loopy
loopy copied to clipboard
inducer
A code generator for array-based code on CPUs and GPUs
TLDR: defining array sizes with unsigned integers and then tiling the resulting kernel will break the per-thread check to see if it is inside the loop domain. This is because...
New draft PR based on [PR 683](https://github.com/inducer/loopy/pull/683). Major change is that the HappensAfter data structure used to represent statement-level dependency relations has been implemented into the InstructionBase data structure.
Since https://github.com/inducer/loopy/pull/710 (perhaps), `memoize_on_disk` produces quite long filenames, for example: ``` .../.cache/pytools/pdict-v4-loopy-memoize-cache-buffer_array-LoopyKeyBuilder.LoopyKeyBuilder-v0-2022.1-islpy2023.1.2-cgen2020.1-8158afdb18f100fb21b6cd86c6178e3c5cd19058-v1-py3.10.12.final.0 ``` Unfortunately, some filesystems have quite short name restrictions. e.g. we had a bug report where this failed...
- Adds `HappensAfter` data structure - Adds routine to compute lexicographic order of statements (found in `loopy/kernel/dependency.py`) - Updates `InstructionBase` to accommodate precise dependency semantics (`depends_on` -> `happens_after`) - Other...
e.g. https://gitlab.tiker.net/inducer/loopy/-/jobs/564054 IMO there are two issues here: - First, how come the CI job is passing when all tests seem to fail? That shouldn't be happening. - Second, do...
For context, in a sumpy P2P kernel I have a temporary of size ``` local_isrc[5, 45] ``` which results in 5 memory loads/stores, but it could be split into ```...
As introduced as a workaround in #768. Needs: - [ ] A permanent solution for https://github.com/inducer/islpy/issues/102 is found.
I don't have a use-case for this, but thought it would be interesting to keep in mind.
