Repository navigation

All pull requests

Labels (70)Milestones (4)New pull request

Open

(169)

Closed

(3,210)

Avoid redundant declaration traversals in DSLX v2 type inference.
#5297copybara-service[bot] opened Nov 7, 2026Bot
[DSLX] Support array_rev and array_size in map IR conversion
10000
#5296cdleary opened Nov 7, 2026Collaborator
[dslx] Preserve signedness of out-of-bounds width slices
#5295cdleary opened Nov 6, 2026Collaborator
Update ReceiveDefaultValueSimplificationPass documentation comment.
#5289copybara-service[bot] opened Nov 6, 2026Bot
fix(noc): correct spelling typos in simulation comments
#5285Jah-yee opened Nov 6, 2026
Precompute fitted estimator model protos to remove scipy from C++ builds.
#5283copybara-service[bot] opened Nov 6, 2026Bot
Name the adjusted loop index in IR for DSLX ranges with a nonzero start
#5281dank-openai opened Nov 5, 2026Collaborator
Resolve a checked DSLX import to its module in the C API
#5280dank-openai opened Nov 5, 2026Collaborator
[dslx] Fix IR conversion for wide decode results
#5278cdleary opened Nov 5, 2026Collaborator
Clean up and reorganize GEMINI.md structure and documentation links.
#5276copybara-service[bot] opened Nov 5, 2026Bot
Resource sharing pass: defer visibility edge computation
#5273copybara-service[bot] opened Nov 4, 2026Bot
Track original requested node names and promote them when freed.
#5272copybara-service[bot] opened Nov 4, 2026Bot
ResourceSharingPass previously restricted multiplication folding to identical
#5269copybara-service[bot] opened Nov 4, 2026Bot
ArraySimplificationPass previously split multi-user array-typed selects only
#5268copybara-service[bot] opened Nov 4, 2026Bot
ShiftEquivalenceMapping::TryCreate previously rejected any folding group
#5267copybara-service[bot] opened Nov 4, 2026Bot
[dslx] Return the explicit for-loop body result in IR
#5262cdleary opened Nov 4, 2026Collaborator
[dslx] #1677: Improve error produced by sizeof applied to a type
#5261axman6 opened Nov 4, 2026
6
[dslx] Fix IR conversion for tuple-valued update indices
#5260cdleary opened Nov 4, 2026Collaborator
Fix nightly builds in CI
#5259rw1nkler opened Nov 3, 2026Contributor
2
Run dslx_fmt --simplify_type_annotations on all xls/examples files.
#5258copybara-service[bot] opened Nov 3, 2026Bot
Narrow logical right shifts used only through low bit slices
#5246Novpolina opened Nov 3, 2026
Strength-reduce umul by variable powers of two
#5245Novpolina opened Nov 3, 2026
[dslx] Issue #683: Add block comments (without fmt support)
#5244axman6 opened Nov 3, 2026
1
[dslx] Fix expansion of signed one-bit ranges
#5243cdleary opened Nov 3, 2026Collaborator
[dslx] preserve range start bits during IR conversion
#5242cdleary opened Nov 3, 2026Collaborator

Navigation Menu

Repository navigation

All pull requests

Avoid redundant declaration traversals in DSLX v2 type inference.

[DSLX] Support array_rev and array_size in map IR conversion

[dslx] Preserve signedness of out-of-bounds width slices

Update ReceiveDefaultValueSimplificationPass documentation comment.

fix(noc): correct spelling typos in simulation comments

Precompute fitted estimator model protos to remove scipy from C++ builds.

Name the adjusted loop index in IR for DSLX ranges with a nonzero start

Resolve a checked DSLX import to its module in the C API

[dslx] Fix IR conversion for wide decode results

Clean up and reorganize GEMINI.md structure and documentation links.

Resource sharing pass: defer visibility edge computation

Track original requested node names and promote them when freed.

`ResourceSharingPass` previously restricted multiplication folding to identical

`ArraySimplificationPass` previously split multi-user array-typed selects only

`ShiftEquivalenceMapping::TryCreate` previously rejected any folding group

[dslx] Return the explicit for-loop body result in IR

[dslx] #1677: Improve error produced by sizeof applied to a type

[dslx] Fix IR conversion for tuple-valued update indices

Fix nightly builds in CI

Run `dslx_fmt --simplify_type_annotations` on all xls/examples files.

Narrow logical right shifts used only through low bit slices

Strength-reduce umul by variable powers of two

[dslx] Issue #683: Add block comments (without fmt support)

[dslx] Fix expansion of signed one-bit ranges

[dslx] preserve range start bits during IR conversion

Pull requests