[wasm] Enable Vector128 fast paths on Wasm via PackedSimd: hex/Guid, UTF-8, SearchValues, Teddy, Adler32/XXH3 by lewing · Pull Request #129838 · dotnet/runtime

lewing · 2026-06-25T04:54:53Z

Enable several Vector128 fast paths in CoreLib and System.IO.Hashing on browser-wasm by adding a PackedSimd.IsSupported branch alongside the existing Sse2/Ssse3/AdvSimd.Arm64 gates. Before these changes, the SIMD code paths were unreachable on Wasm even though the wasm runtime supports PackedSimd (the test pipeline default sets WasmEnableSIMD=true).

Scope and commits

#	Commit	Area
1	`c69998277eb` `0e99f4a194c`	Add Wasm `PackedSimd` path to `Vector128.UnpackLow/UnpackHigh` (composed from two `PackedSimd.Swizzle` calls + OR; the two-vector `PackedSimd.Shuffle` requires constant lane indices and is impractical from generic code).
2	`3dc956b6465`	Enable Vector128 hex/Guid format fast path on Wasm — widen the gates on `HexConverter.AsciiToHexVector128`, `HexConverter.EncodeToUtf8/Utf16`, `HexConverter.EncodeTo_Vector128`, and `Guid.FormatGuidVector128Utf8`. Bodies were already portable (`Vector128.ShuffleNative`, `Vector128.UnpackLow/High`, constant-index `Vector128.Shuffle`).
3	`061b22bffcf` `2543702633`	Vectorize `Utf8Utility.Validation` ASCII fast path on Wasm — add a `PackedSimd` branch alongside `AdvSimd.Arm64`/`Sse2` in the inner ASCII-scan loop of `GetPointerToFirstInvalidChar`, using portable `Vector128.LoadUnsafe` + `ExtractMostSignificantBits`.
4	`da08d905c6e`	Vectorize `Utf8Utility.Transcoding` ASCII fast path on Wasm — the 8-char narrow-store loop and the two 4-char tail stores, using `PackedSimd.ConvertNarrowingSaturateUnsigned`.
5	`2caa5e25997` `0e99f4a194c`	Enable Teddy multi-string search on Wasm — `TeddyHelper.LoadAndPack16AsciiChars`, `GetNibbles`, two-table `Shuffle`, `RightShift1/RightShift2` (composed from two `PackedSimd.Swizzle` calls + OR), plus the `StringSearchValues.CreateFromNormalizedValues` entry gate and `AsciiStringSearchValuesTeddyBase.IndexOfAnyN2/N3` `[CompExactlyDependsOn]`.
6	`819c9fe3960`	Enable `ProbabilisticMap` vectorized `SearchValues<char>` on Wasm. Subtle because the bitmap layout itself branches on the gate (`SetCharBit`/`IsCharBitSet`, marked `[BypassReadyToRun]`). Widens the layout choice, `ContainsMask16Chars` (new PackedSimd narrowing branch), the `IndexOfAny`/`LastIndexOfAny` entry dispatcher, and the worker `[CompExactlyDependsOn]` attributes consistently.
7	`80522ca629b`	`Adler32.UpdateVector128` and `XxHashShared.MultiplyWideningLower`: replace the portable `Vector128.Widen + multiply + add` sequences with `PackedSimd.AddPairwiseWidening` + `MultiplyWideningLower/Upper` and `PackedSimd.MultiplyWideningLower` (`i64x2.extmul_low_i32x4_u`). Already vectorized on Wasm via the generic else branch; this turns 3–5 ops/iter into 1.

Test results

Validated end-to-end with WasmEnableSIMD=true (the browser-wasm test pipeline default).

Suite	Host arm64	browser-wasm (V8 v15)	Coverage
`System.Runtime.Tests`	69,714 / 69,714 ✅	67,835 / 67,835 ✅	Guid format/parse
`System.Runtime.Extensions.Tests`	8,350 / 8,350 ✅	8,224 / 8,224 ✅	`Convert.ToHexString`/`FromHexString`
`System.Memory.Tests`	52,906 / 52,906 ✅	52,249 / 52,249 ✅	UTF-8 validation/transcoding, Ascii, `SearchValues<string>` (Teddy), `SearchValues<char>` (ProbabilisticMap), `SpanHelpers`
`System.IO.Hashing.Tests`	4,196 / 4,196 ✅	4,196 / 4,196 ✅	Adler32, XxHash (lane-order is checked end-to-end via the hash output bytes — a swap would corrupt every hash)

The original push exposed two AOT-only failure modes in browser-wasm Helix legs:

PackedSimd.Shuffle (two-vector i8x16.shuffle) requires compile-time-constant lane indices; Mono AOT can't fold a Vector128.Create(...) operand and throws PlatformNotSupportedException. Fixed in c69998277eb → 0e99f4a194c by composing with single-vector PackedSimd.Swizzle + OR.
The same pattern was originally applied in TeddyHelper.RightShift1/RightShift2 via Vector128.ShuffleNative (a dispatcher with an Ssse3 → AdvSimd.Arm64 → PackedSimd if/else chain). The Mono SIMD intrinsic recognizer doesn't always lower that chain cleanly for less-traveled paths — the safer pattern is to call PackedSimd.Swizzle directly under the PackedSimd.IsSupported branch. Applied consistently in 0e99f4a194c.

Files changed

src/libraries/System.Private.CoreLib/src/System/Runtime/Intrinsics/Vector128.cs
src/libraries/Common/src/System/HexConverter.cs
src/libraries/System.Private.CoreLib/src/System/Guid.cs
src/libraries/System.Private.CoreLib/src/System/Text/Unicode/Utf8Utility.Validation.cs
src/libraries/System.Private.CoreLib/src/System/Text/Unicode/Utf8Utility.Transcoding.cs
src/libraries/System.Private.CoreLib/src/System/SearchValues/Strings/Helpers/TeddyHelper.cs
src/libraries/System.Private.CoreLib/src/System/SearchValues/Strings/StringSearchValues.cs
src/libraries/System.Private.CoreLib/src/System/SearchValues/Strings/AsciiStringSearchValuesTeddyBase.cs
src/libraries/System.Private.CoreLib/src/System/SearchValues/ProbabilisticMap.cs
src/libraries/System.IO.Hashing/src/System/IO/Hashing/Adler32.cs
src/libraries/System.IO.Hashing/src/System/IO/Hashing/XxHashShared.cs

Not in this PR

Base64DecoderHelper / Base64EncoderHelper — need pmaddubsw and pmulhuw analogs composed from MultiplyWideningLower/Upper + AddPairwiseWidening or Dot, with careful 8-short lane preservation. Worth a dedicated follow-up with benchmarks.

Note

This PR description and the commits in this branch were drafted with AI/Copilot assistance.

These internal helpers are used by HexConverter.AsciiToHexVector128 and other byte-interleaving code paths. They previously dispatched only to Sse2.UnpackLow/UnpackHigh or AdvSimd.Arm64.ZipLow/ZipHigh and threw NotSupportedException on platforms without either ISA. With the recent change that enables HexConverter and Guid format on Wasm via PackedSimd, the helpers became reachable on browser-wasm and started throwing at runtime in libraries tests. Lower to PackedSimd.Shuffle with a constant 16-byte index vector (i8x16.shuffle) when PackedSimd is supported. Validated via System.Runtime.Extensions.Tests on browser-wasm (8224 passing, 0 failed) after the previous run failed 132 tests. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

HexConverter.AsciiToHexVector128, HexConverter.EncodeTo_Vector128 and Guid.FormatGuidVector128Utf8 already use only portable Vector128 ops (Vector128.ShuffleNative, Vector128.UnpackLow/High, Vector128.Shuffle with constant indices) plus an optional AdvSimd.Arm64-specific branch. The gates at Convert.ToHexString, EncodeToUtf8/Utf16, and Guid.ToString required Ssse3 or AdvSimd.Arm64, so Wasm fell back to scalar even with PackedSimd. Add PackedSimd.IsSupported to the gates and the [CompExactlyDependsOn] attributes on the helpers. The bodies are unchanged; on Wasm the existing else branch (portable Vector128.Shuffle) is selected. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

GetPointerToFirstInvalidChar's inner ASCII-scan loop dispatched on AdvSimd.Arm64 (with bitmask128) or Sse2 (with MoveMask), falling back to a scalar 4-DWORD-at-a-time path otherwise. On Wasm with PackedSimd, neither SIMD branch was taken, so UTF-8 validation took the scalar path. Add a PackedSimd.IsSupported branch that uses portable Vector128.LoadUnsafe + ExtractMostSignificantBits to compute the same per-byte non-ASCII bitmask used by the Sse2 path. Update the post-loop Debug.Assert to include PackedSimd. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

dotnet-policy-service · 2026-06-25T04:56:01Z

Tagging subscribers to this area: @dotnet/area-system-numerics
See info in area-owners.md if you want to be subscribed.

Copilot

Pull request overview

This PR enables existing Vector128-based fast paths to run on browser-wasm by widening the feature gates to include System.Runtime.Intrinsics.Wasm.PackedSimd and adding a PackedSimd.Shuffle implementation for Vector128.UnpackLow/UnpackHigh so dependent encode/format paths don’t fall into NotSupportedException on Wasm SIMD-enabled runs.

Changes:

Add PackedSimd.IsSupported implementations for Vector128.UnpackLow / UnpackHigh using PackedSimd.Shuffle (two-vector shuffle).
Widen existing hex + Guid formatting SIMD gates / [CompExactlyDependsOn] to include PackedSimd.
Add a PackedSimd vectorized ASCII-scan path in Utf8Utility.Validation using Vector128.LoadUnsafe + ExtractMostSignificantBits.

Reviewed changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 1 comment.

File	Description
src/libraries/System.Private.CoreLib/src/System/Runtime/Intrinsics/Vector128.cs	Adds `PackedSimd` support to `UnpackLow/UnpackHigh` to avoid Wasm falling into the unsupported path.
src/libraries/Common/src/System/HexConverter.cs	Expands encode-side Vector128 gate/attributes to include `PackedSimd` for Wasm SIMD.
src/libraries/System.Private.CoreLib/src/System/Guid.cs	Expands Guid vectorized formatting gate/attributes to include `PackedSimd` on little-endian.
src/libraries/System.Private.CoreLib/src/System/Text/Unicode/Utf8Utility.Validation.cs	Adds a `PackedSimd` ASCII-validation fast path using MSB extraction.

TranscodeToUtf8's 8-char ASCII fast loop used a Vector128<short> read, a mask-and-compare to detect non-ASCII, and a narrow-and-store of 8 bytes using Sse2.PackUnsignedSaturate / AdvSimd.ExtractNarrowingSatura teUnsignedLower. Two follow-on 4-char sites narrowed 4 bytes the same way. All four sites required Sse41.X64 or AdvSimd.Arm64 + LE, so Wasm took the 4-DWORD-at-a-time scalar fallback. Add PackedSimd branches at every dispatch site: - Outer entry gate (declaration + entry condition) - 8-char narrow-store: use the existing portable AND-compare for the non-ASCII test (same code Sse41 already uses) and PackedSimd.Convert NarrowingSaturateUnsigned + scalar extract for the store - 4-char narrow-stores: PackedSimd.ConvertNarrowingSaturateUnsigned + AsUInt32().ToScalar() unaligned write The Sse2.X64.ConvertToUInt64 sub-branch already had an else path that calls AsUInt64().ToScalar(), which works on Wasm without changes. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

SearchValues<string> with 2+ values previously selected the Aho-Corasick implementation on Wasm because the Teddy entry gate in StringSearchValues.cs required Ssse3 or AdvSimd.Arm64. Teddy's core Vector128 primitives in TeddyHelper.cs (LoadAndPack16AsciiChars, the nibble GetNibbles helper, the two-table Shuffle, and RightShift1/2) similarly excluded PackedSimd. Add PackedSimd branches throughout: - LoadAndPack16AsciiChars: PackedSimd.ConvertNarrowingSaturateUnsigned - GetNibbles: PackedSimd needs the explicit '& 0xF' on the low half because Swizzle returns 0 for indices >= 16 (unlike Ssse3's implicit AND of the low 4 bits) - Shuffle: already uses portable Vector128.ShuffleNative which maps to PackedSimd.Swizzle; just widen the [CompExactlyDependsOn] - RightShift1/RightShift2: compose two Vector128.ShuffleNative calls with constant index vectors and OR the halves. PackedSimd.Shuffle (two-vector i8x16.shuffle) is impractical due to constant lane index requirements; Swizzle clamps out-of-range to 0 which makes the OR safe. Widen the entry gate in StringSearchValues.cs.CreateFromNormalizedV alues and the null-char filter in TryGetTeddyAcceleratedValues (PackedSimd shares Ssse3's PackUnsignedSaturate behavior where signed negative inputs become 0, so null-containing needles produce more false positives on both). Widen [CompExactlyDependsOn] on the IndexOfAnyN2/N3 + Vector128 helpers in AsciiStringSearchValuesTeddyBase.cs. Validated: System.Memory.Tests on browser-wasm 52249/52249 passing (covers SearchValues<string> Teddy paths via StringSearchValues tests), host 52905/52906 unchanged. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

SearchValues<char> with values that span more than the ASCII range selects a ProbabilisticMap-based search. The vectorized IndexOfAny / LastIndexOfAny path (using ContainsMask16Chars + IsCharBitNotSet) was previously gated on Sse41 || AdvSimd.Arm64 only, so on Wasm the search fell back to the scalar SimpleLoop even when PackedSimd was available. This change is subtler than the other enablement PRs because the *layout* of the ProbabilisticMap bitmap also branches on the same gate (SetCharBit/IsCharBitSet at the top of the file). The [BypassReadyToRun] comment there warns that the construction and lookup branches must agree at all times during program execution. Widen all three gates (SetCharBit/IsCharBitSet, ContainsMask16Chars, the IndexOfAny/LastIndexOfAny entry dispatcher, and the [CompExactly DependsOn] on the Vector128 worker methods) to include PackedSimd consistently. ContainsMask16Chars gets a PackedSimd branch that mirrors the Sse2 algorithm using PackedSimd.ConvertNarrowingSaturateUnsigned for the two-vector narrowing step. IsCharBitNotSet already had a PackedSimd dependency for the table lookup via Vector128.ShuffleNative. ProbabilisticWithAsciiCharSearchValues already had PackedSimd dispatch. Validated: System.Memory.Tests on browser-wasm 52249/52249 passing, host arm64 52905/52906 unchanged. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Both algorithms were already vectorized on Wasm via the portable Vector128 else branch (Vector128.Widen + multiply + add), but the result was 3-5 portable ops per iteration where PackedSimd has a direct one-instruction equivalent. Adler32.UpdateVector128: add a PackedSimd branch alongside Sse2 and AdvSimd that uses PackedSimd.AddPairwiseWidening (i16x8.extadd_pair wise_i8x16_u and i32x4.extadd_pairwise_i16x8_u) for the s1 sum and PackedSimd.MultiplyWideningLower/Upper + AddPairwiseWidening for the weighted s2 sum. XxHashShared.MultiplyWideningLower: add a PackedSimd branch that computes { source[0]*source[1], source[2]*source[3] } via two shuffles + i64x2.extmul_low_i32x4_u, replacing the portable mask + 64-bit multiply pair. Validated: System.IO.Hashing.Tests 4196/4196 passing on both host arm64 and browser-wasm (the XxHash lane order is checked end-to-end via the algorithm output bytes — a swap would corrupt every hash). Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

lewing · 2026-06-25T05:36:16Z

Added 4 more commits extending the Wasm Vector128 fast-path enablement:

da08d905c6e — Utf8Utility.Transcoding: vectorize the 8-char ASCII fast path in TranscodeToUtf8 and the two follow-on 4-char narrow-stores using PackedSimd.ConvertNarrowingSaturateUnsigned.
2caa5e25997 — Enable Teddy multi-string search on Wasm. Adds PackedSimd dispatch to TeddyHelper.LoadAndPack16AsciiChars (ConvertNarrowingSaturateUnsigned), GetNibbles, the two-table Shuffle, and RightShift1/RightShift2 (two Vector128.ShuffleNative calls + OR — PackedSimd.Shuffle is impractical due to constant lane index requirements). Widens the entry gate in StringSearchValues.CreateFromNormalizedValues, the null-char filter in TryGetTeddyAcceleratedValues, and the [CompExactlyDependsOn] attributes on IndexOfAnyN2/N3 in AsciiStringSearchValuesTeddyBase.cs.
819c9fe3960 — ProbabilisticMap vectorized SearchValues. Subtle because the layout of the bitmap also branches on the gate (SetCharBit/IsCharBitSet marked [BypassReadyToRun]). Widens the layout choice, ContainsMask16Chars (new PackedSimd branch using ConvertNarrowingSaturateUnsigned), the IndexOfAny/LastIndexOfAny entry dispatcher, and the [CompExactlyDependsOn] on the Vector128 worker methods, consistently.
80522ca629b — Adler32.UpdateVector128 and XxHashShared.MultiplyWideningLower: replace the portable Vector128.Widen + multiply + add sequences with PackedSimd.AddPairwiseWidening + MultiplyWideningLower/Upper and PackedSimd.MultiplyWideningLower (i64x2.extmul_low_i32x4_u). Both were already vectorized on Wasm via the generic else branch; this turns 3–5 ops/iter into 1.

Additional test coverage on browser-wasm (V8 v15)

Suite	Result
`System.Memory.Tests` (post-Teddy, post-ProbabilisticMap)	52,249 / 52,249 ✅
`System.Runtime.Tests` (post-Utf8Transcoding)	67,835 / 67,835 ✅
`System.IO.Hashing.Tests` (post-Adler+XxHash)	4,196 / 4,196 ✅
Host arm64 regression checks	All matched baseline counts ✅

The XxHash lane-swap correctness is end-to-end checked: a wrong shuffle order would corrupt every produced hash.

Remaining gap (not in this PR)

Base64DecoderHelper / Base64EncoderHelper — need pmaddubsw and pmulhuw analogs composed from MultiplyWideningLower/Upper + AddPairwiseWidening or Dot, with careful 8-short lane preservation. Worth its own PR with a benchmark to compare against the scalar path.

Note

The new commits in this PR were drafted with AI/Copilot assistance.

…ow/UnpackHigh PackedSimd.Shuffle wraps i8x16.shuffle which requires its 16 lane indices to be compile-time constants. Mono interpreter accepted a Vector128.Create() constant operand at runtime, but Mono AOT cannot fold it and throws PlatformNotSupportedException at runtime. The same impact was already known and avoided in TeddyHelper.Right Shift1/RightShift2 (see preceding commit on this branch) — use two Vector128.ShuffleNative calls (lowering to PackedSimd.Swizzle, which clamps out-of-range indices to 0) and OR the partial results together. Apply the same pattern in Vector128.UnpackLow/UnpackHigh. This was caught by CI as 50 GuidTests + cascaded reflection-invoke failures under the WasmTestOnChrome-MONO-ST (AOT) leg on PR dotnet#129838. On Mono interpreter all callers (HexConverter, Guid.FormatGuid) had already been validated end-to-end. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

lewing · 2026-06-25T12:00:10Z

CI feedback addressed

The previous WasmTestOnChrome-MONO-ST (AOT) leg failed 50 tests in System.Runtime.Tests with PlatformNotSupportedException from Guid formatting + cascaded reflection-invoke failures (StringBuilder.AppendFormat, String.Format, StrongNameKeyPair, etc.).

Root cause: Vector128.UnpackLow/UnpackHigh's PackedSimd branch used PackedSimd.Shuffle (two-vector i8x16.shuffle), whose 16 lane indices must be compile-time constants. The Mono interpreter accepted the Vector128.Create() constant at runtime; the Mono AOT compiler cannot fold it and throws PNS.

Fix (731effa2b3a): replace with two Vector128.ShuffleNative (single-vector, lowers to PackedSimd.Swizzle which clamps out-of-range indices to 0) plus an OR — the same pattern that already worked in TeddyHelper.RightShift1/RightShift2 earlier in this branch.

Unrelated: The Libraries Test Run release coreclr windows x86 Release leg failed two System.Net.Http.Functional.Tests.SocketsHttpHandler_DiagnosticsTest.SendAsync_Success_ConnectionSetupActivityGraphRecorded cases (HTTP networking, untouched by this PR — likely flaky).

Copilot

Pull request overview

Copilot reviewed 11 out of 11 changed files in this pull request and generated 3 comments.

…RightShift Both helpers previously dispatched to PackedSimd via Vector128.ShuffleNative, which itself has a Ssse3 -> AdvSimd.Arm64 -> PackedSimd if/else chain. The Mono SIMD intrinsic recognizer does not always lower that chain cleanly for less-traveled paths, surfacing as NIY interpreter assertions and runtime startup failures. Call PackedSimd.Swizzle (i8x16.swizzle) directly under the PackedSimd.IsSupported branch. The semantics are identical to ShuffleNative on Wasm (clamps indices >= 16 to 0) but the lowering goes through a single recognized intrinsic, avoiding the dispatcher chain. Validated: System.Memory.Tests on browser-wasm V8 interpreter 52249/52249 (covers TeddyHelper.RightShift1/2). The original NIY OutOfMemoryException:.ctor failure seen in System.Runtime.Tests with the prior ShuffleNative version is gone with this change. AOT behaviour will be re-validated by CI on push. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Reword the LoopTerminatedEarlyDueToNonAsciiData label comment to mention that Wasm is also a little-endian-only platform reaching this point through the PackedSimd branch added earlier in this PR. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Copilot

Copilot encountered an error and was unable to review this pull request. You can try again by re-requesting a review.

Wasm is always little-endian by spec, so the BitConverter.IsLittle Endian check on the PackedSimd.IsSupported branch is a no-op. Keep the check on the AdvSimd.Arm64 branch where it actually matters (NEON can be big- or little-endian on some configurations). Mirrors how the existing Sse41.X64 branch in the same gate has no LE check (x86-64 is also always little-endian). Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Copilot

Pull request overview

Copilot reviewed 11 out of 11 changed files in this pull request and generated 1 comment.

lewing and others added 3 commits June 24, 2026 23:54

Copilot AI review requested due to automatic review settings June 25, 2026 04:54

github-actions Bot added the area-System.Numerics label Jun 25, 2026

Copilot started reviewing on behalf of lewing June 25, 2026 04:55 View session

dotnet-policy-service Bot assigned lewing Jun 25, 2026

lewing changed the title ~~Enable Vector128 fast paths on Wasm: hex/Guid format and Utf8 ASCII validation~~ [wasm] Enable Vector128 fast paths on Wasm: hex/Guid format and Utf8 ASCII validation Jun 25, 2026

Copilot AI reviewed Jun 25, 2026

View reviewed changes

Comment thread src/libraries/System.Private.CoreLib/src/System/Text/Unicode/Utf8Utility.Validation.cs Outdated

lewing and others added 4 commits June 25, 2026 00:07

lewing changed the title ~~[wasm] Enable Vector128 fast paths on Wasm: hex/Guid format and Utf8 ASCII validation~~ [wasm] Enable Vector128 fast paths on Wasm via PackedSimd: hex/Guid, UTF-8, SearchValues, Teddy, Adler32/XXH3 Jun 25, 2026

lewing requested a review from tannergooding June 25, 2026 05:42

Copilot AI review requested due to automatic review settings June 25, 2026 12:00

Copilot started reviewing on behalf of lewing June 25, 2026 12:00 View session

Copilot AI reviewed Jun 25, 2026

View reviewed changes

Comment thread src/libraries/System.Private.CoreLib/src/System/SearchValues/ProbabilisticMap.cs

Comment thread src/libraries/System.Private.CoreLib/src/System/Text/Unicode/Utf8Utility.Validation.cs Outdated

Comment thread src/libraries/System.IO.Hashing/src/System/IO/Hashing/Adler32.cs

lewing and others added 2 commits June 25, 2026 07:36

Copilot AI review requested due to automatic review settings June 25, 2026 12:47

Copilot started reviewing on behalf of lewing June 25, 2026 12:48 View session

Copilot AI reviewed Jun 25, 2026

lewing requested a review from Copilot June 25, 2026 13:12

Copilot started reviewing on behalf of lewing June 25, 2026 13:13 View session

Copilot AI reviewed Jun 25, 2026

View reviewed changes

Comment thread src/libraries/System.Private.CoreLib/src/System/SearchValues/Strings/StringSearchValues.cs

lewing requested a review from radekdoulik June 25, 2026 13:32

pavelsavara reviewed Jun 25, 2026

View reviewed changes

Comment thread src/libraries/Common/src/System/HexConverter.cs

pavelsavara approved these changes Jun 25, 2026

View reviewed changes

tannergooding approved these changes Jun 25, 2026

View reviewed changes

lewing mentioned this pull request Jun 25, 2026

[Perf] Linux/x64: 80 Regressions on 6/24/2026 7:31:32 AM +00:00 dotnet/perf-autofiling-issues#75007

Open

lewing enabled auto-merge (squash) June 25, 2026 14:29

lewing mentioned this pull request Jun 25, 2026

[Perf] Linux/x64: 55 Regressions on 6/24/2026 7:31:32 AM +00:00 dotnet/perf-autofiling-issues#74995

Open

radekdoulik reviewed Jun 25, 2026

View reviewed changes

Comment thread src/libraries/System.Private.CoreLib/src/System/SearchValues/ProbabilisticMap.cs

radekdoulik approved these changes Jun 25, 2026

View reviewed changes

LoopedBard3 mentioned this pull request Jun 25, 2026

[Perf] Linux/arm64: 9 Regressions on 6/24/2026 7:31:32 AM +00:00 #129860

Open

build-analysis Bot mentioned this pull request Jun 25, 2026

The Operation will be canceled. The next steps may not contain expected logs. dotnet/dnceng#3008

Open

3 tasks

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

[wasm] Enable Vector128 fast paths on Wasm via PackedSimd: hex/Guid, UTF-8, SearchValues, Teddy, Adler32/XXH3#129838

[wasm] Enable Vector128 fast paths on Wasm via PackedSimd: hex/Guid, UTF-8, SearchValues, Teddy, Adler32/XXH3#129838
lewing wants to merge 11 commits into
dotnet:mainfrom
lewing:wasm-vector128-fastpaths

lewing commented Jun 25, 2026 •

edited

Loading

Uh oh!

dotnet-policy-service Bot commented Jun 25, 2026

Uh oh!

Copilot AI left a comment

Uh oh!

Uh oh!

lewing commented Jun 25, 2026

Uh oh!

lewing commented Jun 25, 2026

Uh oh!

Copilot AI left a comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Uh oh!

Copilot AI left a comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

Uh oh!

Conversation

lewing commented Jun 25, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Scope and commits

Test results

Files changed

Not in this PR

Uh oh!

dotnet-policy-service Bot commented Jun 25, 2026

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Uh oh!

lewing commented Jun 25, 2026

Additional test coverage on browser-wasm (V8 v15)

Remaining gap (not in this PR)

Uh oh!

lewing commented Jun 25, 2026

CI feedback addressed

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

lewing commented Jun 25, 2026 •

edited

Loading