kyrollosyanny
kyrollosyanny
For some simulations and optimizations, cpu is more than enough. If it is possible to add that support in future versions, that would be great. Thanks a lot.
@awni regarding the "metal doesn't support double". Is that a fundamental hardware limitation that will never be overcome or is it the current version of metal? And do you know...
The main advantage I see for cpu float64 mlx vs numpy is the differentiability. For many scientific computing problems (optical ray tracing, nanophotonics, fluid simulations and others), float64 is needed...
Hi @awni, any update on when matrix inversion would be available on GPU? It would be extremely helpful for a lot of applications. For example, using Gauss-Newton second order optimizers...
Interestingly, it doesn't give me any errors if I do this `mx.linalg.inv(JtJ,stream=mx.gpu)` but maybe it is still not using the gpu?
Yes, just updated and gave me the error you mentioned. Only works with cpu stream for now. Thanks for the help.
Got it. Naive question, When you say locally, does it mean there is a way to run MLX on the cloud in higher precision? Thanks a lot