Adding the x86 part of behavioural testing for std::arch intrinsics #1814
+6,227
−6,189
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
trait to make the same dyn-compatible.
intrinsic-test's X86IntrinsicType struct
madhav-madhusoodanan
force-pushed
the
x86_extension_intrinsic_test
branch
from
23674fb to
c80661a
Compare
Since the bootstrap compiler of Rust is bumped to the commit 5dadfd5c417f0b66816cb7ca662859e2c8751fb3 (version 1.88.0-beta.3 2025-05-11), some features should be safe to enable cfg checks. RISC-V Features: * "zicsr" * "zicntr" * "zihpm" * "zifencei" * "zihintntl" * "zihintpause" * "zimop" * "zicboz" * "zicond" * "ztso" * "zfa" * "zca" * "zcb" * "zcmop" * "b" x86 Features: * "amx-avx512" * "amx-fp8" * "amx-movrs" * "amx-tf32" * "amx-transpose"
- The 2nd argument of the LLVM intrinsic should be IMMARG
- use correct intrinsic for unpackl
- fix invalid use of `simd_{or,and,xor}` on floating point vectors
- `vec_search_string` should require `vector-enhancements-2`
- add `-Zverify-llvm-ir` in testsuite
As the author confirmed as in: <https://lists.infradead.org/pipermail/linux-riscv/2025-May/070844.html>, runtime detection of the Zicntr extension (as in the Linux kernel 6.15) is currently (and technically) redundant on the current base IMA behavior (although can be meaningful if new base behavior is added). This commit implies the Zicntr extension from the base IMA behavior.
It modernizes the coding style of the crate stdarch-gen-arm by fixing
Clippy warnings (except clippy::{collapsible_if,obfuscated_if_else} that
might make the program look worse as a result of "fixing" warnings).
Clippy: rust version 1.89.0-nightly (6f6971078 2025-05-28)
Number of Fixed Warnings: 84/84
Note:
Rust Analyzer double counts one of the Clippy warnings so it reduces
85 warnings (as reported by the Rust Analyzer).
This commit also applies similar technique used to resolve Clippy
warnings but also simplifies identifier name formatting and makes
reading easier.
Confirmed that the exact same code will be generated.
It modernizes the coding style of the crate stdarch-gen-loongarch by fixing Clippy warnings. Clippy: rust version 1.89.0-nightly (6f6971078 2025-05-28) Number of Fixed Warnings: 1/1 Confirmed that the exact same code will be generated (note that, generated.rs in the repository is *not* an exact output but some spaces removed).
It modernizes the coding style of the crate stdarch-test by fixing Clippy warnings. Clippy: rust version 1.89.0-nightly (6f6971078 2025-05-28) Number of Fixed Warnings: 1/1
It modernizes the coding style of the crate stdarch-verify by dealing with Clippy warnings (allows clippy::collapsible_if but review may be required for later changes). Clippy: rust version 1.89.0-nightly (6f6971078 2025-05-28) Number of Fixed Warnings: 4/4
It modernizes the coding style of the crate stdarch_examples (an example "connect5") by fixing Clippy warnings (except clippy::manual_range_contains in which "fixing" the warning will complicate the code). Clippy: rust version 1.89.0-nightly (6f6971078 2025-05-28) Number of Fixed Warnings: 6/6
It modernizes the coding style of the crate intrinsic-test by fixing Clippy warnings. Clippy: rust version 1.89.0-nightly (6f6971078 2025-05-28) Number of Fixed Warnings: 36/36
To make the type names to test correct, this commit replaces occurrences of `rust_prefix` to `c_prefix` where necessary.
It seems it returns true when *no* constraints are found, opposite to the expected behavior of the function name. This commit reverses condition as the name suggests.
This commit adds support for `riscv_hwprobe` on the Linux kernel 6.15. It adds feature detection of 8 extensions (4 of them are new in this). Existing RISC-V Extensions: 1. "Zicntr" 2. "Zihpm" 3. "Zalrsc" 4. "Zaamo" New RISC-V Extensions: 5. "Zicbom" 6. "Zfbfmin" 7. "Zvfbfmin" 8. "Zvfbfwma"
madhav-madhusoodanan
force-pushed
the
x86_extension_intrinsic_test
branch
from
ae4cb22 to
b2777ac
Compare
madhav-madhusoodanan
force-pushed
the
x86_extension_intrinsic_test
branch
from
b2777ac to
7047369
Compare
Comment on lines
49
to
60
| match self { | ||
| Self::BFloat => "bfloat", | ||
| Self::Float => "float", | ||
| Self::Int => "int", | ||
| Self::UInt => "uint", | ||
| Self::Double => "double", | ||
| Self::Int(true) => "int", | ||
| Self::Int(false) => "uint", | ||
| Self::Poly => "poly", | ||
| Self::Void => "void", | ||
| Self::Char(true) => "char", | ||
| Self::Char(false) => "unsigned char", | ||
| Self::Short(true) => "short", | ||
| Self::Short(false) => "unsigned short", |
There was a problem hiding this comment.
@Amanieu @adamgemmell Are we using this trait (fmt::Display for TypeKind) anywhere in the crate? I was searching for its usage and I was unable to find any.
If we are not using this trait, may I delete the definition?
There was a problem hiding this comment.
I suspect this may be useful for debug output during development. However since it isn't used for functionality then it's fine to make any changes you want to it.
There was a problem hiding this comment.
Got it.
|
Pardon the git history, I will be sorting that out shortly |
1. filtered any intrinsic arguments where the only parameter is of type "void" 2. Better parsing for intrinsic return type 3. Filtering intrinsics that are difficult to test (void returns) 4. Reduced the variants in TypeKind (which differed only in bit_len)
madhav-madhusoodanan
force-pushed
the
x86_extension_intrinsic_test
branch
from
f2b3b5d to
d92ce71
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
No description provided.