Stream: git-wasmtime

Topic: wasmtime / PR #4401 Convert `scalar_to_vector` to ISLE (A...

Wasmtime GitHub notifications bot (Jul 07 2022 at 14:00):

dheaton-arm opened PR #4401 from isle-scalartovector to main:

Converted the exisiting implementation of scalar_to_vector for AArch64 to
ISLE.

Copyright (c) 2022 Arm Limited

<!--

Please ensure that the following steps are all taken care of before submitting
the PR.

• [ ] This has been discussed in issue #..., or if not, please tell us why
  here.

• [ ] A short description of what this does, why it is needed; if the
  description becomes long, the matter should probably be discussed in an issue
  first.

• [ ] This PR contains test cases, if meaningful.

• [ ] A reviewer from the core maintainer team has been assigned for this PR.
  If you don't know who could review this, please indicate so. The list of
  suggested reviewers on the right can help you.

Please ensure all communication adheres to the code of conduct.
-->

Wasmtime GitHub notifications bot (Jul 07 2022 at 15:13):

alexcrichton submitted PR review.

Wasmtime GitHub notifications bot (Jul 07 2022 at 15:13):

alexcrichton created PR review comment:

Just a drive-by-comment, but I know that historically we've had some issues where instructions were expanded over time to accept more types and we sometimes forget to update previous instruction lowerings to account for the new types flowing in. Could the multi_lane _ _ matcher here be replaced with $I32X4 perhaps to avoid that? Reading over the previous lowering it looks like only I32X4 and I64X2 are supported so it may make sense to only list those two so if other types are ever used in the future an error pops out during lowering rather than the wrong lowering be used and waiting until runtime to have the error.

Wasmtime GitHub notifications bot (Jul 07 2022 at 15:21):

dheaton-arm submitted PR review.

Wasmtime GitHub notifications bot (Jul 07 2022 at 15:21):

dheaton-arm created PR review comment:

The AArch64 backend itself supports i8 -> i8x16 and i16 -> i16x8 using the multi_lane _ _ rule as it is (see runtests/simd-scalartovector-aarch64.clif).

Wasmtime GitHub notifications bot (Jul 07 2022 at 15:22):

alexcrichton submitted PR review.

Wasmtime GitHub notifications bot (Jul 07 2022 at 15:22):

alexcrichton created PR review comment:

Oh sorry I missed that, disregard me then!

Wasmtime GitHub notifications bot (Jul 07 2022 at 17:35):

cfallin submitted PR review.

Wasmtime GitHub notifications bot (Jul 07 2022 at 17:35):

cfallin created PR review comment:

Here as well we'll want a rule priority (as in #4399) so this rule is ensured to be matched after the above case.

Wasmtime GitHub notifications bot (Jul 07 2022 at 17:35):

cfallin submitted PR review.

Wasmtime GitHub notifications bot (Jul 13 2022 at 12:02):

dheaton-arm updated PR #4401 from isle-scalartovector to main.

Wasmtime GitHub notifications bot (Jul 18 2022 at 18:11):

cfallin submitted PR review.

Wasmtime GitHub notifications bot (Jul 18 2022 at 18:11):

cfallin merged PR #4401.

Last updated: Dec 23 2024 at 13:07 UTC