Shreyas Atre
Shreyas Atre
@gevtushenko Hello! Just a gentle ping about this PR. Thanks!
Substituting $ε$ in your equations for double precision i.e. $10^{-16}$ so that it's easier for me to read > The "classic" compensated summation approach of Kahan has ... For IEEE...
We see that using approach 1 and 2 both regresses the performance by 23~24%. On 80GB H100 HBM3: 1. Always: 1
Final benchmarks can be viewed here https://gist.github.com/SAtacker/e3410fdc01a42442645b4890c7d609a6 We found out that it wasn't a proper comparison between Kate's many kernel because it never invoked renorm. Above benchmarks are with same...
> Hi @SAtacker !! > I am a new contributor to WasmEdge. I have experience in python packaging . Can i work on this issue ? sure, please take a...