diff options
author | Szabolcs Nagy <szabolcs.nagy@arm.com> | 2024-01-03 09:06:38 +0000 |
---|---|---|
committer | Szabolcs Nagy <szabolcs.nagy@arm.com> | 2024-01-15 14:54:30 +0000 |
commit | 7100d9ae21e2737f96f824259167788aabd8308f (patch) | |
tree | 81932844c6adce1b0e21f06a3e0887729a0cce8f /NEWS | |
parent | 064c708c78cc2a6b5802dce73108fc0c1c6bfc80 (diff) | |
download | glibc-7100d9ae21e2737f96f824259167788aabd8308f.zip glibc-7100d9ae21e2737f96f824259167788aabd8308f.tar.gz glibc-7100d9ae21e2737f96f824259167788aabd8308f.tar.bz2 |
aarch64: Add NEWS entry about libmvec for 2.39
Auto-vectorizing scalar calls is now supported.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Diffstat (limited to 'NEWS')
-rw-r--r-- | NEWS | 7 |
1 files changed, 7 insertions, 0 deletions
@@ -55,6 +55,13 @@ Major new features: unsigned char, unsigned short, unsigned int, unsigned long int and unsigned long long int, and a type-generic macro. +* On AArch64 new symbols were added to libmvec and now math.h has + annotations to allow GCC 9 or newer to auto-vectorize calls to the + following scalar math functions when -ffast-math is specified: + acos, acosf, asin, asinf, atan, atanf, atan2, atan2f, cos, cosf, + exp, expf, exp10, exp10f, exp2, exp2f, expm1, expm1f, log, logf, + log10, log10f, log1p, log1pf, log2, log2f, sin, sinf, tan, tanf. + Deprecated and removed features, and other changes affecting compatibility: * The ldconfig program now skips file names containing ';' or ending in |