Integrate Arm SVE types and intrinsics into stdarch #149297

GWTINYC · 2025-11-25T03:38:58Z

Background:

RFC 3838 introduces a new internal compiler attribute #[rustc_scalable_vector(N)] to define scalable vector types in the Rust standard library.
PR143924 is a preliminary experimental implementation of RFC 3838, aiming to allow types with a single field to be converted to scalable vectors at the codegen backend.

This PR is based on these previous works.

Summary:

This PR integrates Arm SVE (Scalable Vector Extension) types and intrinsics into the Rust stdarch library, enabling support for scalable vector operations on aarch64 targets.

Key changes:

1. Type System Integration

Added SVE type definitions in aarch64/sve/types.rs using #[rustc_scalable_vector(N)] attributes
Fixed module visibility by changing mod sve; to pub mod sve; in aarch64/mod.rs
Proper re-exports in aarch64/sve/mod.rs to ensure types and functions are accessible

2. Type Conversion Layer

Replaced standard SIMD intrinsics with bit-level reinterpretation using core::mem::transmute_copy
Implemented simd_cast and simd_reinterpret as safe bit-level operations for layout-compatible types

3. Selection Operation Implementation

Direct LLVM intrinsic integration via extern "C" with unique link names for each

Extend parsing of `ReprOptions` with `rustc_scalable_vector(N)` which optionally accepts a single literal integral value - the base multiple of lanes that are in a scalable vector. Can only be applied to structs. Co-authored-by: Jamie Cunliffe <Jamie.Cunliffe@arm.com>

Extend well-formedness checking and HIR analysis to prohibit the use of scalable vectors in structs, enums, unions, tuples and arrays. LLVM does not support scalable vectors being members of other types, so these restrictions are necessary. Co-authored-by: Jamie Cunliffe <Jamie.Cunliffe@arm.com>

`simd_reinterpret` is a replacement for `transmute`, specifically for use with scalable SIMD types. It is used in the tests for scalable vectors and in stdarch. Co-authored-by: Jamie Cunliffe <Jamie.Cunliffe@arm.com>

Introduces `BackendRepr::ScalableVector` corresponding to scalable vector types annotated with `repr(scalable)` which lowers to a scalable vector type in LLVM. Co-authored-by: Jamie Cunliffe <Jamie.Cunliffe@arm.com>

LLVM doesn't handle stores on `<vscale x N x i1>` for `N != 16`, a type used internally in SVE intrinsics. Spilling to the stack to create debuginfo will cause errors during instruction selection. These types that are an internal implementation detail to the intrinsic, so users should never see them types and won't need any debuginfo. Co-authored-by: Jamie Cunliffe <Jamie.Cunliffe@arm.com>

Scalable vectors cannot be members of ADTs and thus cannot be kept over await points in async functions.

Scalable vector types only work with the relevant target features enabled, so require this for any function with the types in its signature.

Trivial changes to rust-analyzer to keep it compiling with changes to `ReprOptions`.

This commit introduces a new file containing a comprehensive set of SVE (Scalable Vector Extension) intrinsics for the AArch64 architecture. The intrinsics include functions for absolute comparisons (greater than, less than, equal to) and arithmetic operations (addition, subtraction) for various data types (float, int) and their respective vector representations. Each function is annotated with documentation links to Arm's official documentation and includes test assertions for validation. The addition of these intrinsics enhances the support for scalable vector types in Rust, aligning with recent changes to the compiler and standard library to accommodate scalable SIMD operations. Co-authored-by: Jamie Cunliffe <Jamie.Cunliffe@arm.com>

This commit introduces a new module for Scalable Vector Extension (SVE) intrinsics specific to the AArch64 architecture. It includes the implementation of scalar type conversions, SIMD functions, and core SVE types such as `svbool_t`, `svint8_t`, and their variants. The new files `mod.rs`, `sve2.rs`, and `types.rs` provide foundational support for SVE operations, enhancing the Rust standard library's capabilities for scalable SIMD programming. The implementation includes traits for scalar conversion and type casting, ensuring compatibility with various data types. This addition lays the groundwork for future enhancements and optimizations in SIMD operations within Rust. Co-authored-by: Jamie Cunliffe <Jamie.Cunliffe@arm.com>

…t` ）（`svcmp*`, `svc*` ）（`svaddv`, `svcnt*`, `svclz/cls` ）

…t` ）（`svcmp*`, `svc*` ）（`svaddv`, `svcnt*`, `svclz/cls` ）（`svadr*`, `svld*`, `svst*` ）（`svbic*`, `svbrk*`, `svcmla`）

…arch

…static dispatch for SVE select operations. Update documentation for clarity and maintainability.

…pe conversion documentation. Refactor existing conversion implementations for clarity and consistency.

rustbot · 2025-11-25T03:39:04Z

rust-analyzer is developed in its own repository. If possible, consider making this change to rust-lang/rust-analyzer instead.

cc @rust-lang/rust-analyzer

stdarch is developed in its own repository. If possible, consider making this change to rust-lang/stdarch instead.

cc @Amanieu, @folkertdev, @sayantn

Some changes occurred to the CTFE machinery

cc @RalfJung, @oli-obk, @lcnr

Some changes occurred in compiler/rustc_attr_parsing

cc @jdonszelmann

Some changes occurred to the CTFE / Miri interpreter

cc @rust-lang/miri

Some changes occurred to the intrinsics. Make sure the CTFE / Miri interpreter
gets adapted for the changes, if necessary.

cc @rust-lang/miri, @RalfJung, @oli-obk, @lcnr

Some changes occurred in compiler/rustc_hir/src/attrs

cc @jdonszelmann

Changes to the size of AST and/or HIR nodes.

cc @nnethercote

Some changes occurred in compiler/rustc_codegen_gcc

cc @antoyo, @GuillaumeGomez

Some changes occurred to MIR optimizations

cc @rust-lang/wg-mir-opt

Some changes occurred to the platform-builtins intrinsics. Make sure the
LLVM backend as well as portable-simd gets adapted for the changes.

cc @antoyo, @GuillaumeGomez, @bjorn3, @calebzulawski, @programmerjake

changes to the core type system

cc @compiler-errors, @lcnr

Some changes occurred in compiler/rustc_passes/src/check_attr.rs

cc @jdonszelmann

This PR changes rustc_public

cc @oli-obk, @celinval, @ouz-a

rustbot · 2025-11-25T03:39:07Z

r? @Mark-Simulacrum

rustbot has assigned @Mark-Simulacrum.
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

rustbot · 2025-11-25T03:39:08Z

⚠️ Warning ⚠️

This PR is based on an upstream commit that is older than 28 days.

It's recommended to update your branch according to the rustc-dev-guide.

oli-obk · 2025-11-25T06:39:02Z

Hi. This is the moderation team of the Rust project. The comments left by you are significantly too verbose. While being detailed is good, please be respectful of reviewer time and avoid verbose text that mostly doesn't convey any useful content.

We are closing and locking this PR, but you're welcome to reopen this contribution in a new PR with more concise wording. It is perfectly fine to just leave a one line comment and bullet points otherwise if this is easier for you.

If you would like some help with that, please ensure you're familiar with our contribution guidelines and feel free to either reach out to moderation or ask on the Zulip stream for general help from other contributors.

If you are unsure how concise to make your message, you can include the full message in a <details> tag and let the reviewer know that some information may be missing from the concise message.

Note that this is a moderation warning. Repetitions of such overly verbose messages will result in a ban from contributing to our project. You can contact the moderation team to discuss your warning.

RalfJung · 2025-11-25T07:37:38Z

I would add, please write commit messages in English, so that one has to learn only one language to get around the Rust project.

davidtwco and others added 27 commits August 12, 2025 10:05

core: add simd_reinterpret

12342ef

`simd_reinterpret` is a replacement for `transmute`, specifically for use with scalable SIMD types. It is used in the tests for scalable vectors and in stdarch. Co-authored-by: Jamie Cunliffe <Jamie.Cunliffe@arm.com>

codegen: implement repr(scalable)

dace4f8

Introduces `BackendRepr::ScalableVector` corresponding to scalable vector types annotated with `repr(scalable)` which lowers to a scalable vector type in LLVM. Co-authored-by: Jamie Cunliffe <Jamie.Cunliffe@arm.com>

mir_transform: prohibit scalable vectors in async

86e1b7f

Scalable vectors cannot be members of ADTs and thus cannot be kept over await points in async functions.

mono: require target feature for scalable vectors

95e7b35

Scalable vector types only work with the relevant target features enabled, so require this for any function with the types in its signature.

rust-analyzer: rustc_scalable_vector

26fb3ae

Trivial changes to rust-analyzer to keep it compiling with changes to `ReprOptions`.

直接引用了SVE类型，但没引用函数，能编译

cf3328e

实现了直接引用SVE类型和Intrinsic函数

f8de7cd

sve.rs中新增了除法函数

82f8b65

对SVE类型进行了适当补全

271aa7b

第3批Intrinsics函数补全，累计已补：（svadd/svsub/svmul, svabd, svabs, `svcno…

bdceb61

…t` ）（`svcmp*`, `svc*` ）（`svaddv`, `svcnt*`, `svclz/cls` ）

第5批Intrinsics函数补全，累计已补：（svadd/svsub/svmul, svabd, svabs, `svcno…

191a1f0

…t` ）（`svcmp*`, `svc*` ）（`svaddv`, `svcnt*`, `svclz/cls` ）（`svadr*`, `svld*`, `svst*` ）（`svbic*`, `svbrk*`, `svcmla`）

11.18更新transmute_copy

c0a0913

11.18更新部分函数

85a7a6b

11.19第1批补全

bb708a1

11.19上午补全

ba00c51

11.19第2批补全

43973b2

11.20上午补全

593cf1f

在标准库中集成了代码生成器的sve.rs，支持了更多Intrinsics

0ea7bce

Add new macros for static assertions and extend feature list in core_…

fe5baaa

…arch

Update SVE2 intrinsics in sve2.rs with new functions and documentation;

6fef83b

Refactor SVE module: Enhance type conversion utilities and implement …

27d3aee

…static dispatch for SVE select operations. Update documentation for clarity and maintainability.

Enhance SVE module: Add predicate generation functions and improve ty…

2bdd4d7

…pe conversion documentation. Refactor existing conversion implementations for clarity and consistency.

rustbot assigned Mark-Simulacrum Nov 25, 2025

rustbot added the A-attributes Area: Attributes (`#[…]`, `#![…]`) label Nov 25, 2025

GWTINYC closed this Nov 25, 2025

rustbot removed the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. label Nov 25, 2025

GWTINYC reopened this Nov 25, 2025

rustbot added the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. label Nov 25, 2025

GWTINYC closed this Nov 25, 2025

rustbot removed the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. label Nov 25, 2025

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Integrate Arm SVE types and intrinsics into stdarch #149297

Integrate Arm SVE types and intrinsics into stdarch #149297

GWTINYC commented Nov 25, 2025

Uh oh!

rustbot commented Nov 25, 2025

Uh oh!

rustbot commented Nov 25, 2025

Uh oh!

rustbot commented Nov 25, 2025

Uh oh!

oli-obk commented Nov 25, 2025

Uh oh!

RalfJung commented Nov 25, 2025

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants

Integrate Arm SVE types and intrinsics into stdarch #149297

Integrate Arm SVE types and intrinsics into stdarch #149297

Conversation

GWTINYC commented Nov 25, 2025

Background:

Summary:

Key changes:

1. Type System Integration

2. Type Conversion Layer

3. Selection Operation Implementation

Uh oh!

rustbot commented Nov 25, 2025

Uh oh!

rustbot commented Nov 25, 2025

Uh oh!

rustbot commented Nov 25, 2025

Uh oh!

oli-obk commented Nov 25, 2025

Uh oh!

RalfJung commented Nov 25, 2025

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants