mirror of
https://github.com/ddidderr/wuffs.git
synced 2026-11-17 16:32:11 +01:00
std/jpeg: add upsample_inv_h2v2 GCC AVX2 comment
This commit is contained in:
1 parent
51bc60ef92
commit
bfc7fc2f45
2 files changed
+6
No files matched your search
@@ -1054,6 +1054,9 @@ wuffs_base__pixel_swizzler__swizzle_ycck(
|
||||
// some unknown reason, GCC performs noticably better on the non-SIMD
|
||||
// version. Possibly because GCC's auto-vectorizer is smarter (just with
|
||||
// SSE2, not AVX2) than our hand-written code, but that's just a guess.
|
||||
//
|
||||
// See commits 51bc60ef9298cb2efc1b29a9681191f66d49820d and
|
||||
// cd769a0cdf1b5affee13f6089b995f3d39569cb4 for benchmark numbers.
|
||||
#else
|
||||
if (wuffs_base__cpu_arch__have_x86_avx2()) {
|
||||
upfuncs[1][1] =
|
||||
|
||||
@@ -24876,6 +24876,9 @@ wuffs_base__pixel_swizzler__swizzle_ycck(
|
||||
// some unknown reason, GCC performs noticably better on the non-SIMD
|
||||
// version. Possibly because GCC's auto-vectorizer is smarter (just with
|
||||
// SSE2, not AVX2) than our hand-written code, but that's just a guess.
|
||||
//
|
||||
// See commits 51bc60ef9298cb2efc1b29a9681191f66d49820d and
|
||||
// cd769a0cdf1b5affee13f6089b995f3d39569cb4 for benchmark numbers.
|
||||
#else
|
||||
if (wuffs_base__cpu_arch__have_x86_avx2()) {
|
||||
upfuncs[1][1] =
|
||||
|
||||
Reference in new issue
Block a user