phvu / cuda-smi Goto Github PK
View Code? Open in Web Editor NEWSmall utility to check CUDA memory on Mac
Small utility to check CUDA memory on Mac
Then trying to open it in terminal it gives me an error:
cuda-smi: command not found
nvcc warning : The 'compute_20', 'sm_20', and 'sm_21' architectures are deprecated, and may be
removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc fatal : The version ('90000') of the host compiler ('Apple clang') is not supported
Thank you very much for this little command line tool – it comes in very handy if you run deep learning projects on your GPU. It's always good to know how much VRAM is available.
What i always was looking for is an OSX GPU monitoring App – where you can see the GPU memory usage for every single process. I'm not going to ask you to write this – don't worry :-). But do you know if this is even possible to write? Is there a Nvidia driver API that could give you detailled info about the VRAM usage?
Could you suggest a documentation where i could look for a solution to this ? I can write Swift and a little Obj-C – i just need a starting point where i can extract the info from.
Thanks for reading!
I got an error saying, Error: CUDA driver version is insufficient for CUDA runtime version in cudaGetDeviceCount at ./cuda-smi.cpp:23
.
I use CUDA 9.2 with driver 396.148 under macOS 10.13.6.
Please have this issue taken a look.
Thanks in advance.
hi @phvu 👋
it would be nice to make a release that contains the binary, to remove the build requirement (since the build fails for people with Apple LLVM version 9.0.0 (clang-900.0.38)
#4
I've made a release on my fork https://github.com/micahstubbs/cuda-smi/releases/tag/v1.0
but of course it would be nice to have a release your main cuda-smi repo 😄
♦ cuda-smi-1.0 / ➞ ./cuda-smi
^C
♦ cuda-smi-1.0 / ➞ sudo ./cuda-smi
Password:
^C
nvcc is installed:
♦ cuda-smi-1.0 / ➞ nvcc --version
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2017 NVIDIA Corporation
Built on Fri_Nov__3_21:08:57_CDT_2017
Cuda compilation tools, release 9.1, V9.1.85
I was able to compile it but after waiting for a minute without output, I had to Ctrl+C and abort.
Every time I run make cuda-smi
I get the following error:
cuda-smi.cpp:4:10: fatal error: 'cuda_runtime_api.h' file not found
Any ideas why this might be?
I'm on a OSX 10.12.6 with a GTX 980 Ti
Thanks!
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.