Vyas Ramasubramani
Vyas Ramasubramani
Hmm I don't think that should be a problem. We only require a Pascal architecture, and the 3060ti is Ampere. What errors are you running into? Did you actually build...
> Hi, just wanted to let you know that I'm still working on this. Sorry for taking so long, this is new for me! No worries! Please feel free to...
@ShahRishi were you able to resolve the build issues? Hope that isn't still holding you up.
Yes, I recognized this problem going into 2.0 and punted on it because it wasn't obvious exactly how we should solve it. I don't think the second solution (keeping a...
@bdice once we've resolved some of the other open PRs, let's chat and try to make a call on this?
Right. I'll rename this issue and update the description to clarify that the purpose has changed a bit, then resolve that PR.
I renamed and updated the description of this issue, but forgot to remove the link when merging #619. The broader issues discussed here still need to be addressed.
If you're doing this, be aware of `jitify2.hpp`, which is now available with a slightly modified API and I believe has CUDA 11.2 support.
Actually, the default is False, and the docstring should be updated to include the usual indicator `Default value: False`. However, I can also see that the naming is a bit...
I think that's fine, as long as we remove documentation of the boolean arguments (and replace with the new keywords) and raise DeprecationWarning whenever booleans are provided. My primary concern...