[libc-commits] [libc] [libc][math] implemented dadd and ddiv (PR #100456)

[via libc-commits]

================
@@ -254,6 +254,10 @@ set(TARGET_LIBM_ENTRYPOINTS
     libc.src.math.cosf
     libc.src.math.cosh
     libc.src.math.coshf
+    libc.src.math.daddl
+    libc.src.math.daddf128
+    libc.src.math.ddivl
+    libc.src.math.ddivf128
----------------
overmighty wrote:

`long double` shouldn't be used on GPUs either, actually. AMDGPU will use a different format than the host, and apparently NVIDIA's proprietary compiler will emit a warning if you use `long double` in device code. So none of these entrypoints should be enabled on GPUs.

https://github.com/llvm/llvm-project/pull/100456