* [Qemu-devel] [PATCH] softfloat: use GCC builtins to count the leading zeros
@ 2011-04-10 19:12 Aurelien Jarno
0 siblings, 0 replies; only message in thread
From: Aurelien Jarno @ 2011-04-10 19:12 UTC (permalink / raw)
To: qemu-devel; +Cc: Peter Maydell, Aurelien Jarno
Softfloat has its own implementation to count the leading zeros. However
a lot of architectures have either a dedicated instruction or an
optimized to do that. When using GCC >= 3.4, this patch uses GCC builtins
instead of the handcoded implementation.
Note that I amware that QEMU_GNUC_PREREQ is defined in osdep.h and that
clz32() and clz64() are defined in host-utils.h, but I think it is better
to keep the softfloat implemntation self contained.
Cc: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
---
fpu/softfloat-macros.h | 29 +++++++++++++++++++++++++++--
1 files changed, 27 insertions(+), 2 deletions(-)
diff --git a/fpu/softfloat-macros.h b/fpu/softfloat-macros.h
index 3128e60..e82ce23 100644
--- a/fpu/softfloat-macros.h
+++ b/fpu/softfloat-macros.h
@@ -36,6 +36,17 @@ these four paragraphs for those parts of this code that are retained.
=============================================================================*/
/*----------------------------------------------------------------------------
+| This macro tests for minimum version of the GNU C compiler.
+*----------------------------------------------------------------------------*/
+#if defined(__GNUC__) && defined(__GNUC_MINOR__)
+# define SOFTFLOAT_GNUC_PREREQ(maj, min) \
+ ((__GNUC__ << 16) + __GNUC_MINOR__ >= ((maj) << 16) + (min))
+#else
+# define SOFTFLOAT_GNUC_PREREQ(maj, min) 0
+#endif
+
+
+/*----------------------------------------------------------------------------
| Shifts `a' right by the number of bits given in `count'. If any nonzero
| bits are shifted off, they are ``jammed'' into the least significant bit of
| the result by setting the least significant bit to 1. The value of `count'
@@ -616,6 +627,13 @@ static uint32_t estimateSqrt32( int16 aExp, uint32_t a )
static int8 countLeadingZeros32( uint32_t a )
{
+#if SOFTFLOAT_GNUC_PREREQ(3, 4)
+ if (a) {
+ return __builtin_clz(a);
+ } else {
+ return 32;
+ }
+#else
static const int8 countLeadingZerosHigh[] = {
8, 7, 6, 6, 5, 5, 5, 5, 4, 4, 4, 4, 4, 4, 4, 4,
3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,
@@ -647,7 +665,7 @@ static int8 countLeadingZeros32( uint32_t a )
}
shiftCount += countLeadingZerosHigh[ a>>24 ];
return shiftCount;
-
+#endif
}
/*----------------------------------------------------------------------------
@@ -657,6 +675,13 @@ static int8 countLeadingZeros32( uint32_t a )
static int8 countLeadingZeros64( uint64_t a )
{
+#if SOFTFLOAT_GNUC_PREREQ(3, 4)
+ if (a) {
+ return __builtin_clzll(a);
+ } else {
+ return 64;
+ }
+#else
int8 shiftCount;
shiftCount = 0;
@@ -668,7 +693,7 @@ static int8 countLeadingZeros64( uint64_t a )
}
shiftCount += countLeadingZeros32( a );
return shiftCount;
-
+#endif
}
/*----------------------------------------------------------------------------
--
1.7.2.3
^ permalink raw reply related [flat|nested] only message in thread
only message in thread, other threads:[~2011-04-10 19:13 UTC | newest]
Thread overview: (only message) (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2011-04-10 19:12 [Qemu-devel] [PATCH] softfloat: use GCC builtins to count the leading zeros Aurelien Jarno
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).