| commit | 8014d1e8321eec40294345d0e855476ea36f5ff6 | |
| author | Joe Ramsay <Joe.Ramsay@arm.com> | |
| Thu, 5 Oct 2023 09:31:38 +0000 (5 10:31 +0100) | ||
| committer | Szabolcs Nagy <szabolcs.nagy@arm.com> | |
| Thu, 5 Oct 2023 15:53:06 +0000 (5 16:53 +0100) | ||
| tree | 6fb25185591415086732b83293eec48e7308d2b4 | treesnapshot (tar.gz zip) |
| parent | 820948edd906ccb475a641ac5c5622e79e7084cf | commitdiff |
aarch64: Improve vecmath sin routines
* Update ULP comment reflecting a new observed max in [-pi/2, pi/2]
* Use the same polynomial in AdvSIMD and SVE, rather than FTRIG instructions
* Improve register use near special-case branch
Also use overloaded intrinsics for SVE.
* Update ULP comment reflecting a new observed max in [-pi/2, pi/2]
* Use the same polynomial in AdvSIMD and SVE, rather than FTRIG instructions
* Improve register use near special-case branch
Also use overloaded intrinsics for SVE.
| sysdeps/aarch64/fpu/sin_advsimd.c | diffblobblamehistory | |
| sysdeps/aarch64/fpu/sin_sve.c | diffblobblamehistory | |
| sysdeps/aarch64/fpu/sinf_sve.c | diffblobblamehistory |