编译安装v2.3.2报错

[kisaragychihaya]

问题描述 Issue Description

[478/2763] Building CUDA object paddle\phi\kernels\CMakeFiles\cross_entropy_kernel.dir\gpu\cross_entropy_kernel.cu.obj
FAILED: paddle/phi/kernels/CMakeFiles/cross_entropy_kernel.dir/gpu/cross_entropy_kernel.cu.obj
C:\PROGRA~1\NVIDIA~2\CUDA\v11.3\bin\nvcc.exe -forward-unknown-to-host-compiler -DBOOST_HAS_STATIC_ASSERT -DCUDA_TOOLKIT_ROOT_DIR="\"C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v11.3\"" -DCUDA_VERSION_MAJOR=\"11\" -DCUDA_VERSION_MINOR=\"3\" -DCUDNN_MAJOR_VERSION=\"8\" -DEIGEN_STRONG_INLINE=inline -DEIGEN_USE_GPU -DGOOGLE_GLOG_DLL_DECL="" -DNEW_RELEASE_PYPI -DNOMINMAX -DPADDLE_DISABLE_PROFILER -DPADDLE_DLL_EXPORT -DPADDLE_DLL_INFERENCE -DPADDLE_ON_INFERENCE -DPADDLE_USE_OPENBLAS -DPADDLE_VERSION=2.3.2 -DPADDLE_VERSION_INTEGER=2003002 -DPADDLE_WITH_AVX -DPADDLE_WITH_CRYPTO -DPADDLE_WITH_CUDA -DPADDLE_WITH_POCKETFFT -DPADDLE_WITH_SSE3 -DPADDLE_WITH_TENSORRT -DTRT_PLUGIN_FP16_AVALIABLE -DUTF8PROC_STATIC -D_XKEYCHECK_H -IE:\tmp\Paddle\build -IE:\tmp\Paddle\paddle\fluid\framework\io -IE:\TensorRT-8.2.3.0\include -IE:\tmp\Paddle\build\third_party\install\zlib\include -IE:\tmp\Paddle\build\third_party\install -IE:\tmp\Paddle\build\third_party\install\gflags\include -IE:\tmp\Paddle\build\third_party\install\glog\include -IE:\tmp\Paddle\build\third_party\boost\src\extern_boost -IE:\tmp\Paddle\build\third_party\eigen3\src\extern_eigen3 -IE:\tmp\Paddle\build\third_party\threadpool\src\extern_threadpool -IE:\tmp\Paddle\build\third_party\dlpack\src\extern_dlpack\include -IE:\tmp\Paddle\build\third_party\install\xxhash\include -IE:\tmp\Paddle\build\third_party\install\warpctc\include -IE:\tmp\Paddle\build\third_party\install\utf8proc\include -IE:\tmp\Paddle\build\third_party\install\openblas\include\openblas -IE:\tmp\Paddle\build\third_party\install\protobuf\include -IC:\Python38\include -IC:\Python38\Lib\site-packages\numpy\core\include -IE:\tmp\Paddle\build\third_party\pybind\src\extern_pybind\include -IE:\tmp\Paddle\build\third_party\install\cryptopp\include -IE:\tmp\Paddle\build\third_party\pocketfft\src -IE:\tmp\Paddle\build\third_party\dirent\src\extern_dirent\include -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.3\include" -IE:\tmp\Paddle -IE:\tmp\Paddle\build\..\paddle\fluid\framework\io -D_WINDOWS -Xcompiler=" /W0  /GR /EHsc" -D_MWAITXINTRIN_H_INCLUDED -D__STRICT_ANSI__ -Wno-deprecated-gpu-targets  -gencode arch=compute_61,code=sm_61 -gencode arch=compute_72,code=sm_72 -gencode arch=compute_75,code=sm_75 -gencode arch=compute_86,code=sm_86 -w --expt-relaxed-constexpr --expt-extended-lambda -Xcompiler "/wd4244 /wd4267 /wd4819 " -Xcompiler /bigobj  -Xcompiler="/arch:AVX" -Xcompiler="-MT -O2 -Ob2" -DNDEBUG -std=c++14 -MD -MT paddle\phi\kernels\CMakeFiles\cross_entropy_kernel.dir\gpu\cross_entropy_kernel.cu.obj -MF paddle\phi\kernels\CMakeFiles\cross_entropy_kernel.dir\gpu\cross_entropy_kernel.cu.obj.d -x cu -c E:\tmp\Paddle\paddle\phi\kernels\gpu\cross_entropy_kernel.cu -o paddle\phi\kernels\CMakeFiles\cross_entropy_kernel.dir\gpu\cross_entropy_kernel.cu.obj -Xcompiler=-Fdpaddle\phi\kernels\CMakeFiles\cross_entropy_kernel.dir\cross_entropy_kernel.pdb,-FS

Wrote crash dump file "C:\Users\chen\AppData\Local\Temp\cicc.exe-9bc143.dmp"
0x00007FF64B448D80 (0x000001D5703FC6F0 0x0000000000000000 0x0000000000000000 0x0000000000000000)
0x00007FF64B44701B (0x0010248000000000 0x000001D5703FCA40 0x00000032000C8A51 0x0000000000000000)
0x00007FF64B48ECF6 (0x000001D5703FC860 0x0000000000000000 0x00000057617F9F21 0x0000000000000000)
0x00007FF64B3B394A (0x000001D5703FCEA8 0x0000000000000000 0x00000057617FA148 0x0000000000000000)
0x00007FF64B39EFC9 (0x0000000000000000 0x0000000000000000 0x00000057617FA274 0x0000000200000000)
0x00007FF64B39F649 (0x0000000000000000 0x0000000000000000 0x00000057617FA300 0x00000003000C8A51)
0x00007FF64B3A2821 (0x000001D500000000 0x0000000000000000 0x0000000000000000 0x000001D500000001)
0x00007FF64B294BD6 (0x00000003000C8A51 0x00000057617FA5D0 0x0000000008080002 0x0000000000000000)
0x00007FF64B290935 (0x00007FF608080002 0x0000000000000000 0x0000000000000000 0x0000000000000000)
0x00007FF64B287CCE (0x0000000000000000 0x000001D518A34270 0x000001D518A40058 0x0000000000000000)
0x00007FF64B28633D (0x000001D5703FC860 0x00000057617FA881 0x0000000000000000 0x0000000000000000)
0x00007FF64B3B313C (0x000001D5703FCEA8 0x0000000000000000 0x00000057617FAAA8 0x0000000000000000)
0x00007FF64B39EFC9 (0x0000000000000000 0x0000000000000000 0x00000057617FABD4 0x0000000200000000)
0x00007FF64B39F649 (0x0000000000000000 0x0000000000000000 0x00000057617FAC60 0x00000002000C8A50)
0x00007FF64B3A2821 (0x0000000000000000 0x0000000000000000 0x0000005700000000 0x0000000000000001)
0x00007FF64B23DE9B (0x0000000C000C8A57 0x0000000C000C8A57 0x0000000C000C8A57 0x0000000C000C8A57)
0x00007FF64B47D7DF (0x0000002400000000 0x0000005700000000 0x000000000001622B 0x00007FF600000000)
0x00007FF64B21B68B (0x0000000000000000 0x000001D5189CE438 0x00000024000C8A57 0x00000057617FB729)
0x00007FF64B23C195 (0x00000057617FB820 0x00007FF64B456D00 0x0000000000000001 0x00007FF600000000)
0x00007FF64B23D009 (0x0000000C000C9000 0x0000000C000C9000 0x0000000C000C9000 0x0000000C000C9000)
0x00007FF64B47D7DF (0x0000003D00000000 0x0000005700000000 0x0000000000016236 0x00007FF600000000)
0x00007FF64B21B68B (0x0000000000000000 0x000001D5189CE170 0x0000003D000C9000 0x00000057617FC1C9)
0x00007FF64B23C195 (0x00000057617FC2C0 0x00000057617FC400 0x00000057617F0001 0x00007FF600000000)
0x00007FF64B23D009 (0x000000140013D8F0 0x0000001A0013D8F0 0x0000001A0013D8F0 0x0000001A0013D8F0)
0x00007FF64B47D7DF (0x0000005700000000 0x0000000000000000 0x00000000000163B1 0x0000005700000000)
0x00007FF64B21DA7B (0x00000057617FCDD0 0x0000000000000000 0x00000057617FD038 0x0000000000000000)
0x00007FF64B22D2E3 (0x0000000000000000 0x00007FF600000000 0x000001D500000000 0x00000057617FD1A0)
0x00007FF64B240883 (0x00000057617FD1A0 0x0000000000000000 0x00000057617FCDC0 0x00000057617FCD60)
0x00007FF64B2259FF (0x0000000000000000 0x0000000000000000 0x0000000000000000 0x0000000000000001)
0x00007FF64B23E019 (0x00000057617FDC40 0x0000000000000000 0x000001D5189CC978 0x00000057617FDC40)
0x00007FF64B28E0CC (0x00000057617FD8A0 0x00000057617FD9C0 0x000001D5752B1891 0x00007FF64B3AF01B)
0x00007FF64B2917EB (0x0000005708A008B3 0x0000000000000000 0x00000057617FDB70 0x0000000000000000)
0x00007FF64B2855C7 (0x0000000000000000 0x0000000000000000 0x0000000000000000 0x0000000100000001)
0x00007FF64B27AC21 (0x000000150013D8A1 0x00000057617FDEF1 0x000001D57A126000 0x0000000000000000)
0x00007FF64B28105F (0x00000057617FE000 0x0000000000000000 0x0000000000000000 0x0000000000000000)
0x00007FF64B272D7B (0x00000057617FE0D0 0x0000000000000000 0x0000000000000000 0x0000000000000000)
0x00007FF64B2853F1 (0x0000000000000000 0x00007FF64B3BFF4A 0x0000000000000000 0x00007FF64B45F2F0)
0x00007FF64B27AC21 (0x000000120013D8A0 0x00000057617FE491 0x000001D5729749C8 0x0000000000000000)
0x00007FF64B28105F (0x00000057617FE5A0 0x0000000000000000 0x0000000000000000 0x0000000000000000)
0x00007FF64B272D7B (0x00000057617FE670 0x00000057617FEBD4 0x0000000000000000 0x0000000000000000)
0x00007FF64B2853F1 (0x000001D56C674238 0x0000000000000000 0x00000057617FEBD4 0x0000000100000AB4)
0x00007FF64B2875BE (0x000001D56C674238 0x000001D500000000 0x0000000000000000 0x000000000000001F)
0x00007FF64B4AA4D2 (0x000001D56C5EB560 0x00000057617FEB29 0x000001D56C5F7E80 0x000001D500000000)
0x00007FF64B207F91 (0x0000000000000000 0x000001D56C5EB560 0x00007FF64C1FB070 0x0000000000000000)
0x00007FF64B2080F9 (0x000000570000001F 0x000001D56C5EB560 0x000001D56C5F7E80 0x00000057617FEA38)
0x00007FF64B4D2E97 (0x000001D56C5F7E80 0x00007FF64CA5CAB8 0x0000000000000000 0x000001D56C606480)
0x00007FF64B4CB80F (0x00007FF64CA5CAB8 0x0000000000000295 0x0000000000000295 0x00007FF64CA5CAB8)
0x00007FF64B4D7C3F (0x00007FF64CA5CAB8 0x0000000000000000 0x000001D56C606480 0x0000000000000000)
0x00007FF64C0721C4 (0x0000000000000000 0x0000000000000000 0x0000000000000000 0x0000000000000000)
0x00007FFBC8DA7034 (0x0000000000000000 0x0000000000000000 0x0000000000000000 0x0000000000000000), BaseThreadInitThunk() + 0x14 bytes(s)
0x00007FFBC9F62651 (0x0000000000000000 0x0000000000000000 0x0000000000000000 0x0000000000000000), RtlUserThreadStart() + 0x21 bytes(s)
conv_grad_grad_kernel.cu
nvcc error   : 'cicc' died with status 0xC0000005 (ACCESS_VIOLATION)
[474/2763] Building CUDA object paddle\phi\kernels\CMakeFiles\full_kernel_base.dir\gpu\full_kernel.cu.obj
full_kernel.cu

CMakeCache.txt cicc.exe-9bc143.zip

版本&环境信息 Version & Environment Information

Paddle 源码Version 2.3.2 CUDA:11.3.58+cudnn8.2.1 系统 Win10 64bit 编译环境 VS2017+Ninja

---

---

Paddle version: None Paddle With CUDA: None

OS: Windows 10 Python version: 3.8.10

CUDA version: 11.3.58 Build cuda_11.3.r11.3/compiler.29745058_0 cuDNN version: None.None.None Nvidia driver version: 516.40

---

编译过程出错

[paddle-bot[bot]]

您好，我们已经收到了您的问题，会安排技术人员尽快解答您的问题，请耐心等待。请您再次检查是否提供了清晰的问题描述、复现代码、环境&版本、报错信息等。同时，您也可以通过查看官网API文档、常见问题、历史Issue、AI社区来寻求解答。祝您生活愉快～

Hi! We've received your issue and please be patient to get responded. We will arrange technicians to answer your questions as soon as possible. Please make sure that you have posted enough message to demo your request. You may also check out the API，FAQ，Github Issue and AI community to get the answer.Have a nice day!

[betterpig]

为复现问题，请提供编译使用的完整cmake命令

[kisaragychihaya]

为复现问题，请提供编译使用的完整cmake命令

cmake .. -GNinja -DWITH_MKL=OFF -DWITH_MKLDNN=OFF -DWITH_TESTING=OFF -DCMAKE_BUILD_TYPE=Release -DON_INFER=ON -DWITH_LITE=ON -DWITH_TENSORRT=ON -DTENSORRT_ROOT_DIR="E:\tmp\TensorRT-8.2.3.0" -DCUDA_ARCH_NAME="Manual" -DNEW_RELEASE_PYPI=ON -DCUDA_ARCH_BIN="61 72 75 86" -DWITH_UNITY_BUILD=ON

[leiqing1]

kisaragychihaya这边的软硬件环境是： AMD r9 5950X+ RTX3080 win10 64bit 1）需求反馈：现在家用电脑和云端不少锐龙和EPYC的机器 这些机器MKL应该都跑不了 2）现在需要个openblas+cuda11.6或者11.2的包试试看 预编译的那些在amd zen3架构cpu上面没法跑在cpu模式下，电脑上python38的paddle的pdi的cpu模式不是段错误就是avxcore出问题 只能用GPU模式完成开发了

[Ligoml]

@pangyoki 辛苦协助复现和定位一下问题

[zhwesky2010]

@kisaragychihaya 你好，CUDA11.3需要使用VS2019来编译，可以在 https://visualstudio.microsoft.com/zh-hans/vs/older-downloads/ 中下载vs2019社区版并安装

[paddle-bot[bot]]

Since you haven\'t replied for more than a year, we have closed this issue/pr. If the problem is not solved or there is a follow-up one, please reopen it at any time and we will continue to follow up. 由于您超过一年未回复，我们将关闭这个issue/pr。 若问题未解决或有后续问题，请随时重新打开，我们会继续跟进。