDLE_WITH_CUDA -DEIGEN_USE_GPU -ccbin cc -Xcompiler -fPIC --expt-relaxed-constexpr -DNVCC -gencode arch=compute_80,code=sm_80 -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__ -Igpu -Igpu/cutlass_kernels -Igpu/fp8_gemm_with_cutlass -Igpu/cutlass_kernels/fp8_gemm_fused/autogen -Ithird_party/cutlass/include -Ithird_party/nlohmann_json/single_include -Igpu/sample_kernels -w -DPADDLE_WITH_CUSTOM_KERNEL -D_GLIBCXX_USE_CXX11_ABI=1 -std=c++14
[2024-09-26 17:35:24,001] [    INFO] spawn.py:38 - /usr/local/cuda/bin/nvcc -I/root/anaconda3/envs/paddle-nlp/lib/python3.10/site-packages/paddle/include -I/root/anaconda3/envs/paddle-nlp/lib/python3.10/site-packages/paddle/include/third_party -I/usr/local/cuda/include -I/root/anaconda3/envs/paddle-nlp/include/python3.10 -I/root/anaconda3/envs/paddle-nlp/include/python3.10 -c /data01/hyq/PaddleNLP/csrc/gpu/rebuild_padding.cu -o /data01/hyq/PaddleNLP/csrc/build/paddlenlp_ops/lib.linux-x86_64-cpython-310/rebuild_padding.cu.o -DPADDLE_WITH_CUDA -DEIGEN_USE_GPU -ccbin cc -Xcompiler -fPIC --expt-relaxed-constexpr -DNVCC -gencode arch=compute_80,code=sm_80 -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__ -Igpu -Igpu/cutlass_kernels -Igpu/fp8_gemm_with_cutlass -Igpu/cutlass_kernels/fp8_gemm_fused/autogen -Ithird_party/cutlass/include -Ithird_party/nlohmann_json/single_include -Igpu/sample_kernels -w -DPADDLE_WITH_CUSTOM_KERNEL -D_GLIBCXX_USE_CXX11_ABI=1 -std=c++14
[2024-09-26 17:35:44,299] [    INFO] spawn.py:38 - /usr/local/cuda/bin/nvcc -I/root/anaconda3/envs/paddle-nlp/lib/python3.10/site-packages/paddle/include -I/root/anaconda3/envs/paddle-nlp/lib/python3.10/site-packages/paddle/include/third_party -I/usr/local/cuda/include -I/root/anaconda3/envs/paddle-nlp/include/python3.10 -I/root/anaconda3/envs/paddle-nlp/include/python3.10 -c /data01/hyq/PaddleNLP/csrc/gpu/rebuild_padding_v2.cu -o /data01/hyq/PaddleNLP/csrc/build/paddlenlp_ops/lib.linux-x86_64-cpython-310/rebuild_padding_v2.cu.o -DPADDLE_WITH_CUDA -DEIGEN_USE_GPU -ccbin cc -Xcompiler -fPIC --expt-relaxed-constexpr -DNVCC -gencode arch=compute_80,code=sm_80 -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__ -Igpu -Igpu/cutlass_kernels -Igpu/fp8_gemm_with_cutlass -Igpu/cutlass_kernels/fp8_gemm_fused/autogen -Ithird_party/cutlass/include -Ithird_party/nlohmann_json/single_include -Igpu/sample_kernels -w -DPADDLE_WITH_CUSTOM_KERNEL -D_GLIBCXX_USE_CXX11_ABI=1 -std=c++14
[2024-09-26 17:36:05,960] [    INFO] spawn.py:38 - /usr/local/cuda/bin/nvcc -I/root/anaconda3/envs/paddle-nlp/lib/python3.10/site-packages/paddle/include -I/root/anaconda3/envs/paddle-nlp/lib/python3.10/site-packages/paddle/include/third_party -I/usr/local/cuda/include -I/root/anaconda3/envs/paddle-nlp/include/python3.10 -I/root/anaconda3/envs/paddle-nlp/include/python3.10 -c /data01/hyq/PaddleNLP/csrc/gpu/sample_kernels/top_p_sampling_reject.cu -o /data01/hyq/PaddleNLP/csrc/build/paddlenlp_ops/lib.linux-x86_64-cpython-310/top_p_sampling_reject.cu.o -DPADDLE_WITH_CUDA -DEIGEN_USE_GPU -ccbin cc -Xcompiler -fPIC --expt-relaxed-constexpr -DNVCC -gencode arch=compute_80,code=sm_80 -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__ -Igpu -Igpu/cutlass_kernels -Igpu/fp8_gemm_with_cutlass -Igpu/cutlass_kernels/fp8_gemm_fused/autogen -Ithird_party/cutlass/include -Ithird_party/nlohmann_json/single_include -Igpu/sample_kernels -w -DPADDLE_WITH_CUSTOM_KERNEL -D_GLIBCXX_USE_CXX11_ABI=1 -std=c++14
gpu/sample_kernels/sampling.cuh(392): error: namespace "std" has no member "gcd"
    const uint32_t vec_size = std::gcd(16 / sizeof(T), d);
                                   ^

稳定复现步骤 & 代码

编译按照readme步骤进行。

[DrownFish19]

辛苦提供编译环境下的具体gcc和cmake版本或者采用Paddle docker镜像进行编译安装。

[Jakin-huang]

辛苦提供编译环境下的具体gcc和cmake版本或者采用Paddle docker镜像进行编译安装。

gcc (GCC) 10.1.0
cmake version 3.22.2

[DrownFish19]

辛苦提供编译环境下的具体gcc和cmake版本或者采用Paddle docker镜像进行编译安装。

gcc (GCC) 10.1.0 cmake version 3.22.2

这个版本应该是能正常编译的，是否是因为使用anaconda或者系统中存在多个版本gcc的原因，编译时未使用高版本的gcc。