CUDA: Add accessors for function attributes beyond just register use per thread

[gmarkall]

PR #6447 adds a public API to get the maximum number of registers per thread (numba.cuda.Dispatcher.get_regs_per_thread()). There are other attributes that might be nice to provide - shared memory per block, local memory per thread, const memory usage, maximum block size.

These are all available in the FuncAttr named tuple: https://github.com/numba/numba/blob/master/numba/cuda/cudadrv/driver.py#L2023, set by _read_func_attr_all: https://github.com/numba/numba/blob/master/numba/cuda/cudadrv/driver.py#L2067

Cross ref: adding these would provide some of the info in the ptxas output requested in #3482.

[edwinlim0919]

Hi Graham! I'm a student @ The University of California, Berkeley and this is my first time contributing to open source. I am interested in addressing this issue.

[Stark-developer01]

PR #6447 adds a public API to get the maximum number of registers per thread (numba.cuda.Dispatcher.get_regs_per_thread()). There are other attributes that might be nice to provide - shared memory per block, local memory per thread, const memory usage, maximum block size.

These are all available in the FuncAttr named tuple: https://github.com/numba/numba/blob/master/numba/cuda/cudadrv/driver.py#L2023, set by _read_func_attr_all: https://github.com/numba/numba/blob/master/numba/cuda/cudadrv/driver.py#L2067

Cross ref: adding these would provide some of the info in the ptxas output requested in #3482.

I want to address this issue @gmarkall