loopy
loopy copied to clipboard
A code generator for array-based code on CPUs and GPUs
Draft because: - [x] Incomplete implementation - [ ] Depends on #350
Suggested by @kaushikcfd in https://github.com/inducer/loopy/pull/631#discussion_r907887383. The original thinking behind it was transform usability. Suppose one splits an iname. The one thing that this currently enables is the use of the...
It either shouldn't, or it shouldn't be a public API.
```python import loopy as lp knl = lp.make_kernel( "{[i]: 0
This is a blind fix (for the superficial issue, but might be symptomatic of some other problem) for, e.g., the following kernel: ```python import loopy as lp knl = lp.make_kernel(...
``` import loopy as lp knl = lp.make_kernel( "{[i, j]: 0
It's perfectly valid for one kernel to compute the launch bounds of the next. https://github.com/inducer/loopy/blob/9c1e37dc859e1132741025ac05dfa00622b08e0d/loopy/check.py#L486-L509
Provided test case fails on `main`. Draft because: - Fix for #600 must land first as it has broken out CI.
x-ref: https://gist.github.com/1st1/292e3f0bbe43bd65ff3256f80aa2637d?permalink_comment_id=2399355 x-ref: https://github.com/inducer/pytato/pull/277#discussion_r837867347 x-ref: https://github.com/inducer/pytato/issues/303 cc @kaushikcfd