glibc/sysdeps
Joseph Myers ae8372d7e4 Add SSE4.1 trunc, truncf (bug 20142).
This patch adds SSE4.1 versions of trunc and truncf, using the roundsd
/ roundss instructions, similar to the versions of ceil, floor, rint
and nearbyint functions we already have.  In my testing with the glibc
benchtests these are about 30% faster than the C versions for double,
20% faster for float.

Tested for x86_64.

	[BZ #20142]
	* sysdeps/x86_64/fpu/multiarch/Makefile (libm-sysdep_routines):
	Add s_trunc-c, s_truncf-c, s_trunc-sse4_1 and s_truncf-sse4_1.
	* sysdeps/x86_64/fpu/multiarch/s_trunc-c.c: New file.
	* sysdeps/x86_64/fpu/multiarch/s_trunc-sse4_1.S: Likewise.
	* sysdeps/x86_64/fpu/multiarch/s_trunc.c: Likewise.
	* sysdeps/x86_64/fpu/multiarch/s_truncf-c.c: Likewise.
	* sysdeps/x86_64/fpu/multiarch/s_truncf-sse4_1.S: Likewise.
	* sysdeps/x86_64/fpu/multiarch/s_truncf.c: Likewise.
2017-09-20 16:54:05 +00:00
..
aarch64 Enable unwind info in libc-start.c and backtrace.c 2017-09-19 15:07:58 +01:00
alpha Obsolete pow10 functions. 2017-09-01 21:13:18 +00:00
arm Enable unwind info in libc-start.c and backtrace.c 2017-09-19 15:07:58 +01:00
generic Define and use libm_alias_float128. 2017-09-14 01:11:46 +00:00
gnu posix: Sync glob with gnulib [BZ #1062] 2017-09-08 09:39:13 +02:00
hppa Obsolete pow10 functions. 2017-09-01 21:13:18 +00:00
i386 Enable unwind info in libc-start.c and backtrace.c 2017-09-19 15:07:58 +01:00
ia64 Prefer new libm function wrappers for !LIBM_SVID_COMPAT. 2017-09-05 23:35:55 +00:00
ieee754 Fix powerpc64le problem from last ldbl-opt patch. 2017-09-18 23:26:35 +00:00
init_array Update copyright dates with scripts/update-copyrights. 2017-01-01 00:14:16 +00:00
m68k Enable unwind info in libc-start.c and backtrace.c 2017-09-19 15:07:58 +01:00
mach Add missing libc_hidden_weak/def calls 2017-09-17 21:57:39 +02:00
microblaze Obsolete pow10 functions. 2017-09-01 21:13:18 +00:00
mips Enable unwind info in libc-start.c and backtrace.c 2017-09-19 15:07:58 +01:00
nios2 Enable unwind info in libc-start.c and backtrace.c 2017-09-19 15:07:58 +01:00
nptl Remove extra semicolons in struct pthread_mutex (bug 21804) 2017-07-24 12:22:05 +02:00
posix Add missing libc_hidden_weak/def calls 2017-09-17 21:57:39 +02:00
powerpc powerpc: Avoid misaligned stores in memset 2017-09-19 13:55:49 +05:30
pthread aio: Remove internal_function function attribute 2017-08-31 15:59:06 +02:00
s390 Obsolete pow10 functions. 2017-09-01 21:13:18 +00:00
sh Enable unwind info in libc-start.c and backtrace.c 2017-09-19 15:07:58 +01:00
sparc Enable unwind info in libc-start.c and backtrace.c 2017-09-19 15:07:58 +01:00
tile Obsolete pow10 functions. 2017-09-01 21:13:18 +00:00
unix Fix fexecve build where syscall macros call sizeof. 2017-09-19 15:50:38 +00:00
wordsize-32 Build divdi3 only for architecture that required it 2017-04-06 15:14:34 -03:00
wordsize-64 posix: Consolidate Linux glob implementation 2017-09-08 16:34:02 +02:00
x86 x86: Add MathVec_Prefer_No_AVX512 to cpu-features [BZ #21967] 2017-09-12 07:54:47 -07:00
x86_64 Add SSE4.1 trunc, truncf (bug 20142). 2017-09-20 16:54:05 +00:00