torch / ezinstall

One-line install scripts for Torch.
75 stars 79 forks source link

no CUDA-capable device is detected #96

Closed mrtortong closed 8 years ago

mrtortong commented 8 years ago

็็Hi.

Please recommend.

root@tong:/desired/absolute/path/to/deepmask/pretrained/sharpmask# cd $DEEPMASK root@tong:/desired/absolute/path/to/deepmask# th computeProposals.lua $DEEPMASK/pretrained/deepmask # run DeepMask THCudaCheck FAIL file=/tmp/luarocks_cutorch-scm-1-3624/cutorch/lib/THC/THCGeneral.c line=20 error=38 : no CUDA-capable device is detected /usr/local/bin/luajit: /usr/local/share/lua/5.1/trepl/init.lua:384: cuda runtime error (38) : no CUDA-capable device is detected at /tmp/luarocks_cutorch-scm-1-3624/cutorch/lib/THC/THCGeneral.c:20 stack traceback: [C]: in function 'error' /usr/local/share/lua/5.1/trepl/init.lua:384: in function 'require' computeProposals.lua:11: in main chunk [C]: in function 'dofile' /usr/local/lib/luarocks/rocks/trepl/scm-1/bin/th:145: in main chunk [C]: at 0x00406260

Thank in advance. mrtortong.

albanD commented 8 years ago

no CUDA-capable device is detected do you have a CUDA GPU on your machine?

mrtortong commented 8 years ago

Hi albanD Please recommend. my notebook is http://specsnotebook.blogspot.com/2011/03/laptop-asus-a42j.html Video Type ATI Mobility Radeon HD 5470 1GB

it can use

CUDA-capable GPU A supported version of Linux with a gcc compiler and toolchain NVIDIA CUDA Toolkit (available at http://developer.nvidia.com/cuda-downloads)

Or not.

Thank you. mrtortong

albanD commented 8 years ago

You GPU is an ATI one. Thus does not support Nvidia's CUDA. Unfortunately all main torch GPU packages are based on CUDA.

If you need GPU acceleration for your own project, you should look into cltorch that is based on OpenCL and thus should work on your GPU.

mrtortong commented 8 years ago

albanD

           Thank you very much.If i have any question.please recommend me.

I am fresh developer

mrtortong

albanD commented 8 years ago

No problem. Also next time please open a SINGLE issue, don't open one in every single repo. It is quite annoying. Or even better, use the mailing list here https://groups.google.com/forum/#!forum/torch7