[libc-commits] [libc] [libc][math] Implement double precision sin correctly rounded to all rounding modes. (PR #95736)

[via libc-commits]

lntue wrote:

> How does this perform on targets without 128-bit float support? Guessing those skip the final accurate pass?

The 128-bit precision floating point type that we use for accurate pass is not quad precision, and it is emulated using `uint64_t` internally: https://github.com/llvm/llvm-project/blob/main/libc/src/__support/FPUtil/dyadic_float.h#L33

https://github.com/llvm/llvm-project/pull/95736