Andreas Klöckner

`cuFuncSetCacheConfig` is supported: https://documen.tician.de/pycuda/driver.html#pycuda.driver.Function.set_cache_config For the other functions that you find missing, I would welcome a PR.

I haven't tried myself, but I haven't heard any complaints. Can you run other CUDA software in the container? How about Nvidia's SDK samples that use the driver API? You...

Works fine when building from source: https://gitlab.tiker.net/inducer/pyopencl/-/jobs/194730 I don't have the spare bandwidth to engineer the wheel build to do Pypy wheels. I've turned on PR builds on Travis. If...

@kif Two minor clarifications. - Intel is still producing releases for their ICD. They're a bit hidden, but they exist. The most recent release is in the `oclcpuexp` tarball in...

Thanks for working on this! > I am confused by the branch for `pow` - it seems to generate `powf32` or `powf64` based on the type, each of which are...

I'm not in principle against making casts explicit in the code, with the caveat that I'd like the SNR of the generated code to remain reasonably high. If the code...

Thanks for taking the time to describe the instances of casting. While some aren't ideal, I can live with all of them. > casting conditions inside ternaries to char (shouldn't...

> Should the input types be chosen in the same way as, e.g., other OpenCLCallables, via numpy.find_common_type and the like? (Not to append yet another change to this PR...) I'd...

Invited you to the repo, so that you can do that yourself. :)

Thanks for working on this. I see you've taken the approach of pushing the `bool` -> `int8` mapping "up" into type inference. That really worries me, because now our type...