Stowage/go
2
0
Fork 0
mirror of https://github.com/golang/go.git synced 2025-12-08 06:10:04 +00:00
Code Issues Projects Releases Packages Wiki Activity

[dev.simd] simd, cmd/compile: add 128 bit select-from-pair

Browse source 
Using this name until a better one appears:
   x.Select128FromPair(3, 2, y)

Includes test for constant and variable case.
Checks for unexpected immediates (using the zeroing flag,
which is not supported for this intrinsic) and panics.

Change-Id: I9249475d6572968c127b4ee9e00328d717c07578
Reviewed-on: https://go-review.googlesource.com/c/go/+/705496
Reviewed-by: Junyang Shao <shaojunyang@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
This commit is contained in:
David Chase 2025-09-19 13:07:59 -04:00
parent f0e281e693
commit 25c36b95d1
18 changed files with 369 additions and 5 deletions
Download patch file Download diff file Expand all files Collapse all files

56
src/simd/ops_amd64.go
View file

@ -5576,6 +5576,62 @@ func (x Float64x4) Scale(y Float64x4) Float64x4
// Asm: VSCALEFPD, CPU Feature: AVX512
func (x Float64x8) Scale(y Float64x8) Float64x8
/* Select128FromPair */
// Select128FromPair selects the low and high 128-bit halves from the 128-bit halves
// of its two 256-bit inputs, numbering those halves 0, 1, 2, 3.
//
// lo, hi result in better performance when they are constants, non-constant values will be translated into a jump table.
// lo, hi should be between 0 and 3, inclusive; other values will result in a runtime panic.
//
// Asm: VPERM2F128, CPU Feature: AVX
func (x Float32x8) Select128FromPair(lo, hi uint8, y Float32x8) Float32x8
// Select128FromPair selects the low and high 128-bit halves from the 128-bit halves
// of its two 256-bit inputs, numbering those halves 0, 1, 2, 3.
//
// lo, hi result in better performance when they are constants, non-constant values will be translated into a jump table.
// lo, hi should be between 0 and 3, inclusive; other values will result in a runtime panic.
//
// Asm: VPERM2F128, CPU Feature: AVX
func (x Float64x4) Select128FromPair(lo, hi uint8, y Float64x4) Float64x4
// Select128FromPair selects the low and high 128-bit halves from the 128-bit halves
// of its two 256-bit inputs, numbering those halves 0, 1, 2, 3.
//
// lo, hi result in better performance when they are constants, non-constant values will be translated into a jump table.
// lo, hi should be between 0 and 3, inclusive; other values will result in a runtime panic.
//
// Asm: VPERM2I128, CPU Feature: AVX2
func (x Int32x8) Select128FromPair(lo, hi uint8, y Int32x8) Int32x8
// Select128FromPair selects the low and high 128-bit halves from the 128-bit halves
// of its two 256-bit inputs, numbering those halves 0, 1, 2, 3.
//
// lo, hi result in better performance when they are constants, non-constant values will be translated into a jump table.
// lo, hi should be between 0 and 3, inclusive; other values will result in a runtime panic.
//
// Asm: VPERM2I128, CPU Feature: AVX2
func (x Int64x4) Select128FromPair(lo, hi uint8, y Int64x4) Int64x4
// Select128FromPair selects the low and high 128-bit halves from the 128-bit halves
// of its two 256-bit inputs, numbering those halves 0, 1, 2, 3.
//
// lo, hi result in better performance when they are constants, non-constant values will be translated into a jump table.
// lo, hi should be between 0 and 3, inclusive; other values will result in a runtime panic.
//
// Asm: VPERM2I128, CPU Feature: AVX2
func (x Uint32x8) Select128FromPair(lo, hi uint8, y Uint32x8) Uint32x8
// Select128FromPair selects the low and high 128-bit halves from the 128-bit halves
// of its two 256-bit inputs, numbering those halves 0, 1, 2, 3.
//
// lo, hi result in better performance when they are constants, non-constant values will be translated into a jump table.
// lo, hi should be between 0 and 3, inclusive; other values will result in a runtime panic.
//
// Asm: VPERM2I128, CPU Feature: AVX2
func (x Uint64x4) Select128FromPair(lo, hi uint8, y Uint64x4) Uint64x4
/* SetElem */
// SetElem sets a single constant-indexed element's value.