8000 ENH: Convert comparison from C universal intrinsics to C++ using Highway by ixgbe · Pull Request #28490 · numpy/numpy · GitHub
[go: up one dir, main page]
Skip to content

ENH: Convert comparison from C universal intrinsics to C++ using Highway #28490

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
wants to merge 14 commits into
base: main
Choose a base branch
from
Open

ENH: Convert comparison from C universal intrinsics to C++ using Highway #28490

Changes from 1 commit
Commits
Show all changes
14 commits
Select commit Hold shift + click to select a range
98c0cc2
ENH: Convert comparison from C universal intrinsics to C++ using Highway
ixgbe Mar 13, 2025
1066b49
Replace OrderedTruncate2To with OrderedDemote2To
ixgbe Mar 13, 2025
faf1bdf
MAINT: Enable building loop_comparision on vector length agnostic arc…
ixgbe Mar 14, 2025
46792b4
fix compile error[1]
ixgbe Mar 15, 2025
62cea2a
fix compile error[2]: cannot convert long unsigned int to unsigned int
ixgbe Mar 17, 2025
2f6aae9
fix test failed[1]: Non-bitwise-equal booleans test failed
ixgbe Mar 17, 2025
fb47971
fix test failed[2]: Non-bitwise-equal booleans test failed
ixgbe Mar 17, 2025
365ca32
fix test failed[3]: Non-bitwise-equal booleans test failed
ixgbe Mar 18, 2025
30dcc0a
fix internal compiler error: in sign_mask, at wide-int.h:855 on ARM S…
ixgbe Mar 18, 2025
e2a3cc6
fix compile error[3]:OrderedTruncate2To is not a member of hn
ixgbe Mar 19, 2025
8008e40
fix compile error[4]:OrderedTruncate2To is not a member of hn
ixgbe Mar 19, 2025
e1c3730
Add Highway tag wrapper to simplify template function development
ixgbe Mar 26, 2025
e07bd6e
Fix compile error for RISC-V architecture
ixgbe Mar 26, 2025
fb9cbd9
Optimize the code based on the Highway wrapper
ixgbe Jun 9, 2025
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
Prev Previous commit
Next Next commit
fix compile error[2]: cannot convert long unsigned int to unsigned int
  • Loading branch information
@ixgbe
ixgbe committed Mar 17, 2025
commit 62cea2aa8d169d3bef88adcca6b4c8445425d465
10 changes: 5 additions & 5 deletions numpy/_core/src/umath/loops_comparison.dispatch.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -32,17 +32,17 @@ using vec_s64 = hn::Vec<decltype(s64)>;
using vec_f32 = hn::Vec<decltype(f32)>;
using vec_f64 = hn::Vec<decltype(f64)>;

HWY_INLINE vec_u8 simd_pack_b8_b16(vec_u16 a, vec_u16 b) {
HWY_INLINE HWY_ATTR vec_u8 simd_pack_b8_b16(vec_u16 a, vec_u16 b) {
return hn::OrderedTruncate2To(u8, a, b);
}

HWY_INLINE vec_u8 simd_pack_b8_b32(vec_u32 a, vec_u32 b, vec_u32 c, vec_u32 d) {
HWY_INLINE HWY_ATTR vec_u8 simd_pack_b8_b32(vec_u32 a, vec_u32 b, vec_u32 c, vec_u32 d) {
auto ab = hn::OrderedTruncate2To(u16, a, b);
auto cd = hn::OrderedTruncate2To(u16, c, d);
return simd_pack_b8_b16(ab, cd);
}

HWY_INLINE vec_u8 simd_pack_b8_b64(vec_u64 a, vec_u64 b, vec_u64 c, vec_u64 d,
HWY_INLINE HWY_ATTR vec_u8 simd_pack_b8_b64(vec_u64 a, vec_u64 b, vec_u64 c, vec_u64 d,
vec_u64 e, vec_u64 f, vec_u64 g, vec_u64 h) {
auto ab = hn::OrderedTruncate2To(u32, a, b);
auto cd = hn::OrderedTruncate2To(u32, c, d);
Expand Down Expand Up @@ -124,7 +124,7 @@ struct TypeTraits<int16_t> {

template<>
struct TypeTraits<uint32_t> {
using ScalarType = npyv_lanetype_u32;
using ScalarType = uint32_t;
using ScalarType2 = npy_uint;
using VecType = vec_u32;
static constexpr auto Tag = u32;
Expand All @@ -137,7 +137,7 @@ struct TypeTraits<uint32_t> {

template<>
struct TypeTraits<int32_t> {
using ScalarType = npyv_lanetype_s32;
using ScalarType = int32_t;
using ScalarType2 = npy_int;
using VecType = vec_s32;
static constexpr auto Tag = s32;
Expand Down
Loading
0