diff options
author | Daniel Baumann <daniel.baumann@progress-linux.org> | 2024-04-17 12:18:58 +0000 |
---|---|---|
committer | Daniel Baumann <daniel.baumann@progress-linux.org> | 2024-04-17 12:18:58 +0000 |
commit | a4b7ed7a42c716ab9f05e351f003d589124fd55d (patch) | |
tree | b620cd3f223850b28716e474e80c58059dca5dd4 /tests/codegen/simd-wide-sum.rs | |
parent | Adding upstream version 1.67.1+dfsg1. (diff) | |
download | rustc-a4b7ed7a42c716ab9f05e351f003d589124fd55d.tar.xz rustc-a4b7ed7a42c716ab9f05e351f003d589124fd55d.zip |
Adding upstream version 1.68.2+dfsg1.upstream/1.68.2+dfsg1
Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>
Diffstat (limited to 'tests/codegen/simd-wide-sum.rs')
-rw-r--r-- | tests/codegen/simd-wide-sum.rs | 55 |
1 files changed, 55 insertions, 0 deletions
diff --git a/tests/codegen/simd-wide-sum.rs b/tests/codegen/simd-wide-sum.rs new file mode 100644 index 000000000..04314dc29 --- /dev/null +++ b/tests/codegen/simd-wide-sum.rs @@ -0,0 +1,55 @@ +// compile-flags: -C opt-level=3 --edition=2021 +// only-x86_64 +// ignore-debug: the debug assertions get in the way + +#![crate_type = "lib"] +#![feature(portable_simd)] + +use std::simd::{Simd, SimdUint}; +const N: usize = 8; + +#[no_mangle] +// CHECK-LABEL: @wider_reduce_simd +pub fn wider_reduce_simd(x: Simd<u8, N>) -> u16 { + // CHECK: zext <8 x i8> + // CHECK-SAME: to <8 x i16> + // CHECK: call i16 @llvm.vector.reduce.add.v8i16(<8 x i16> + let x: Simd<u16, N> = x.cast(); + x.reduce_sum() +} + +#[no_mangle] +// CHECK-LABEL: @wider_reduce_loop +pub fn wider_reduce_loop(x: Simd<u8, N>) -> u16 { + // CHECK: zext <8 x i8> + // CHECK-SAME: to <8 x i16> + // CHECK: call i16 @llvm.vector.reduce.add.v8i16(<8 x i16> + let mut sum = 0_u16; + for i in 0..N { + sum += u16::from(x[i]); + } + sum +} + +#[no_mangle] +// CHECK-LABEL: @wider_reduce_iter +pub fn wider_reduce_iter(x: Simd<u8, N>) -> u16 { + // CHECK: zext <8 x i8> + // CHECK-SAME: to <8 x i16> + // CHECK: call i16 @llvm.vector.reduce.add.v8i16(<8 x i16> + x.as_array().iter().copied().map(u16::from).sum() +} + +// This iterator one is the most interesting, as it's the one +// which used to not auto-vectorize due to a suboptimality in the +// `<array::IntoIter as Iterator>::fold` implementation. + +#[no_mangle] +// CHECK-LABEL: @wider_reduce_into_iter +pub fn wider_reduce_into_iter(x: Simd<u8, N>) -> u16 { + // FIXME MIR inlining messes up LLVM optimizations. + // WOULD-CHECK: zext <8 x i8> + // WOULD-CHECK-SAME: to <8 x i16> + // WOULD-CHECK: call i16 @llvm.vector.reduce.add.v8i16(<8 x i16> + x.to_array().into_iter().map(u16::from).sum() +} |