1
0
Fork 0
mirror of https://github.com/sockspls/badfish synced 2025-05-01 01:03:09 +00:00

ARM lsb/msb assembly

Implement lsb/msb using armv7 assembly instructions.
msb is the easiest one, using a gcc intrinsic that generates
code using the ARM's clz instruction. lsb is also using this
clz instruction, but with the help of ARM's 'rbit' (bit
reversing) instruction. This leads to a >2% speed gain.

I also renamed 'arm-32' to the more meaningfull 'armv7' in the Makefile

No functional change.
This commit is contained in:
Jean-Francois Romang 2012-10-11 22:55:25 +08:00 committed by Marco Costalba
parent 4e7da9be3d
commit 7f9ebf8e86
2 changed files with 22 additions and 7 deletions

View file

@ -126,12 +126,12 @@ ifeq ($(ARCH),x86-32-old)
endif
#arm section
ifeq ($(ARCH),arm-32)
arch = arm-32
ifeq ($(ARCH),armv7)
arch = armv7
os = any
bits = 32
prefetch = yes
bsfq = no
bsfq = yes
popcnt = no
endif
@ -276,7 +276,7 @@ ifeq ($(optimize),yes)
endif
endif
ifeq ($(arch),arm-32)
ifeq ($(arch),armv7)
CXXFLAGS += -fno-gcse
endif
endif
@ -315,7 +315,7 @@ endif
### 3.7 prefetch
ifeq ($(prefetch),yes)
ifneq ($(arch),arm-32)
ifneq ($(arch),armv7)
CXXFLAGS += -msse
DEPENDFLAGS += -msse
endif
@ -376,7 +376,7 @@ help:
@echo "osx-ppc-32 > PPC-Mac OS X 32 bit"
@echo "osx-x86-64 > x86-Mac OS X 64 bit"
@echo "osx-x86-32 > x86-Mac OS X 32 bit"
@echo "arm-32 > ARM 32 bit"
@echo "armv7 > ARMv7 32 bit"
@echo "general-64 > unspecified 64-bit"
@echo "general-32 > unspecified 32-bit"
@echo ""
@ -466,7 +466,7 @@ config-sanity:
@test "$(debug)" = "yes" || test "$(debug)" = "no"
@test "$(optimize)" = "yes" || test "$(optimize)" = "no"
@test "$(arch)" = "any" || test "$(arch)" = "x86_64" || test "$(arch)" = "i386" || \
test "$(arch)" = "ppc64" || test "$(arch)" = "ppc" || test "$(arch)" = "arm-32"
test "$(arch)" = "ppc64" || test "$(arch)" = "ppc" || test "$(arch)" = "armv7"
@test "$(os)" = "any" || test "$(os)" = "osx"
@test "$(bits)" = "32" || test "$(bits)" = "64"
@test "$(prefetch)" = "yes" || test "$(prefetch)" = "no"

View file

@ -247,6 +247,21 @@ FORCE_INLINE Square msb(Bitboard b) {
return (Square) index;
}
# elif defined(__arm__)
FORCE_INLINE int lsb32(uint32_t v) {
__asm__("rbit %0, %1" : "=r"(v) : "r"(v));
return __builtin_clz(v);
}
FORCE_INLINE Square msb(Bitboard b) {
return (Square) (63 - __builtin_clzll(b));
}
FORCE_INLINE Square lsb(Bitboard b) {
return (Square) (uint32_t(b) ? lsb32(uint32_t(b)) : 32 + lsb32(uint32_t(b >> 32)));
}
# else
FORCE_INLINE Square lsb(Bitboard b) { // Assembly code by Heinz van Saanen