Skip to content

Add missing runtime tests for SSE alias intrinsics #2041

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
ArunTamil21 wants to merge 4 commits into
base: main
Choose a base branch
from
+95 −59
Open

Add missing runtime tests for SSE alias intrinsics #2041

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
de85b1c
Add missing runtime tests for alias intrinsics: _mm_cvt_ss2si, _mm_cv…
ArunTamil21 Feb 26, 2026
b397dac
Refactor alias tests using meta function pattern and add missing test…
ArunTamil21 Feb 26, 2026
b5b01c0
Restore deleted test_mm_store_ps
ArunTamil21 Feb 26, 2026
1a3e798
Remove redundant tests for _mm_prefetch and _mm_undefined_ps
ArunTamil21 Feb 27, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
146 changes: 92 additions & 54 deletions crates/core_arch/src/x86/sse.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2997,24 +2997,28 @@ mod tests {
}
}

#[simd_test(enable = "sse")]
fn test_mm_cvtss_si32() {
fn test_mm_cvtss_si32_impl(f: fn(__m128) -> i32) {
let inputs = &[42.0f32, -3.1, 4.0e10, 4.0e-20, NAN, 2147483500.1];
let result = &[42i32, -3, i32::MIN, 0, i32::MIN, 2147483520];
for i in 0..inputs.len() {
let x = _mm_setr_ps(inputs[i], 1.0, 3.0, 4.0);
let x = unsafe { _mm_setr_ps(inputs[i], 1.0, 3.0, 4.0) };
let e = result[i];
let r = _mm_cvtss_si32(x);
assert_eq!(
e, r,
"TestCase #{} _mm_cvtss_si32({:?}) = {}, expected: {}",
i, x, r, e
);
let r = f(x);
assert_eq!(e, r, "TestCase #{} f({:?}) = {}, expected: {}", i, x, r, e);
}
}

#[simd_test(enable = "sse")]
fn test_mm_cvttss_si32() {
fn test_mm_cvtss_si32() {
test_mm_cvtss_si32_impl(_mm_cvtss_si32);
}

#[simd_test(enable = "sse")]
fn test_mm_cvt_ss2si() {
test_mm_cvtss_si32_impl(_mm_cvt_ss2si);
}

fn test_cvttss_si32_impl(f: fn(__m128) -> i32) {
let inputs = &[
(42.0f32, 42i32),
(-31.4, -31),
Expand All @@ -3028,35 +3032,48 @@ mod tests {
(2147483500.1, 2147483520),
];
for (i, &(xi, e)) in inputs.iter().enumerate() {
let x = _mm_setr_ps(xi, 1.0, 3.0, 4.0);
let r = _mm_cvttss_si32(x);
assert_eq!(
e, r,
"TestCase #{} _mm_cvttss_si32({:?}) = {}, expected: {}",
i, x, r, e
);
let x = unsafe { _mm_setr_ps(xi, 1.0, 3.0, 4.0) };
let r = f(x);
assert_eq!(e, r, "TestCase #{} f({:?}) = {}, expected: {}", i, x, r, e);
}
}

#[simd_test(enable = "sse")]
const fn test_mm_cvtsi32_ss() {
let a = _mm_setr_ps(5.0, 6.0, 7.0, 8.0);
fn test_mm_cvttss_si32() {
test_cvttss_si32_impl(_mm_cvttss_si32);
}

let r = _mm_cvtsi32_ss(a, 4555);
let e = _mm_setr_ps(4555.0, 6.0, 7.0, 8.0);
assert_eq_m128(e, r);
#[simd_test(enable = "sse")]
fn test_mm_cvtt_ss2si() {
test_cvttss_si32_impl(_mm_cvtt_ss2si)
}

let r = _mm_cvtsi32_ss(a, 322223333);
let e = _mm_setr_ps(322223333.0, 6.0, 7.0, 8.0);
assert_eq_m128(e, r);
fn test_mm_cvtsi32_ss_impl(f: fn(__m128, i32) -> __m128) {
unsafe {
let a = _mm_setr_ps(5.0, 6.0, 7.0, 8.0);

let r = _mm_cvtsi32_ss(a, -432);
let e = _mm_setr_ps(-432.0, 6.0, 7.0, 8.0);
assert_eq_m128(e, r);
let r = f(a, 4555);
assert_eq_m128(_mm_setr_ps(4555.0, 6.0, 7.0, 8.0), r);

let r = _mm_cvtsi32_ss(a, -322223333);
let e = _mm_setr_ps(-322223333.0, 6.0, 7.0, 8.0);
assert_eq_m128(e, r);
let r = f(a, 322223333);
assert_eq_m128(_mm_setr_ps(322223333.0, 6.0, 7.0, 8.0), r);

let r = f(a, -432);
assert_eq_m128(_mm_setr_ps(-432.0, 6.0, 7.0, 8.0), r);

let r = f(a, -322223333);
assert_eq_m128(_mm_setr_ps(-322223333.0, 6.0, 7.0, 8.0), r);
}
}

#[simd_test(enable = "sse")]
fn test_mm_cvtsi32_ss() {
test_mm_cvtsi32_ss_impl(_mm_cvtsi32_ss);
}

#[simd_test(enable = "sse")]
fn test_mm_cvt_si2ss() {
test_mm_cvtsi32_ss_impl(_mm_cvt_si2ss);
}

#[simd_test(enable = "sse")]
Expand All @@ -3071,18 +3088,25 @@ mod tests {
assert_eq_m128(r, _mm_setr_ps(4.25, 0.0, 0.0, 0.0));
}

fn test_mm_set1_ps_impl(f: fn(f32) -> __m128) {
unsafe {
let r = f(black_box(4.25));
assert_eq!(get_m128(r, 0), 4.25);
assert_eq!(get_m128(r, 1), 4.25);
assert_eq!(get_m128(r, 2), 4.25);
assert_eq!(get_m128(r, 3), 4.25);
}
}

#[simd_test(enable = "sse")]
fn test_mm_set1_ps() {
test_mm_set1_ps_impl(_mm_set1_ps);
test_mm_set1_ps_impl(_mm_set_ps1);
}

#[simd_test(enable = "sse")]
const fn test_mm_set1_ps() {
let r1 = _mm_set1_ps(black_box(4.25));
let r2 = _mm_set_ps1(black_box(4.25));
assert_eq!(get_m128(r1, 0), 4.25);
assert_eq!(get_m128(r1, 1), 4.25);
assert_eq!(get_m128(r1, 2), 4.25);
assert_eq!(get_m128(r1, 3), 4.25);
assert_eq!(get_m128(r2, 0), 4.25);
assert_eq!(get_m128(r2, 1), 4.25);
assert_eq!(get_m128(r2, 2), 4.25);
assert_eq!(get_m128(r2, 3), 4.25);
fn test_mm_set_ps1() {
test_mm_set1_ps_impl(_mm_set_ps1);
}

#[simd_test(enable = "sse")]
Expand Down Expand Up @@ -3171,11 +3195,20 @@ mod tests {
assert_eq_m128(r, _mm_setr_ps(42.0, 0.0, 0.0, 0.0));
}

#[simd_test(enable = "sse")]
const fn test_mm_load1_ps() {
fn test_mm_load1_ps_impl(f: unsafe fn(*const f32) -> __m128) {
let a = 42.0f32;
let r = unsafe { _mm_load1_ps(ptr::addr_of!(a)) };
assert_eq_m128(r, _mm_setr_ps(42.0, 42.0, 42.0, 42.0));
let r = unsafe { f(ptr::addr_of!(a)) };
unsafe { assert_eq_m128(r, _mm_setr_ps(42.0, 42.0, 42.0, 42.0)) };
}

#[simd_test(enable = "sse")]
fn test_mm_load1_ps() {
test_mm_load1_ps_impl(_mm_load1_ps);
}

#[simd_test(enable = "sse")]
fn test_mm_load_ps1() {
test_mm_load1_ps_impl(_mm_load_ps1);
}

#[simd_test(enable = "sse")]
Expand Down Expand Up @@ -3227,21 +3260,21 @@ mod tests {
assert_eq!(vals[2], 0.0);
}

#[simd_test(enable = "sse")]
const fn test_mm_store1_ps() {
fn test_mm_store1_ps_impl(f: unsafe fn(*mut f32, __m128)) {
let mut vals = Memory { data: [0.0f32; 4] };
let a = _mm_setr_ps(1.0, 2.0, 3.0, 4.0);

// guaranteed to be aligned to 16 bytes
let a = unsafe { _mm_setr_ps(1.0, 2.0, 3.0, 4.0) };
let p = vals.data.as_mut_ptr();

unsafe {
_mm_store1_ps(p, *black_box(&a));
f(p, *black_box(&a));
}

assert_eq!(vals.data, [1.0, 1.0, 1.0, 1.0]);
}

#[simd_test(enable = "sse")]
fn test_mm_store1_ps() {
test_mm_store1_ps_impl(_mm_store1_ps);
}

#[simd_test(enable = "sse")]
const fn test_mm_store_ps() {
let mut vals = Memory { data: [0.0f32; 4] };
Expand All @@ -3257,6 +3290,11 @@ mod tests {
assert_eq!(vals.data, [1.0, 2.0, 3.0, 4.0]);
}

#[simd_test(enable = "sse")]
fn test_mm_store_ps1() {
test_mm_store1_ps_impl(_mm_store_ps1);
}

#[simd_test(enable = "sse")]
const fn test_mm_storer_ps() {
let mut vals = Memory { data: [0.0f32; 4] };
Expand Down
8 changes: 3 additions & 5 deletions crates/stdarch-verify/tests/x86-intel.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -209,11 +209,11 @@ fn verify_all_signatures() {
"_rdseed16_step",
"_rdseed32_step",
"_rdseed64_step",
// Prefetch
"_mm_prefetch",
// CMPXCHG
"cmpxchg16b",
// Undefined
//PREFETCH
"_mm_prefetch",
// Undefined,
"_mm_undefined_ps",
"_mm_undefined_pd",
"_mm_undefined_si128",
Expand Down Expand Up @@ -250,8 +250,6 @@ fn verify_all_signatures() {
"_mm_cvtt_ss2si",
"_mm_cvt_si2ss",
"_mm_set_ps1",
"_mm_load_ps1",
"_mm_store_ps1",
"_mm_bslli_si128",
"_mm_bsrli_si128",
"_bextr2_u32",
Expand Down