Add LSX accelerated implementation for source file analysis #145963

heiher · 2025-08-28T15:43:20Z

This patch introduces an LSX-optimized version of analyze_source_file for the loongarch64 target. Similar to existing SSE2 implementation for x86, this version:

Processes 16-byte chunks at a time using LSX vector intrinsics.
Quickly identifies newlines in ASCII-only chunks.
Falls back to the generic implementation when multi-byte UTF-8 characters are detected or in the tail portion.

rustbot · 2025-08-28T15:43:24Z

r? @jackh726

rustbot has assigned @jackh726.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

lqd · 2025-08-28T16:29:04Z

I’m not your reviewer but please add a PR description describing what it’s doing, or what LSX is, to avoid having to infer it from the code, and also why (with some before/after numbers if it’s indeed an optimization like it seems to be)

@rustbot author

rustbot · 2025-08-28T18:04:27Z

Reminder, once the PR becomes ready for a review, use @rustbot ready.

heiher · 2025-08-29T10:02:02Z

@rustbot ready r? @lqd

bors · 2025-08-30T05:34:56Z

☔ The latest upstream changes (presumably #146023) made this pull request unmergeable. Please resolve the merge conflicts.

rustbot · 2025-08-30T06:35:54Z

This PR was rebased onto a different master commit. Here's a range-diff highlighting what actually changed.

Rebasing is a normal part of keeping PRs up to date, so no action is needed—this note is just to help reviewers.

lqd · 2025-09-02T09:45:29Z

compiler/rustc_span/src/analyze_source_file.rs

+            for (chunk_index, chunk) in chunks.iter().enumerate() {
+                let chunk = unsafe { lsx_vld::<0>(chunk.as_ptr() as *const i8) };
+
+                // For character in the chunk, see if its byte value is < 0, which


nit: this is pre-existing from the SSE2 impl, so probably fix it there as well.

Suggested change

// For character in the chunk, see if its byte value is < 0, which

// For each character in the chunk, see if its byte value is < 0, which

lqd · 2025-09-02T09:53:03Z

compiler/rustc_span/src/analyze_source_file.rs

+    }
    _ => {
-        // The target (or compiler version) does not support SSE2 ...
+        // The target (or compiler version) does not support SSE2/LSX ...


something like

Suggested change

// The target (or compiler version) does not support SSE2/LSX ...

// The target (or compiler version) does not support vector instructions our specialized implementations need (x86 SSE2, loongarch64 LSX)...

lqd · 2025-09-02T10:06:02Z

compiler/rustc_span/src/analyze_source_file.rs

+            let mut intra_chunk_offset = 0;
+
+            for (chunk_index, chunk) in chunks.iter().enumerate() {
+                let chunk = unsafe { lsx_vld::<0>(chunk.as_ptr() as *const i8) };


Probably add a comment about vld not having alignment requirements here?

lqd · 2025-09-02T10:10:28Z

Does this yield observable high level gains on the rustc-perf suite?

I don't feel we need a microbenchmark to see how fast the compiler is at recognizing new lines, though it's good to have that data. Thanks for providing it.

I don't know anything about LSX so I may not be a suited reviewer for this, but the implementation matches the SSE2 version very well (to the point of me wondering whether we'll be able to use std::simd one day to remove the duplication; there's only like 3 intrinsics in these functions at most). The intrinsics use seem equivalent to me from docs, and sse to lsx migration guides.

So yeah, this seems good in my uninformed opinion, and we can drop the first commit and fix the couple nits.

@rustbot author

This patch introduces an LSX-optimized version of `analyze_source_file` for the `loongarch64` target. Similar to existing SSE2 implementation for x86, this version: - Processes 16-byte chunks at a time using LSX vector intrinsics. - Quickly identifies newlines in ASCII-only chunks. - Falls back to the generic implementation when multi-byte UTF-8 characters are detected or in the tail portion.

heiher · 2025-09-02T13:04:15Z

Thanks for your review! @rustbot ready

Does this yield observable high level gains on the rustc-perf suite?

I'm not sure, but it's worth checking. I don't know whether rustc-perf is fully usable on LoongArch, or if the overhead changes here would show up in its benchmarks.

Add LSX accelerated implementation for source file analysis This patch introduces an LSX-optimized version of `analyze_source_file` for the `loongarch64` target. Similar to existing SSE2 implementation for x86, this version: - Processes 16-byte chunks at a time using LSX vector intrinsics. - Quickly identifies newlines in ASCII-only chunks. - Falls back to the generic implementation when multi-byte UTF-8 characters are detected or in the tail portion.

Rollup of 9 pull requests Successful merges: - #143725 (core: add Peekable::next_if_map) - #145209 (Stabilize `path_add_extension`) - #145750 (raw_vec.rs: Remove superfluous fn alloc_guard) - #145962 (Ensure we emit an allocator shim when only some crate types need one) - #145963 (Add LSX accelerated implementation for source file analysis) - #146054 (add `#[must_use]` to `array::repeat`) - #146090 (Derive `PartialEq` for `InvisibleOrigin`) - #146120 (Correct typo in `rustc_errors` comment) - #146131 (rustdoc-search: add test case for indexing every item type) r? `@ghost` `@rustbot` modify labels: rollup

Add LSX accelerated implementation for source file analysis This patch introduces an LSX-optimized version of `analyze_source_file` for the `loongarch64` target. Similar to existing SSE2 implementation for x86, this version: - Processes 16-byte chunks at a time using LSX vector intrinsics. - Quickly identifies newlines in ASCII-only chunks. - Falls back to the generic implementation when multi-byte UTF-8 characters are detected or in the tail portion.

Rollup of 15 pull requests Successful merges: - #143725 (core: add Peekable::next_if_map) - #145209 (Stabilize `path_add_extension`) - #145750 (raw_vec.rs: Remove superfluous fn alloc_guard) - #145962 (Ensure we emit an allocator shim when only some crate types need one) - #145963 (Add LSX accelerated implementation for source file analysis) - #146054 (add `#[must_use]` to `array::repeat`) - #146090 (Derive `PartialEq` for `InvisibleOrigin`) - #146120 (Correct typo in `rustc_errors` comment) - #146127 (Rename `ToolRustc` to `ToolRustcPrivate`) - #146131 (rustdoc-search: add test case for indexing every item type) - #146133 (Revert "Make `lto` and `linker-plugin-lto` work the same for `compiler_builtins`) - #146134 (llvm: nvptx: Layout update to match LLVM) - #146136 (docs(std): add missing closing code block fences in doc comments) - #146137 (Disallow frontmatter in `--cfg` and `--check-cfg` arguments) - #146140 (compiletest: cygwin follows windows in using PATH for dynamic libraries) r? `@ghost` `@rustbot` modify labels: rollup

Add LSX accelerated implementation for source file analysis This patch introduces an LSX-optimized version of `analyze_source_file` for the `loongarch64` target. Similar to existing SSE2 implementation for x86, this version: - Processes 16-byte chunks at a time using LSX vector intrinsics. - Quickly identifies newlines in ASCII-only chunks. - Falls back to the generic implementation when multi-byte UTF-8 characters are detected or in the tail portion.

Rollup of 16 pull requests Successful merges: - #143725 (core: add Peekable::next_if_map) - #145209 (Stabilize `path_add_extension`) - #145342 (fix drop scope for `super let` bindings within `if let`) - #145750 (raw_vec.rs: Remove superfluous fn alloc_guard) - #145962 (Ensure we emit an allocator shim when only some crate types need one) - #145963 (Add LSX accelerated implementation for source file analysis) - #146054 (add `#[must_use]` to `array::repeat`) - #146090 (Derive `PartialEq` for `InvisibleOrigin`) - #146120 (Correct typo in `rustc_errors` comment) - #146127 (Rename `ToolRustc` to `ToolRustcPrivate`) - #146133 (Revert "Make `lto` and `linker-plugin-lto` work the same for `compiler_builtins`) - #146134 (llvm: nvptx: Layout update to match LLVM) - #146136 (docs(std): add missing closing code block fences in doc comments) - #146137 (Disallow frontmatter in `--cfg` and `--check-cfg` arguments) - #146140 (compiletest: cygwin follows windows in using PATH for dynamic libraries) - #146156 (miri subtree update) r? `@ghost` `@rustbot` modify labels: rollup

Add LSX accelerated implementation for source file analysis This patch introduces an LSX-optimized version of `analyze_source_file` for the `loongarch64` target. Similar to existing SSE2 implementation for x86, this version: - Processes 16-byte chunks at a time using LSX vector intrinsics. - Quickly identifies newlines in ASCII-only chunks. - Falls back to the generic implementation when multi-byte UTF-8 characters are detected or in the tail portion.

Rollup of 24 pull requests Successful merges: - #140459 (Add `read_buf` equivalents for positioned reads) - #143725 (core: add Peekable::next_if_map) - #145209 (Stabilize `path_add_extension`) - #145342 (fix drop scope for `super let` bindings within `if let`) - #145750 (raw_vec.rs: Remove superfluous fn alloc_guard) - #145827 (On unused binding or binding not present in all patterns, suggest potential typo of unit struct/variant or const) - #145932 (Allow `inline(always)` with a target feature behind a unstable feature `target_feature_inline_always`.) - #145962 (Ensure we emit an allocator shim when only some crate types need one) - #145963 (Add LSX accelerated implementation for source file analysis) - #146054 (add `#[must_use]` to `array::repeat`) - #146090 (Derive `PartialEq` for `InvisibleOrigin`) - #146112 (don't uppercase error messages) - #146120 (Correct typo in `rustc_errors` comment) - #146124 (Test `rustc-dev` in `distcheck`) - #146127 (Rename `ToolRustc` to `ToolRustcPrivate`) - #146131 (rustdoc-search: add test case for indexing every item type) - #146134 (llvm: nvptx: Layout update to match LLVM) - #146136 (docs(std): add missing closing code block fences in doc comments) - #146137 (Disallow frontmatter in `--cfg` and `--check-cfg` arguments) - #146140 (compiletest: cygwin follows windows in using PATH for dynamic libraries) - #146150 (fix(rustdoc): match rustc `--emit` precedence ) - #146155 (Make bootstrap self test parallel) - #146161 ([rustdoc] Uncomment code to add scraped rustdoc examples in loaded paths) - #146172 (triagebot: configure some pings when certain attributes are used) r? `@ghost` `@rustbot` modify labels: rollup

Rollup merge of #145963 - heiher:src-analysis-lsx, r=lqd Add LSX accelerated implementation for source file analysis This patch introduces an LSX-optimized version of `analyze_source_file` for the `loongarch64` target. Similar to existing SSE2 implementation for x86, this version: - Processes 16-byte chunks at a time using LSX vector intrinsics. - Quickly identifies newlines in ASCII-only chunks. - Falls back to the generic implementation when multi-byte UTF-8 characters are detected or in the tail portion.

Rollup of 24 pull requests Successful merges: - rust-lang/rust#140459 (Add `read_buf` equivalents for positioned reads) - rust-lang/rust#143725 (core: add Peekable::next_if_map) - rust-lang/rust#145209 (Stabilize `path_add_extension`) - rust-lang/rust#145342 (fix drop scope for `super let` bindings within `if let`) - rust-lang/rust#145750 (raw_vec.rs: Remove superfluous fn alloc_guard) - rust-lang/rust#145827 (On unused binding or binding not present in all patterns, suggest potential typo of unit struct/variant or const) - rust-lang/rust#145932 (Allow `inline(always)` with a target feature behind a unstable feature `target_feature_inline_always`.) - rust-lang/rust#145962 (Ensure we emit an allocator shim when only some crate types need one) - rust-lang/rust#145963 (Add LSX accelerated implementation for source file analysis) - rust-lang/rust#146054 (add `#[must_use]` to `array::repeat`) - rust-lang/rust#146090 (Derive `PartialEq` for `InvisibleOrigin`) - rust-lang/rust#146112 (don't uppercase error messages) - rust-lang/rust#146120 (Correct typo in `rustc_errors` comment) - rust-lang/rust#146124 (Test `rustc-dev` in `distcheck`) - rust-lang/rust#146127 (Rename `ToolRustc` to `ToolRustcPrivate`) - rust-lang/rust#146131 (rustdoc-search: add test case for indexing every item type) - rust-lang/rust#146134 (llvm: nvptx: Layout update to match LLVM) - rust-lang/rust#146136 (docs(std): add missing closing code block fences in doc comments) - rust-lang/rust#146137 (Disallow frontmatter in `--cfg` and `--check-cfg` arguments) - rust-lang/rust#146140 (compiletest: cygwin follows windows in using PATH for dynamic libraries) - rust-lang/rust#146150 (fix(rustdoc): match rustc `--emit` precedence ) - rust-lang/rust#146155 (Make bootstrap self test parallel) - rust-lang/rust#146161 ([rustdoc] Uncomment code to add scraped rustdoc examples in loaded paths) - rust-lang/rust#146172 (triagebot: configure some pings when certain attributes are used) r? `@ghost` `@rustbot` modify labels: rollup

Gelbpunkt · 2025-09-06T20:47:43Z

This change seems to cause compile errors for me with HEAD @ bea625f:

error[E0308]: mismatched types
   --> compiler/rustc_span/src/analyze_source_file.rs:189:60
    |
189 |                 let multibyte_mask = lsx_vpickve2gr_w::<0>(multibyte_mask);
    |                                      --------------------- ^^^^^^^^^^^^^^ expected `v4i32`, found `v16i8`
    |                                      |
    |                                      arguments to this function are incorrect
    |
note: function defined here
   --> /rustc/788da80fcfcef3f34c90def5baa32813e39a1a41/library/core/src/../../stdarch/crates/core_arch/src/loongarch64/lsx/generated.rs:3825:8

error[E0308]: mismatched types
   --> compiler/rustc_span/src/analyze_source_file.rs:198:67
    |
198 |                     let mut newlines_mask = lsx_vpickve2gr_w::<0>(newlines_mask);
    |                                             --------------------- ^^^^^^^^^^^^^ expected `v4i32`, found `v16i8`
    |                                             |
    |                                             arguments to this function are incorrect
    |
note: function defined here
   --> /rustc/788da80fcfcef3f34c90def5baa32813e39a1a41/library/core/src/../../stdarch/crates/core_arch/src/loongarch64/lsx/generated.rs:3825:8

heiher · 2025-09-07T02:48:48Z

This change seems to cause compile errors for me with HEAD @ bea625f

I’m puzzled why this issue didn’t get caught in CI. More precisely, this patch depends on the compiler after #145042. Both #145042 and the current patch have been merged into the 1.91.0 release line, which seems problematic. Should we delay this patch to the next release instead?

heiher · 2025-09-07T03:18:30Z

I think I've figured out why this issue wasn't caught in CI: LoongArch64, being a cross-compilation target, isn't built with the stage0 compiler. I'm confident this is a real issue in native LoongArch64 builds, so I'll revert it until stage0 is bumped. Sorry for the noise.

Revert "Add LSX accelerated implementation for source file analysis" This reverts commit 5b43244 to fix native build failures on LoongArch. Link: rust-lang#145963 (comment) Link: rust-lang#145963 (comment)

Rollup merge of #146290 - heiher:r-src-analysis-lsx, r=lqd Revert "Add LSX accelerated implementation for source file analysis" This reverts commit 5b43244 to fix native build failures on LoongArch. Link: #145963 (comment) Link: #145963 (comment)

Rollup of 24 pull requests Successful merges: - rust-lang/rust#140459 (Add `read_buf` equivalents for positioned reads) - rust-lang/rust#143725 (core: add Peekable::next_if_map) - rust-lang/rust#145209 (Stabilize `path_add_extension`) - rust-lang/rust#145342 (fix drop scope for `super let` bindings within `if let`) - rust-lang/rust#145750 (raw_vec.rs: Remove superfluous fn alloc_guard) - rust-lang/rust#145827 (On unused binding or binding not present in all patterns, suggest potential typo of unit struct/variant or const) - rust-lang/rust#145932 (Allow `inline(always)` with a target feature behind a unstable feature `target_feature_inline_always`.) - rust-lang/rust#145962 (Ensure we emit an allocator shim when only some crate types need one) - rust-lang/rust#145963 (Add LSX accelerated implementation for source file analysis) - rust-lang/rust#146054 (add `#[must_use]` to `array::repeat`) - rust-lang/rust#146090 (Derive `PartialEq` for `InvisibleOrigin`) - rust-lang/rust#146112 (don't uppercase error messages) - rust-lang/rust#146120 (Correct typo in `rustc_errors` comment) - rust-lang/rust#146124 (Test `rustc-dev` in `distcheck`) - rust-lang/rust#146127 (Rename `ToolRustc` to `ToolRustcPrivate`) - rust-lang/rust#146131 (rustdoc-search: add test case for indexing every item type) - rust-lang/rust#146134 (llvm: nvptx: Layout update to match LLVM) - rust-lang/rust#146136 (docs(std): add missing closing code block fences in doc comments) - rust-lang/rust#146137 (Disallow frontmatter in `--cfg` and `--check-cfg` arguments) - rust-lang/rust#146140 (compiletest: cygwin follows windows in using PATH for dynamic libraries) - rust-lang/rust#146150 (fix(rustdoc): match rustc `--emit` precedence ) - rust-lang/rust#146155 (Make bootstrap self test parallel) - rust-lang/rust#146161 ([rustdoc] Uncomment code to add scraped rustdoc examples in loaded paths) - rust-lang/rust#146172 (triagebot: configure some pings when certain attributes are used) r? `@ghost` `@rustbot` modify labels: rollup

Rollup of 24 pull requests Successful merges: - rust-lang#140459 (Add `read_buf` equivalents for positioned reads) - rust-lang#143725 (core: add Peekable::next_if_map) - rust-lang#145209 (Stabilize `path_add_extension`) - rust-lang#145342 (fix drop scope for `super let` bindings within `if let`) - rust-lang#145750 (raw_vec.rs: Remove superfluous fn alloc_guard) - rust-lang#145827 (On unused binding or binding not present in all patterns, suggest potential typo of unit struct/variant or const) - rust-lang#145932 (Allow `inline(always)` with a target feature behind a unstable feature `target_feature_inline_always`.) - rust-lang#145962 (Ensure we emit an allocator shim when only some crate types need one) - rust-lang#145963 (Add LSX accelerated implementation for source file analysis) - rust-lang#146054 (add `#[must_use]` to `array::repeat`) - rust-lang#146090 (Derive `PartialEq` for `InvisibleOrigin`) - rust-lang#146112 (don't uppercase error messages) - rust-lang#146120 (Correct typo in `rustc_errors` comment) - rust-lang#146124 (Test `rustc-dev` in `distcheck`) - rust-lang#146127 (Rename `ToolRustc` to `ToolRustcPrivate`) - rust-lang#146131 (rustdoc-search: add test case for indexing every item type) - rust-lang#146134 (llvm: nvptx: Layout update to match LLVM) - rust-lang#146136 (docs(std): add missing closing code block fences in doc comments) - rust-lang#146137 (Disallow frontmatter in `--cfg` and `--check-cfg` arguments) - rust-lang#146140 (compiletest: cygwin follows windows in using PATH for dynamic libraries) - rust-lang#146150 (fix(rustdoc): match rustc `--emit` precedence ) - rust-lang#146155 (Make bootstrap self test parallel) - rust-lang#146161 ([rustdoc] Uncomment code to add scraped rustdoc examples in loaded paths) - rust-lang#146172 (triagebot: configure some pings when certain attributes are used) r? `@ghost` `@rustbot` modify labels: rollup

Reland "Add LSX accelerated implementation for source file analysis" This patch introduces an LSX-optimized version of `analyze_source_file` for the `loongarch64` target. Similar to existing SSE2 implementation for x86, this version: - Processes 16-byte chunks at a time using LSX vector intrinsics. - Quickly identifies newlines in ASCII-only chunks. - Falls back to the generic implementation when multi-byte UTF-8 characters are detected or in the tail portion. Reland rust-lang#145963 r? `@lqd`

Reland "Add LSX accelerated implementation for source file analysis" This patch introduces an LSX-optimized version of `analyze_source_file` for the `loongarch64` target. Similar to existing SSE2 implementation for x86, this version: - Processes 16-byte chunks at a time using LSX vector intrinsics. - Quickly identifies newlines in ASCII-only chunks. - Falls back to the generic implementation when multi-byte UTF-8 characters are detected or in the tail portion. Reland rust-lang#145963 r? ``@lqd``

Rollup merge of #147113 - heiher:src-analysis-lsx, r=lqd Reland "Add LSX accelerated implementation for source file analysis" This patch introduces an LSX-optimized version of `analyze_source_file` for the `loongarch64` target. Similar to existing SSE2 implementation for x86, this version: - Processes 16-byte chunks at a time using LSX vector intrinsics. - Quickly identifies newlines in ASCII-only chunks. - Falls back to the generic implementation when multi-byte UTF-8 characters are detected or in the tail portion. Reland #145963 r? ``@lqd``

rustbot assigned jackh726 Aug 28, 2025

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. T-compiler Relevant to the compiler team, which will review and decide on the PR/issue. labels Aug 28, 2025

This comment has been minimized.

Sign in to view

heiher force-pushed the src-analysis-lsx branch from 532e46b to a0b4864 Compare August 28, 2025 15:57

rustbot added S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Aug 28, 2025

heiher force-pushed the src-analysis-lsx branch from a0b4864 to 73287f3 Compare August 29, 2025 08:16

This comment has been minimized.

Sign in to view

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. and removed S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Aug 29, 2025

rustbot assigned lqd and unassigned jackh726 Aug 29, 2025

heiher force-pushed the src-analysis-lsx branch from 73287f3 to 2b6510e Compare August 30, 2025 06:35

lqd reviewed Sep 2, 2025

View reviewed changes

rustbot added S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Sep 2, 2025

heiher force-pushed the src-analysis-lsx branch from 2b6510e to 5b43244 Compare September 2, 2025 13:02

rustbot removed the S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. label Sep 2, 2025

tgross35 mentioned this pull request Sep 3, 2025

Rollup of 9 pull requests #146142

Closed

tgross35 mentioned this pull request Sep 3, 2025

Rollup of 9 pull requests #146147

Closed

jhpratt mentioned this pull request Sep 3, 2025

Rollup of 15 pull requests #146153

Closed

Zalathar mentioned this pull request Sep 3, 2025

Rollup of 16 pull requests #146158

Closed

Zalathar mentioned this pull request Sep 4, 2025

Rollup of 24 pull requests #146185

Merged

bors merged commit 2952b99 into rust-lang:master Sep 4, 2025
10 checks passed

rustbot added this to the 1.91.0 milestone Sep 4, 2025

heiher deleted the src-analysis-lsx branch September 5, 2025 01:37

heiher mentioned this pull request Sep 7, 2025

Revert "Add LSX accelerated implementation for source file analysis" #146290

Merged

heiher mentioned this pull request Sep 28, 2025

Reland "Add LSX accelerated implementation for source file analysis" #147113

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Add LSX accelerated implementation for source file analysis #145963

Add LSX accelerated implementation for source file analysis #145963

heiher commented Aug 28, 2025 •

edited

Loading

Uh oh!

rustbot commented Aug 28, 2025

Uh oh!

This comment has been minimized.

lqd commented Aug 28, 2025 •

edited

Loading

Uh oh!

rustbot commented Aug 28, 2025

Uh oh!

This comment has been minimized.

heiher commented Aug 29, 2025

Uh oh!

bors commented Aug 30, 2025

Uh oh!

rustbot commented Aug 30, 2025

Uh oh!

lqd Sep 2, 2025

Uh oh!

lqd Sep 2, 2025

Uh oh!

lqd Sep 2, 2025

Uh oh!

lqd commented Sep 2, 2025

Uh oh!

heiher commented Sep 2, 2025

Uh oh!

Uh oh!

Gelbpunkt commented Sep 6, 2025

Uh oh!

heiher commented Sep 7, 2025

Uh oh!

heiher commented Sep 7, 2025

Uh oh!

Uh oh!

	// For character in the chunk, see if its byte value is < 0, which
	// For each character in the chunk, see if its byte value is < 0, which

	// The target (or compiler version) does not support SSE2/LSX ...
	// The target (or compiler version) does not support vector instructions our specialized implementations need (x86 SSE2, loongarch64 LSX)...

Add LSX accelerated implementation for source file analysis #145963

Add LSX accelerated implementation for source file analysis #145963

Conversation

heiher commented Aug 28, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rustbot commented Aug 28, 2025

Uh oh!

This comment has been minimized.

lqd commented Aug 28, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rustbot commented Aug 28, 2025

Uh oh!

This comment has been minimized.

heiher commented Aug 29, 2025

Uh oh!

bors commented Aug 30, 2025

Uh oh!

rustbot commented Aug 30, 2025

Uh oh!

lqd Sep 2, 2025

Choose a reason for hiding this comment

Uh oh!

lqd Sep 2, 2025

Choose a reason for hiding this comment

Uh oh!

lqd Sep 2, 2025

Choose a reason for hiding this comment

Uh oh!

lqd commented Sep 2, 2025

Uh oh!

heiher commented Sep 2, 2025

Uh oh!

Uh oh!

Gelbpunkt commented Sep 6, 2025

Uh oh!

heiher commented Sep 7, 2025

Uh oh!

heiher commented Sep 7, 2025

Uh oh!

Uh oh!

heiher commented Aug 28, 2025 •

edited

Loading

lqd commented Aug 28, 2025 •

edited

Loading