Andreas Klöckner
Andreas Klöckner
Agree with @kaushikcfd here: Hacking the `CTarget` to do the same thing as the CL one would be a way to achieve this. CL also gives control of the allocator,...
My main objection to the static arrays would be that they fall apart if their size is parameter-dependent, which IMO is a fairly common use case. (The AST handling from...
The main place to look is this bit of code in the PyOpenCL Python Target: https://github.com/inducer/loopy/blob/e3431a17bcbcdf9a37b4812bf997729536b232ee/loopy/target/pyopencl.py#L658-L828 (That generates the host code that actually allocates and passes the global temporaries.) An...
Thanks for the report and the suggested fix. What'll happen if "unsafe" characters are found is that the affected argument gets quoted. Are you saying that the quoting is breaking...
Yeah, I can confirm your observation. So it seems the idea of quoting is just thoroughly broken. Which begs the question what the correct behavior would be if PyOpenCL encounters...
The core issue is that PyOpenCL needs to be able to combine sets of build options (its own default ones and the user-supplied ones, for instance). Since the current quoting...
How come you didn't trigger this on the array having dtype of `np.bool`?
Does. Thanks for the hint. As a design matter, is `ls` really a good default command, since (in the way described above) it makes the whole thing a little less...
IMO that'd be more user-friendly, but that's obviously also in the eye of the beholder.
It appears that this is a regression. mypy 0.942 does not object to the code in the OP.