search

apache / mxnet

Lightweight, Portable, Flexible Distributed/Mobile Deep Learning with Dynamic, Mutation-aware Dataflow Dep Scheduler; for Python, R, Julia, Scala, Go, Javascript and more
https://mxnet.apache.org
Apache License 2.0
20.76k stars 6.8k forks source link

Build Mxnet C++ failed with "nvcc fatal : Unsupported gpu architecture 'compute_92'" #13999

Open zhizec opened 5 years ago

zhizec commented 5 years ago

Hi

I want build mxnet c++ from source with GPU 1060 6g and cuda9.2 on Win10， but report:

Building NVCC (Device) object CMakeFiles/cuda_compile_1.dir/src/common/Release/cuda_compile_1_generated_random_generator.cu.obj nvcc fatal : Unsupported gpu architecture 'compute_92' CMake Error at cuda_compile_1_generated_random_generator.cu.obj.Release.cmake:222 (message): Error generating D:/study/mxnet_build/incubator-mxnet/build/CMakeFiles/cuda_compile_1.dir/src/common/Release/cuda_compile_1_generated_random_generator.cu.o

I think it's cuda version not matched with GPU arch in CMakeFiles. How can I solve this issue? Thank you!

mxnet-label-bot commented 5 years ago

Hey, this is the MXNet Label Bot. Thank you for submitting the issue! I will try and suggest some labels so that the appropriate MXNet community members can help resolve it. Here are my recommended labels: Installation, C++, Build

frankfliu commented 5 years ago

@mxnet-label-bot add [build, Windows, cuda]

zhizec commented 5 years ago

I try uninstall cuda9.2 and replace with cuda8.0, now can compile, but slow and finally report:

fatal error C1002: compiler is out of heap space in pass 2.

Do you know how to fix it? Seems other guys also saw this issue... Thank you.

access2rohit commented 5 years ago

Hi, Did you make changes the codebase before building ?

access2rohit commented 5 years ago

@zhizec ^

piyushghai commented 5 years ago

@zhizec I followed the instructions to build MXNet and I did not face the error. http://mxnet.incubator.apache.org/versions/master/install/windows_setup.html

Can you try following the steps for Build from Source on this page ?

IAGLivingInDream commented 3 years ago

GPU 1060 6g -> compute_61,sm_61