Ryan Stocks

> This is a bit more complicated than [wavefunction91/ExchCXX#39](https://github.com/wavefunction91/ExchCXX/pull/39), I'm happy to work with you on getting this fixed up, but the HIP implementation is GauXC is a bit delicate...

@wavefunction91 @ajaypanyala I have fixed the merge conflicts so that it successfully compiles with HIP, however it fails a lot of tests, would be great to get some more experienced...

@wavefunction91 hope Microsoft is treating you well! (Is this project likely to continue much development?) The HIP code seems to be correct for LDA functionals so I think I must...

@wavefunction91 thank you very much for the detailed response. I don't think we can check the grads on AMD as I get a "Generic GauXC Exception (LDA Grad NYI for...

I think I have now got gauxc running correctly with HIP on a NVIDIA machine - the problem was the vvar grad kernel which required the thread block to be...

@wavefunction91 I have finally had some time and have fixed up the HIP kernels so that they now work (though likely not as efficiently as the CUDA implementations). This is...

@mikovtun good pickup! I think you might be on to something - If I use CUDA on a 4080 I get exactly the same 0.0000001368 errors, however I do not...