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

I can't get native linux MKL provider to work with MathNet v5 #946

Open
atlemann opened this issue Sep 1, 2022 · 2 comments
Open

I can't get native linux MKL provider to work with MathNet v5 #946

atlemann opened this issue Sep 1, 2022 · 2 comments

Comments

@atlemann
Copy link

atlemann commented Sep 1, 2022

The release notes says something about making it better for x-plat, but I'm unable to make native providers work in linux at all. Is anyone else able to?

Should it be possible to include both MathNet.Numerics.MKL.Linux-x64 and MathNet.Numerics.MKL.Win-x64 in my application now so I can run it on both platforms?

@petterton
Copy link

I have the same problem. When trying to call LinearAlgebraControl.UseNativeMKL() on Linux I get System.NotSupportedException: Native Provider Probing failed to resolve creator. Did you make this work @atlemann?

@atlemann
Copy link
Author

No, unfortunately not. The actual application has been running mostly in a Windows environment due to other reasons, although I develop on Linux. I didn't have time to test too much so I just gave up. But we want the app to run on linux as well, so I'll have to look into it again some time. It's a bit annoying how it works in MathNet, since AFAIK you cannot have both linux and windows native libs available, since the files have the same name. At least before MathNet v5.

We're building our own custom native Intel.MKL to get some other functions and we pack both windows and linux artifacts in the same NuGet package. If you call your artifacts "MyCoolLib.dll" and "libMyCoolLib.so", you can use DllImport("MyCoolLib") and .NET will find the correct file for each platform itself. Last time I checked it seemed MathNet was using a very complicated way of finding these libs.

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