feilongjiang requested elliottt for a review on PR #8393.
feilongjiang opened PR #8393 from feilongjiang:fix_pcc
to bytecodealliance:main
:
Hi team, please consider.
As described in #8380, when running PCC validation on workloads like PolyBenchC, it will report anUnsupportedFact
error.
Based on the analysis of trace logs, I thinkattach_constant_fact
introduced by #8173 should also consider the type of constant when adding facts for iconsts. After #8173, we may generate the following facts for i32 constants:block9: v4339 ! range(64, 0xea28, 0xea28) = iconst.i32 0xea28 v2858 ! range(64, 0x8, 0x8) = iconst.i32 8 v4406 ! range(64, 0x115a0, 0x115a0) = iconst.i32 0x0001_15a0 v4407 ! range(64, 0x115a8, 0x115a8) = iadd v4406, v2858 ; v4406 = 0x0001_15a0, v2858 = 8
These i32 constants have their
bit_width
set to 64, causing the error.#8173 also exposes another issue on AArch64 when generating facts for
MovK
, we should calculate the running value forMovK
with the field mask.<!--
Please make sure you include the following information:
If this work has been discussed elsewhere, please include a link to that
conversation. If it was discussed in an issue, just mention "issue #...".Explain why this change is needed. If the details are in an issue already,
this can be brief.Our development process is documented in the Wasmtime book:
https://docs.wasmtime.dev/contributing-development-process.htmlPlease ensure all communication follows the code of conduct:
https://github.com/bytecodealliance/wasmtime/blob/main/CODE_OF_CONDUCT.md
-->
feilongjiang requested wasmtime-compiler-reviewers for a review on PR #8393.
github-actions[bot] commented on PR #8393:
Subscribe to Label Action
cc @cfallin, @fitzgen
<details>
This issue or pull request has been labeled: "cranelift", "cranelift:area:aarch64", "isle"Thus the following users have been cc'd because of the following labels:
- cfallin: isle
- fitzgen: isle
To subscribe or unsubscribe from this label, edit the <code>.github/subscribe-to-label.json</code> configuration file.
Learn more.
</details>
jameysharp submitted PR review:
Thank you! This looks good to me, though I'd like @cfallin to take a look before merging it, and he may not be able to get to it until next week.
jameysharp submitted PR review:
Thank you! This looks good to me, though I'd like @cfallin to take a look before merging it, and he may not be able to get to it until next week.
jameysharp created PR review comment:
I'd prefer parentheses here even if they're not strictly required:
let constant = (input_constant & !mask) | constant;
feilongjiang updated PR #8393.
feilongjiang commented on PR #8393:
@jameysharp Thanks for the review!
and he may not be able to get to it until next week.
That’s okay, we can wait until @cfallin comes back!
feilongjiang updated PR #8393.
cfallin submitted PR review:
LGTM, thanks!
cfallin merged PR #8393.
Last updated: Dec 23 2024 at 13:07 UTC