Skip to content

sincos: Call into libquadmath instead of calling sin and cos - #119

Merged
ararslan merged 1 commit into
masterfrom
aa/sincos
Oct 2, 2026
Merged

ararslan merged 1 commit into
masterfrom
aa/sincos

Conversation

@ararslan

Copy link
Copy Markdown
Member

On macOS AArch64 (libquadmath sincosq) with Julia 1.12.7, this provides a performance improvement: median 477 ns on this branch vs. 880 ns on master. On FreeBSD AArch64 (libm sincosl) with Julia 1.14.0-DEV.3218, the performance on this branch is identical to master, about 610 ns for both. So wrapping the library function is as good or better than calling sin and cos separately, which is not surprising.

On macOS AArch64 with libquadmath, this provides a performance
improvement: median 477 ns on this branch vs. 880 ns on master. On
FreeBSD AArch64 with libm, the performance on this branch is identical
to master, about 610 ns for both. (Note that the Julia version used on
macOS and FreeBSD here differed: 1.12.7 on macOS and 1.14.0-DEV.3218 on
FreeBSD.)
@ararslan
ararslan merged commit 78df0c1 into master Oct 2, 2026
17 checks passed
@ararslan
ararslan deleted the aa/sincos branch October 2, 2026 16:56
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants