Andreas Klöckner
Andreas Klöckner
I'm having a hard time understanding what you're asking for. If you look at `test_array.py`, you can see sample code using, e.g. `cltypes.float4`. Does that help?
Could you draft a PR that improves the documentation in the way you're suggesting?
Could you explain what the presumed benefit is? An immediate downside is that devices without high-performance double support (Nvidia...) will suffer a decrease in performance. As a result, maybe this...
I'd be happy to take a PR to realize this.
Just add handling for the real-valued case here: https://github.com/inducer/loopy/blob/718c503b68e0843d58b8e049a80e0072eaea88e6/loopy/target/pyopencl.py#L239-L266
Well, in an ideal world, Apple would just fix their compiler. That said, they're not [the only CL implementation](https://github.com/pocl/pocl/issues/1) struggling with that. If you can make it so the header...
Interesting problem. If we built a mechanism for dealing with this right into PyOpenCL, how would you handle the case of being passed an array with a dtype that doesn't...
Would that be detected (at call time)? At what cost? How would type-dependent code generation be handled? Some caches may be keyed off of the type identifier and not the...
I'm honestly still not sure what the advantage of this scheme is over one where different types (on different devices) have different names--that seems reasonably manageable from an application perspective,...
That's clearly a bug. Believe it or not, PyCUDA predates `__truediv__`, I think. :astonished: