This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
[commit] Re-enable two float conversion tests on SPU
- From: "Ulrich Weigand" <uweigand at de dot ibm dot com>
- To: gcc-patches at gcc dot gnu dot org
- Cc: rguenther at suse dot de, joseph at codesourcery dot com
- Date: Tue, 19 Aug 2008 18:15:54 +0200 (CEST)
- Subject: [commit] Re-enable two float conversion tests on SPU
Joseph S. Myers wrote:
> I notice you do not re-enable gcc.dg/torture/fp-int-convert-float.c for
> SPU. As this is a test of consistency of compile-time and run-time
> conversions, I expect it should be possible to re-enable it after this
> patch. This is of course a separate matter from the 4.3 backport.
I forgot about those, thanks. These tests do indeed pass on spu-elf
now on mainline. I've committed the following patch to re-enable
the test cases on SPU on mainline:
gcc.dg/torture/fp-int-convert-float.c
gcc.dg/torture/fp-int-convert-timode.c
These other test cases that I had also disabled for SPU as part of
http://gcc.gnu.org/ml/gcc-patches/2008-06/msg00967.html
remain disabled, as they test specifically for round-to-nearest
behaviour:
gcc.c-torture/execute/ieee/920518-1.c
gcc.c-torture/execute/ieee/20010114-2.c
gcc.c-torture/execute/ieee/20030331-1.c
Bye,
Ulrich
testsuite/ChangeLog:
* gcc.dg/torture/fp-int-convert-float.c: Reenable test on SPU.
* gcc.dg/torture/fp-int-convert-timode.c: Reenable "float" test on SPU.
diff -urNp gcc-head-orig/gcc/testsuite/gcc.dg/torture/fp-int-convert-float.c gcc-head/gcc/testsuite/gcc.dg/torture/fp-int-convert-float.c
--- gcc-head-orig/gcc/testsuite/gcc.dg/torture/fp-int-convert-float.c 2008-07-02 17:26:34.000000000 +0200
+++ gcc-head/gcc/testsuite/gcc.dg/torture/fp-int-convert-float.c 2008-08-19 17:38:36.000000000 +0200
@@ -2,7 +2,6 @@
/* Origin: Joseph Myers <joseph@codesourcery.com> */
/* { dg-do run } */
/* { dg-options "" } */
-/* { dg-skip-if "Round to zero" { spu-*-* } } */
#include <float.h>
#include "fp-int-convert.h"
diff -urNp gcc-head-orig/gcc/testsuite/gcc.dg/torture/fp-int-convert-timode.c gcc-head/gcc/testsuite/gcc.dg/torture/fp-int-convert-timode.c
--- gcc-head-orig/gcc/testsuite/gcc.dg/torture/fp-int-convert-timode.c 2008-07-02 17:26:34.000000000 +0200
+++ gcc-head/gcc/testsuite/gcc.dg/torture/fp-int-convert-timode.c 2008-08-19 17:38:51.000000000 +0200
@@ -9,10 +9,7 @@
int
main (void)
{
-#ifndef __SPU__
- /* Single-precision floating point on SPU always rounds to zero. */
TEST_I_F(TItype, UTItype, float, FLT_MANT_DIG);
-#endif
TEST_I_F(TItype, UTItype, double, DBL_MANT_DIG);
/* Disable the long double tests when using IBM Extended Doubles.
They have variable precision, but constants calculated by gcc's
--
Dr. Ulrich Weigand
GNU Toolchain for Linux on System z and Cell BE
Ulrich.Weigand@de.ibm.com