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

building on windows #2

Open
RyanHope opened this issue Feb 7, 2017 · 8 comments
Open

building on windows #2

RyanHope opened this issue Feb 7, 2017 · 8 comments

Comments

@RyanHope
Copy link

RyanHope commented Feb 7, 2017

At one point I thought you had a windows binary available but maybe I am mistaken. Anyway, I am all setup with a msys2/mingw64 build environment and CUDA and visual studio 12. How do I get this to compile? I was able to hobble my way through the configure but then it still fails on the build because of sys/mman.h missing and unistd.h.

@fraimondo
Copy link
Owner

I've never even tried building it on windows.

It can be ported. I guess CUDA behaves the same, it's just a matter of low level IO functions.

@RyanHope
Copy link
Author

RyanHope commented Feb 9, 2017

Yeah I bet it is just a few lines of code. Let me know if there is anything I can do to help out... if you need any more detailed build logs.

@chrisblais
Copy link

Super interested in a version that runs on windows.

@RyanHope
Copy link
Author

Any update on a windows build attempt?

@fraimondo
Copy link
Owner

fraimondo commented Jan 16, 2018 via email

@CloudyDory
Copy link

I have successfully build CUDAICA on Windows 10 using Visual Studio 2015, and it seems running fast and well now.

@fraimondo If you have time, could you please help to review the code a little bit? Both the low-level IO part and the CUDA and BLAS part have issues, and I used some workarounds to bypass them. If you agree, I will upload the code on GitHub.

Thanks,

@fraimondo
Copy link
Owner

fraimondo commented Aug 23, 2018 via email

@CloudyDory
Copy link

Hello Fraimondo,

I have uploaded modified version of CUDAICA on https://github.com/yhz-1995/cudaica_win.

I find that the cublas part of "step3" in infomax.cu is very slow, On my machine (i7 7700HQ + Nvidia 1050Ti) a lot of time is spent on transform the triangular matrix "u*u'" into a dense matrix. Instead, your "step3<<<>>>" kernel is very fast, but somehow you disabled it. Once I reenable this kernel, I get a 5x speed up on my machine.

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

4 participants