Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Teddy example crashes when using Nvidia Quadro K420 with 1GB memory #42

Closed
hling opened this issue Mar 10, 2016 · 3 comments
Closed

Teddy example crashes when using Nvidia Quadro K420 with 1GB memory #42

hling opened this issue Mar 10, 2016 · 3 comments

Comments

@hling
Copy link

hling commented Mar 10, 2016

When I compile InfiniTAM with CPU setting on Ubuntu Linux 14.04, I can successfully run the Teddy example. However, when I compile it with CUDA 7.5 and run it with an Nvidia Quadro K420 card (with 1GB memory) on the same machine, the program crashes with an out of memory error as follows:
InfiniTAM/ITMLib/Engine/../Utils/../Objects/../../ORUtils/MemoryBlock.h(217) : cudaSafeCall() Runtime API error : out of memory.

Has anyone else seen the same problem? Do you have any suggestion to resolve it? I am pretty sure my CUDA drive is installed correctly. I successfully ran all the NVIDIA_CUDA-7.5 Samples and saw the GPU version gains significant speedup compared to the CPU version on all sample programs.

BTW, I use ./InfiniTAM Teddy/calib.txt Teddy/Frames/%04i.ppm Teddy/Frames/%04i.pgm command to run the Teddy example.

Thanks!

@sgolodetz
Copy link
Collaborator

You could try changing the settings in ITMVoxelBlockHash.h - if it's just a case of it running out of memory then decreasing the number of stored blocks should help.

@hling
Copy link
Author

hling commented Mar 11, 2016

Stuart, thanks for your quick response. Yes, it is working now after changing the parameters that are set in ../Utils/ITMLibDefines.h.

@sgolodetz
Copy link
Collaborator

Glad that helped (and sorry for the wrong file reference - I was looking at the latest internal branch rather than the public one).

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants