bjoernager/rust
bjoernager/rust
1
Fork 0
Code Issues Pull requests Activity
rust/tests/assembly/simd-intrinsic-mask-reduce.rs

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

60 lines
1.8 KiB
Rust
Raw Permalink Normal View History

Add tests for the generated assembly of mask related simd instructions. The tests show that the code generation currently uses the least significant bits of <iX x N> vector masks when converting to <i1 xN>. This leads to an additional left shift operation in the assembly for x86, since mask operations on x86 operate based on the most significant bit. On aarch64 the left shift is followed by a comparison against zero, which repeats the sign bit across the whole lane. The exception, which does not introduce an unneeded shift, is simd_bitmask, because the code generation already shifts before truncating. By using the "C" calling convention the tests should be stable regarding changes in register allocation, but it is possible that future llvm updates will require updating some of the checks. This additional instruction would be removed by the fix in #104693, which uses the most significant bit for all mask operations.
2024-03-02 10:59:11 +01:00
// verify that simd mask reductions do not introduce additional bit shift operations
tests: use minicore more minicore makes it much easier to add new language items to all of the existing `no_core` tests.
2025-02-24 09:26:54 +00:00
//@ add-core-stubs
Add tests for the generated assembly of mask related simd instructions. The tests show that the code generation currently uses the least significant bits of <iX x N> vector masks when converting to <i1 xN>. This leads to an additional left shift operation in the assembly for x86, since mask operations on x86 operate based on the most significant bit. On aarch64 the left shift is followed by a comparison against zero, which repeats the sign bit across the whole lane. The exception, which does not introduce an unneeded shift, is simd_bitmask, because the code generation already shifts before truncating. By using the "C" calling convention the tests should be stable regarding changes in register allocation, but it is possible that future llvm updates will require updating some of the checks. This additional instruction would be removed by the fix in #104693, which uses the most significant bit for all mask operations.
2024-03-02 10:59:11 +01:00
//@ revisions: x86 aarch64
//@ [x86] compile-flags: --target=x86_64-unknown-linux-gnu -C llvm-args=-x86-asm-syntax=intel
Add inline comments why we're forcing the target cpu
2024-05-01 16:54:20 -07:00
// Set the base cpu explicitly, in case the default has been changed.
Use an explicit x86-64 cpu in tests that are sensitive to it There are a few tests that depend on some target features **not** being enabled by default, and usually they are correct with the default x86-64 target CPU. However, in downstream builds we have modified the default to fit our distros -- `x86-64-v2` in RHEL 9 and `x86-64-v3` in RHEL 10 -- and the latter especially trips tests that expect not to have AVX. These cases are few enough that we can just set them back explicitly.
2024-05-01 15:25:26 -07:00
//@ [x86] compile-flags: -C target-cpu=x86-64
Add tests for the generated assembly of mask related simd instructions. The tests show that the code generation currently uses the least significant bits of <iX x N> vector masks when converting to <i1 xN>. This leads to an additional left shift operation in the assembly for x86, since mask operations on x86 operate based on the most significant bit. On aarch64 the left shift is followed by a comparison against zero, which repeats the sign bit across the whole lane. The exception, which does not introduce an unneeded shift, is simd_bitmask, because the code generation already shifts before truncating. By using the "C" calling convention the tests should be stable regarding changes in register allocation, but it is possible that future llvm updates will require updating some of the checks. This additional instruction would be removed by the fix in #104693, which uses the most significant bit for all mask operations.
2024-03-02 10:59:11 +01:00
//@ [x86] needs-llvm-components: x86
//@ [aarch64] compile-flags: --target=aarch64-unknown-linux-gnu
//@ [aarch64] needs-llvm-components: aarch64
//@ assembly-output: emit-asm
tests/assembly: use -Copt-level=3 instead of -O
2025-02-08 18:56:57 -08:00
//@ compile-flags: --crate-type=lib -Copt-level=3 -C panic=abort
Add tests for the generated assembly of mask related simd instructions. The tests show that the code generation currently uses the least significant bits of <iX x N> vector masks when converting to <i1 xN>. This leads to an additional left shift operation in the assembly for x86, since mask operations on x86 operate based on the most significant bit. On aarch64 the left shift is followed by a comparison against zero, which repeats the sign bit across the whole lane. The exception, which does not introduce an unneeded shift, is simd_bitmask, because the code generation already shifts before truncating. By using the "C" calling convention the tests should be stable regarding changes in register allocation, but it is possible that future llvm updates will require updating some of the checks. This additional instruction would be removed by the fix in #104693, which uses the most significant bit for all mask operations.
2024-03-02 10:59:11 +01:00
#![feature(no_core, lang_items, repr_simd, intrinsics)]
#![no_core]
#![allow(non_camel_case_types)]
tests: use minicore more minicore makes it much easier to add new language items to all of the existing `no_core` tests.
2025-02-24 09:26:54 +00:00
extern crate minicore;
use minicore::*;
Add tests for the generated assembly of mask related simd instructions. The tests show that the code generation currently uses the least significant bits of <iX x N> vector masks when converting to <i1 xN>. This leads to an additional left shift operation in the assembly for x86, since mask operations on x86 operate based on the most significant bit. On aarch64 the left shift is followed by a comparison against zero, which repeats the sign bit across the whole lane. The exception, which does not introduce an unneeded shift, is simd_bitmask, because the code generation already shifts before truncating. By using the "C" calling convention the tests should be stable regarding changes in register allocation, but it is possible that future llvm updates will require updating some of the checks. This additional instruction would be removed by the fix in #104693, which uses the most significant bit for all mask operations.
2024-03-02 10:59:11 +01:00
#[repr(simd)]
pub struct mask8x16([i8; 16]);
update/bless tests
2025-04-06 15:12:24 +02:00
#[rustc_intrinsic]
unsafe fn simd_reduce_all<T>(x: T) -> bool;
#[rustc_intrinsic]
unsafe fn simd_reduce_any<T>(x: T) -> bool;
Add tests for the generated assembly of mask related simd instructions. The tests show that the code generation currently uses the least significant bits of <iX x N> vector masks when converting to <i1 xN>. This leads to an additional left shift operation in the assembly for x86, since mask operations on x86 operate based on the most significant bit. On aarch64 the left shift is followed by a comparison against zero, which repeats the sign bit across the whole lane. The exception, which does not introduce an unneeded shift, is simd_bitmask, because the code generation already shifts before truncating. By using the "C" calling convention the tests should be stable regarding changes in register allocation, but it is possible that future llvm updates will require updating some of the checks. This additional instruction would be removed by the fix in #104693, which uses the most significant bit for all mask operations.
2024-03-02 10:59:11 +01:00
// CHECK-LABEL: mask_reduce_all:
#[no_mangle]
pub unsafe extern "C" fn mask_reduce_all(m: mask8x16) -> bool {
Consistently use the most significant bit of vector masks This improves the codegen for vector `select`, `gather`, `scatter` and boolean reduction intrinsics and fixes rust-lang/portable-simd#316. The current behavior of most mask operations during llvm codegen is to truncate the mask vector to <N x i1>, telling llvm to use the least significat bit. The exception is the `simd_bitmask` intrinsics, which already used the most signifiant bit. Since sse/avx instructions are defined to use the most significant bit, truncating means that llvm has to insert a left shift to move the bit into the most significant position, before the mask can actually be used. Similarly on aarch64, mask operations like blend work bit by bit, repeating the least significant bit across the whole lane involves shifting it into the sign position and then comparing against zero. By shifting before truncating to <N x i1>, we tell llvm that we only consider the most significant bit, removing the need for additional shift instructions in the assembly.
2023-01-04 23:55:40 +01:00
// x86-NOT: psllw
// x86: pmovmskb eax, xmm0
// x86-NEXT: {{cmp ax, -1|cmp eax, 65535|xor eax, 65535}}
Add tests for the generated assembly of mask related simd instructions. The tests show that the code generation currently uses the least significant bits of <iX x N> vector masks when converting to <i1 xN>. This leads to an additional left shift operation in the assembly for x86, since mask operations on x86 operate based on the most significant bit. On aarch64 the left shift is followed by a comparison against zero, which repeats the sign bit across the whole lane. The exception, which does not introduce an unneeded shift, is simd_bitmask, because the code generation already shifts before truncating. By using the "C" calling convention the tests should be stable regarding changes in register allocation, but it is possible that future llvm updates will require updating some of the checks. This additional instruction would be removed by the fix in #104693, which uses the most significant bit for all mask operations.
2024-03-02 10:59:11 +01:00
// x86-NEXT: sete al
//
Consistently use the most significant bit of vector masks This improves the codegen for vector `select`, `gather`, `scatter` and boolean reduction intrinsics and fixes rust-lang/portable-simd#316. The current behavior of most mask operations during llvm codegen is to truncate the mask vector to <N x i1>, telling llvm to use the least significat bit. The exception is the `simd_bitmask` intrinsics, which already used the most signifiant bit. Since sse/avx instructions are defined to use the most significant bit, truncating means that llvm has to insert a left shift to move the bit into the most significant position, before the mask can actually be used. Similarly on aarch64, mask operations like blend work bit by bit, repeating the least significant bit across the whole lane involves shifting it into the sign position and then comparing against zero. By shifting before truncating to <N x i1>, we tell llvm that we only consider the most significant bit, removing the need for additional shift instructions in the assembly.
2023-01-04 23:55:40 +01:00
// aarch64-NOT: shl
// aarch64: cmge v0.16b, v0.16b, #0
// aarch64-DAG: mov [[REG1:[a-z0-9]+]], #1
// aarch64-DAG: umaxv b0, v0.16b
// aarch64-NEXT: fmov [[REG2:[a-z0-9]+]], s0
// aarch64-NEXT: bic w0, [[REG1]], [[REG2]]
Add tests for the generated assembly of mask related simd instructions. The tests show that the code generation currently uses the least significant bits of <iX x N> vector masks when converting to <i1 xN>. This leads to an additional left shift operation in the assembly for x86, since mask operations on x86 operate based on the most significant bit. On aarch64 the left shift is followed by a comparison against zero, which repeats the sign bit across the whole lane. The exception, which does not introduce an unneeded shift, is simd_bitmask, because the code generation already shifts before truncating. By using the "C" calling convention the tests should be stable regarding changes in register allocation, but it is possible that future llvm updates will require updating some of the checks. This additional instruction would be removed by the fix in #104693, which uses the most significant bit for all mask operations.
2024-03-02 10:59:11 +01:00
simd_reduce_all(m)
}
// CHECK-LABEL: mask_reduce_any:
#[no_mangle]
pub unsafe extern "C" fn mask_reduce_any(m: mask8x16) -> bool {
Consistently use the most significant bit of vector masks This improves the codegen for vector `select`, `gather`, `scatter` and boolean reduction intrinsics and fixes rust-lang/portable-simd#316. The current behavior of most mask operations during llvm codegen is to truncate the mask vector to <N x i1>, telling llvm to use the least significat bit. The exception is the `simd_bitmask` intrinsics, which already used the most signifiant bit. Since sse/avx instructions are defined to use the most significant bit, truncating means that llvm has to insert a left shift to move the bit into the most significant position, before the mask can actually be used. Similarly on aarch64, mask operations like blend work bit by bit, repeating the least significant bit across the whole lane involves shifting it into the sign position and then comparing against zero. By shifting before truncating to <N x i1>, we tell llvm that we only consider the most significant bit, removing the need for additional shift instructions in the assembly.
2023-01-04 23:55:40 +01:00
// x86-NOT: psllw
// x86: pmovmskb
Add tests for the generated assembly of mask related simd instructions. The tests show that the code generation currently uses the least significant bits of <iX x N> vector masks when converting to <i1 xN>. This leads to an additional left shift operation in the assembly for x86, since mask operations on x86 operate based on the most significant bit. On aarch64 the left shift is followed by a comparison against zero, which repeats the sign bit across the whole lane. The exception, which does not introduce an unneeded shift, is simd_bitmask, because the code generation already shifts before truncating. By using the "C" calling convention the tests should be stable regarding changes in register allocation, but it is possible that future llvm updates will require updating some of the checks. This additional instruction would be removed by the fix in #104693, which uses the most significant bit for all mask operations.
2024-03-02 10:59:11 +01:00
// x86-NEXT: test eax, eax
// x86-NEXT: setne al
//
Consistently use the most significant bit of vector masks This improves the codegen for vector `select`, `gather`, `scatter` and boolean reduction intrinsics and fixes rust-lang/portable-simd#316. The current behavior of most mask operations during llvm codegen is to truncate the mask vector to <N x i1>, telling llvm to use the least significat bit. The exception is the `simd_bitmask` intrinsics, which already used the most signifiant bit. Since sse/avx instructions are defined to use the most significant bit, truncating means that llvm has to insert a left shift to move the bit into the most significant position, before the mask can actually be used. Similarly on aarch64, mask operations like blend work bit by bit, repeating the least significant bit across the whole lane involves shifting it into the sign position and then comparing against zero. By shifting before truncating to <N x i1>, we tell llvm that we only consider the most significant bit, removing the need for additional shift instructions in the assembly.
2023-01-04 23:55:40 +01:00
// aarch64-NOT: shl
// aarch64: cmlt v0.16b, v0.16b, #0
Add tests for the generated assembly of mask related simd instructions. The tests show that the code generation currently uses the least significant bits of <iX x N> vector masks when converting to <i1 xN>. This leads to an additional left shift operation in the assembly for x86, since mask operations on x86 operate based on the most significant bit. On aarch64 the left shift is followed by a comparison against zero, which repeats the sign bit across the whole lane. The exception, which does not introduce an unneeded shift, is simd_bitmask, because the code generation already shifts before truncating. By using the "C" calling convention the tests should be stable regarding changes in register allocation, but it is possible that future llvm updates will require updating some of the checks. This additional instruction would be removed by the fix in #104693, which uses the most significant bit for all mask operations.
2024-03-02 10:59:11 +01:00
// aarch64-NEXT: umaxv b0, v0.16b
// aarch64-NEXT: fmov [[REG:[a-z0-9]+]], s0
// aarch64-NEXT: and w0, [[REG]], #0x1
simd_reduce_any(m)
}
Reference in a new issue Copy permalink