On 21.10.2024 03:25, LIU Hao wrote:
在 2024-10-19 21:49, Morilli 写道:
This speeds up all floating-point rounding and exception functions
by only querying cpuid once and storing its result.
This helps when such functions are repeatedly called.
This fixes https://sourceforge.net/p/mingw-w64/bugs/995/ .
Signed-off-by: Moritz Bender <35152647+mori...@users.noreply.github.com>
---
diff --git a/mingw-w64-crt/misc/feclearexcept.c
b/mingw-w64-crt/misc/feclearexcept.c
index 673528b20..259809d58 100644
--- a/mingw-w64-crt/misc/feclearexcept.c
+++ b/mingw-w64-crt/misc/feclearexcept.c
@@ -7,9 +7,13 @@
Do you think it's acceptable to drop this check, assuming SSE is
available everywhere on x86?
I'm not sure about 32-bit targets, I'd expect that it would be mostly
fine assumption, but it would contradict efforts to support crts from
win9x era. For 64-bit it's always available, SSE is a requirement there,
we should be able to just skip runtime check there.
BTW, affected functions seem to be available in modern msvcrt (msvcr120
and UCRT), so another solution for the bug would be to move those
function out of mingwex and use implementation provided by OS when possible.
Thanks,
Jacek
_______________________________________________
Mingw-w64-public mailing list
Mingw-w64-public@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/mingw-w64-public
