Use -fno-builtin for sqrt benchmark.

This patch makes the sqrt benchmark use -fno-builtin, as already done
for benchmarks of ffs and ffsll, so that it actually benchmarks the
glibc function as (presumably) intended even in the presence of the
compiler inlining sqrt.

Tested for x86_64 and also used for benchmarking my ARM sqrt patch.

	* benchtests/Makefile (CFLAGS-bench-sqrt.c): New variable.

diff --git a/ChangeLog b/ChangeLog
index d2b0b765785b6d92517714dee020cf53463ad2b0..e41c3e733074d6a7d62e532b72c273868b47e969 100644 (file)
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,7 @@
+2016-10-21  Joseph Myers  <joseph@codesourcery.com>
+
+       * benchtests/Makefile (CFLAGS-bench-sqrt.c): New variable.
+
 2016-10-21  Paul E. Murphy  <murphyp@linux.vnet.ibm.com>
 
        * sysdeps/powerpc/fpu/fenv_private.h:

diff --git a/benchtests/Makefile b/benchtests/Makefile
index 144b32ea3891ce181c54426d5656a787bee03966..5a61522689cb368c9825a0f9a146321a7adee430 100644 (file)
--- a/benchtests/Makefile
+++ b/benchtests/Makefile
@@ -72,6 +72,7 @@ benchset := $(string-benchset-all) $(stdlib-benchset) $(stdio-common-benchset) \
 
 CFLAGS-bench-ffs.c += -fno-builtin
 CFLAGS-bench-ffsll.c += -fno-builtin
+CFLAGS-bench-sqrt.c += -fno-builtin
 
 bench-malloc := malloc-thread