Claudia Comito
Claudia Comito
@krajsek can we break this in smaller tasks?
~~Torch binary operations seem to be much more efficient on scalars than on torch tensors.~~ EDIT: `torch.pow()` seems to be a lot more efficient when the exponent is a scalar,...
Some single-process test runs (all arrays `float32`): (**EDIT**: values corrected after fixing mistake in test code, see below, thanks @fschlimb ) | Operation | Heat runtime (s) | NumPy runtime...
> what is your data size for the benchmarks? Hey Daniel, I'm running the exact same example above for different operations, only difference I'm setting `dtype`.
> Hm, the difference I see is not just a few %, it's a factor of 3 slower than numpy and 10x slower than torch. Something else might be going...
@fschlimb @coquelin77 thanks again for all this work. What are the next steps here?
What should be done here @coquelin77 ?
Hi @Dhruv454000 , thanks for taking this on. Make sure you check out the documentation for [`heat.array`](https://heat.readthedocs.io/en/latest/autoapi/heat/core/factories/index.html#heat.core.factories.array) and the [DNDarray](https://heat.readthedocs.io/en/latest/autoapi/heat/core/dndarray/index.html) class. What we're trying to do here is identify those...
> > @ClaudiaComito I have added a check to confirm it is a scalar dndarray , Is anything wrong over there? @Dhruv454000 no nothing wrong, we can simplify things a...
run tests