search

weiliu89 / caffe

Caffe: a fast open framework for deep learning.
http://caffe.berkeleyvision.org/
Other
4.77k stars 1.68k forks source link

Check failed: error == cudaSuccess (8 vs. 0) invalid device function #336

Open yghcats opened 7 years ago

yghcats commented 7 years ago

Hi, I'm performing: python examples/ssd/ssd_pascal.py this shell command

error：im2col.cu:61] Check failed: error == cudaSuccess (8 vs. 0) invalid device function

The system is Ubuntu 14.04, with one GPU Can anyone give me some advice, thank you！

eonurk commented 7 years ago

What is your GPU?

yghcats commented 7 years ago

@eonurk Has been done, is due to the lack of ability to calculate the graphics（GeForce GTX TiTan X） card.

Modify Makefile.config solution

Thank you！

michael5lin commented 7 years ago

@yghcats even TiTan X cannot do the jobs? no need to change parameter in makefile.config if your ablility is enough

xhzzhang commented 6 years ago

Hello, I'm performing: python examples/ssd/ssd_pascal_video.py(or python examples/ssd/ssd_pascal_webcam.py): got below error message:

F0301 12:35:45.320708 14370 pooling_layer.cu:212] Check failed: error == cudaSuccess (8 vs. 0) invalid device function Check failure stack trace: @ 0x7f83914b94 (unknown) @ 0x7f83914aac (unknown) @ 0x7f83914500 (unknown) @ 0x7f8391717c (unknown) @ 0x7f83dea394 caffe::PoolingLayer<>::Forward_gpu() @ 0x7f83beab18 caffe::Net<>::ForwardFromTo() @ 0x7f83beae04 caffe::Net<>::Forward() @ 0x408dec test() @ 0x4074a0 main @ 0x7f8353de08 (unknown) Aborted

And my GPU is Nvidia TX1 and TX2.

xhzzhang commented 6 years ago

I tested add this code into Makefile.config in TX2, then can fix it:

CUDA_ARCH := -gencode arch=compute_62,code=[sm_62,compute_62]