Skip to content

[GR-52711] Implement GraalWasm SIMD proposal using the Vector API. #11630

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Open
wants to merge 101 commits into
base: master
Choose a base branch
from
Open

[GR-52711] Implement GraalWasm SIMD proposal using the Vector API. #11630

wants to merge 101 commits into from

Conversation

graalvmbot
Copy link
Collaborator

@graalvmbot graalvmbot commented Jul 15, 2025 •
edited by fniephaus
Loading

Fixes #9767

@oracle-contributor-agreement oracle-contributor-agreement bot added the OCA Verified All contributors have signed the Oracle Contributor Agreement. label Jul 15, 2025
@graalvmbot graalvmbot force-pushed the jmarsik/wasm-simd-vector-api branch 5 times, most recently from df0b8f6 to 85bfd80 Compare July 21, 2025 09:30
jirkamarsik added 24 commits July 22, 2025 13:16
@jirkamarsik
Use Vector API for v128_any_true, f32x4_binop, f64x2_binop
36932d3
@jirkamarsik
Use Vector API for i32x4_all_true, i32x4_binop and i32x4_relop
1e77bdd
@jirkamarsik
Use Vector values on the GraalWasm stack directly
8fb06a2
@jirkamarsik
Cast to Byte128Vector on SIMD operation boundaries
31b9f1a
@jirkamarsik
Avoid allocating byte[] in v128.const
d324d02
@jirkamarsik
Use a more efficient implementation for v128.any_true
133c3e1
@jirkamarsik
Use a more efficient implementation for i32x4.all_true
f1d51ba
@jirkamarsik
Implement SIMD load instructions using Vector API
f377341
@jirkamarsik
Implement i8x16.shuffle using the Vector API
d9e76ca
@jirkamarsik
Use Vector API SIMD ops for extending loads
d765016
@jirkamarsik
Implement v128.store using Vector API
ef4ae71
@jirkamarsik
Implement extract_lane and replace_lane using Vector API
cfd8598
@jirkamarsik
Implement shift ops using the Vector API
534c534
@jirkamarsik
Remove redundant ExplodeLoop annotations
23fe4ea
@jirkamarsik
Implement wasm SIMD ternary operators using Vector API
e9ded4c
@jirkamarsik
Implement wasm SIMD bitwise binary ops using Vector API
943bf82
@jirkamarsik
Implement splat operations using Vector API
85d59fc
@jirkamarsik
Implement unary operators using Vector API
8080df2
@jirkamarsik
Implement load_lane and store_lane ops using Vector API
4c13777
@jirkamarsik
Implement binary ops using Vector API
8ff6d3a
@jirkamarsik
Implement vectorToInt ops using Vector API, reorder Vector128Ops
97ea714
@jirkamarsik
SIMD implementation bug fixes
be625a7
@jirkamarsik
Avoid using reinterpretShape
19452ab
@jirkamarsik
Use module path for GraalWasm launcher and tests
bb30262
jirkamarsik added 15 commits July 22, 2025 13:34
@jirkamarsik
Use Locale.ROOT during image build-time operations
fd2d9c3
@jirkamarsik
Split off VectorAPI TruffleBoundaries to separate file
15d5246
@jirkamarsik
Populate cache without calling ImplCache.find
22bb2bb 
Avoids the isNonCapturingLambda assertion.
@jirkamarsik
Do not call makeConv to create Identity between different types
89c67e9
@jirkamarsik
Avoid calling vector constructors during Vector128Ops clinit
cf8465c 
This avoids running into the isNonCapturingLambda assertion.
@jirkamarsik
Fix failing MaxShuffle VLENGTH assertion
7777588
@jirkamarsik
Handle optional directories in PolyBenchmarkSuite
b5da00f
@jirkamarsik
Move all jdk.incubator.vector usage to Vector128Ops
ec1796a
@jirkamarsik
Move Vector API SIMD implementation to separate project
9b962de
@jirkamarsik
Provide both a Vector API and a fallback impl of GraalWasm SIMD
84b560b
@jirkamarsik
Fix API mismatch in Vector128Ops
d4e4883
@jirkamarsik
Do not set Vector API specific options for libwasmvm on pre-JDK25 Gra…
0b5594a 
…alVMs
@jirkamarsik
Add VectorMathLibrary#lookup to jdk_allowed_methods
95b886c
@jirkamarsik
Add changelog entry for Wasm SIMD Vector API implementation
0f75443
@jirkamarsik
Update method signatures in VectorAPITruffleBoundaries for JDK 26+6
2729ca2
@graalvmbot graalvmbot force-pushed the jmarsik/wasm-simd-vector-api branch from bc71e96 to 55fb802 Compare July 22, 2025 11:42
@jirkamarsik
Make the Vector API an optional dependency of GraalWasm
3a4e016 
GraalWasm now requires jdk.incubator.vector using "requires static". If
the jdk.incubator.vector was not provided, we fall back to our previous
implementation. This makes incubator warnings opt in.
@graalvmbot graalvmbot force-pushed the jmarsik/wasm-simd-vector-api branch from 55fb802 to 3a4e016 Compare July 22, 2025 12:02
@graalvmbot graalvmbot force-pushed the jmarsik/wasm-simd-vector-api branch from 41c71b6 to 97b4cb3 Compare July 22, 2025 23:59
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers
No reviews
Assignees

@jirkamarsik jirkamarsik

Labels
OCA Verified All contributors have signed the Oracle Contributor Agreement.
Projects
None yet
Milestone
25.0.0 Release (September 16, 2025)
Development

Successfully merging this pull request may close these issues.

[GR-52711] Add compiler intrinsics for GraalWasm SIMD instructions
5 participants
@graalvmbot @jirkamarsik @fniephaus @gergo- @woess