public inbox for gcc-patches@gcc.gnu.org
 help / color / mirror / Atom feed
* Re: [PATCH,TESTSUITE]Require 32 bit support for gcc.dg/graphite tests
       [not found]   ` <cb9d34b20911241927i205a5067l31f7215f72108285@mail.gmail.com>
@ 2009-11-25  4:09     ` Sebastian Pop
  0 siblings, 0 replies; 5+ messages in thread
From: Sebastian Pop @ 2009-11-25  4:09 UTC (permalink / raw)
  To: Tobias Grosser; +Cc: andrewhutchinson, GCC Patches

On Tue, Nov 24, 2009 at 07:17, Tobias Grosser <grosser@fim.uni-passau.de> wrote:
>> The patch only affects  for 16 bit targets - which can only have 32767 BYTES in array and have  int range -32768 to +32767.
>
> Yes, that's what I understood, so it is fine with me. I do not
> understand Sebastian concerns. I think you can keep the patch until
> Sebastian explains what is wrong with it.

I misunderstood the intent of this patch.  The patch is ok.

Sebastian

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH,TESTSUITE]Require 32 bit support for gcc.dg/graphite  tests
  2009-11-24  8:20   ` Sebastian Pop
@ 2009-11-24  9:49     ` Tobias Grosser
  0 siblings, 0 replies; 5+ messages in thread
From: Tobias Grosser @ 2009-11-24  9:49 UTC (permalink / raw)
  To: Sebastian Pop; +Cc: janis187, Andrew Hutchinson, gcc-patches

On Tue, 2009-11-24 at 00:47 -0600, Sebastian Pop wrote:
> On Mon, Nov 23, 2009 at 11:47, Janis Johnson <janis187@us.ibm.com> wrote:
> > On Sat, 2009-11-21 at 11:14 -0500, Andrew Hutchinson wrote:
> >> The following patch cleans up tests which must have large array or 32
> >> bit integer support by using require-effective-target size32plus and
> >> int32plus.
> >>
> >>
> >> 2009-11-21  Andy Hutchinson   <hutchinsonandy@gcc.gnu.org>
> >>
> >>     * gcc.dg/graphite/block-1.c: Require 32-bit or larger array support.
> >>     * gcc.dg/graphite/block-3.c: Ditto.
> >>     * gcc.dg/graphite/block-4.c: Ditto.
> >>     * gcc.dg/graphite/block-5.c: Ditto.
> >>     * gcc.dg/graphite/block-6.c: Ditto.
> >>     * gcc.dg/graphite/interchange-0.c: Ditto.
> >>     * gcc.dg/graphite/interchange-1.c: Ditto.
> >>     * gcc.dg/graphite/interchange-2.c: Ditto.
> >>     * gcc.dg/graphite/interchange-3.c: Ditto.
> >>     * gcc.dg/graphite/interchange-4.c: Ditto.
> >>     * gcc.dg/graphite/interchange-5.c: Ditto.
> >>     * gcc.dg/graphite/interchange-6.c: Ditto.
> >>     * gcc.dg/graphite/run-id-1.c: Ditto.
> >>     * gcc.dg/graphite/scop-16.c: Ditto.
> >>     * gcc.dg/graphite/scop-17.c: Ditto.
> >>     * gcc.dg/graphite/scop-18.c: Ditto.
> >>     * gcc.dg/graphite/scop-matmult.c: Ditto.
> >>     * gcc.dg/graphite/pr40157.c: Require 32bit or larger int.
> >>     * gcc.dg/graphite/id-15.c: Ditto.
> >>
> >> Ok for 4.5 head?
> >
> > OK.
> >
> > Janis
> >
> 
> This is not OK.
> 
> We fixed these in the graphite branch by adjusting the size of the arrays.
> Please revert your patch if you applied it to trunk.

Did we? I fixed some test cases to work with 32bit, however it seems
this is for backends that do not even support 32bit.

Tobi

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH,TESTSUITE]Require 32 bit support for gcc.dg/graphite tests
  2009-11-23 17:50 ` Janis Johnson
@ 2009-11-24  8:20   ` Sebastian Pop
  2009-11-24  9:49     ` Tobias Grosser
  0 siblings, 1 reply; 5+ messages in thread
From: Sebastian Pop @ 2009-11-24  8:20 UTC (permalink / raw)
  To: janis187; +Cc: Andrew Hutchinson, gcc-patches

On Mon, Nov 23, 2009 at 11:47, Janis Johnson <janis187@us.ibm.com> wrote:
> On Sat, 2009-11-21 at 11:14 -0500, Andrew Hutchinson wrote:
>> The following patch cleans up tests which must have large array or 32
>> bit integer support by using require-effective-target size32plus and
>> int32plus.
>>
>>
>> 2009-11-21  Andy Hutchinson   <hutchinsonandy@gcc.gnu.org>
>>
>>     * gcc.dg/graphite/block-1.c: Require 32-bit or larger array support.
>>     * gcc.dg/graphite/block-3.c: Ditto.
>>     * gcc.dg/graphite/block-4.c: Ditto.
>>     * gcc.dg/graphite/block-5.c: Ditto.
>>     * gcc.dg/graphite/block-6.c: Ditto.
>>     * gcc.dg/graphite/interchange-0.c: Ditto.
>>     * gcc.dg/graphite/interchange-1.c: Ditto.
>>     * gcc.dg/graphite/interchange-2.c: Ditto.
>>     * gcc.dg/graphite/interchange-3.c: Ditto.
>>     * gcc.dg/graphite/interchange-4.c: Ditto.
>>     * gcc.dg/graphite/interchange-5.c: Ditto.
>>     * gcc.dg/graphite/interchange-6.c: Ditto.
>>     * gcc.dg/graphite/run-id-1.c: Ditto.
>>     * gcc.dg/graphite/scop-16.c: Ditto.
>>     * gcc.dg/graphite/scop-17.c: Ditto.
>>     * gcc.dg/graphite/scop-18.c: Ditto.
>>     * gcc.dg/graphite/scop-matmult.c: Ditto.
>>     * gcc.dg/graphite/pr40157.c: Require 32bit or larger int.
>>     * gcc.dg/graphite/id-15.c: Ditto.
>>
>> Ok for 4.5 head?
>
> OK.
>
> Janis
>

This is not OK.

We fixed these in the graphite branch by adjusting the size of the arrays.
Please revert your patch if you applied it to trunk.

Thanks,
Sebastian

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH,TESTSUITE]Require 32 bit support for gcc.dg/graphite  tests
  2009-11-21 16:37 Andrew Hutchinson
@ 2009-11-23 17:50 ` Janis Johnson
  2009-11-24  8:20   ` Sebastian Pop
  0 siblings, 1 reply; 5+ messages in thread
From: Janis Johnson @ 2009-11-23 17:50 UTC (permalink / raw)
  To: Andrew Hutchinson; +Cc: gcc-patches

On Sat, 2009-11-21 at 11:14 -0500, Andrew Hutchinson wrote:
> The following patch cleans up tests which must have large array or 32 
> bit integer support by using require-effective-target size32plus and 
> int32plus.
> 
> 
> 2009-11-21  Andy Hutchinson   <hutchinsonandy@gcc.gnu.org>
> 
>     * gcc.dg/graphite/block-1.c: Require 32-bit or larger array support.
>     * gcc.dg/graphite/block-3.c: Ditto.
>     * gcc.dg/graphite/block-4.c: Ditto.
>     * gcc.dg/graphite/block-5.c: Ditto.
>     * gcc.dg/graphite/block-6.c: Ditto.
>     * gcc.dg/graphite/interchange-0.c: Ditto.
>     * gcc.dg/graphite/interchange-1.c: Ditto.
>     * gcc.dg/graphite/interchange-2.c: Ditto.
>     * gcc.dg/graphite/interchange-3.c: Ditto.
>     * gcc.dg/graphite/interchange-4.c: Ditto.
>     * gcc.dg/graphite/interchange-5.c: Ditto.
>     * gcc.dg/graphite/interchange-6.c: Ditto.
>     * gcc.dg/graphite/run-id-1.c: Ditto.
>     * gcc.dg/graphite/scop-16.c: Ditto.
>     * gcc.dg/graphite/scop-17.c: Ditto.
>     * gcc.dg/graphite/scop-18.c: Ditto.
>     * gcc.dg/graphite/scop-matmult.c: Ditto.
>     * gcc.dg/graphite/pr40157.c: Require 32bit or larger int.
>     * gcc.dg/graphite/id-15.c: Ditto.
> 
> Ok for 4.5 head?

OK.

Janis

^ permalink raw reply	[flat|nested] 5+ messages in thread

* [PATCH,TESTSUITE]Require 32 bit support for gcc.dg/graphite tests
@ 2009-11-21 16:37 Andrew Hutchinson
  2009-11-23 17:50 ` Janis Johnson
  0 siblings, 1 reply; 5+ messages in thread
From: Andrew Hutchinson @ 2009-11-21 16:37 UTC (permalink / raw)
  To: gcc-patches, janis187

[-- Attachment #1: Type: text/plain, Size: 1097 bytes --]

The following patch cleans up tests which must have large array or 32 
bit integer support by using require-effective-target size32plus and 
int32plus.


2009-11-21  Andy Hutchinson   <hutchinsonandy@gcc.gnu.org>

    * gcc.dg/graphite/block-1.c: Require 32-bit or larger array support.
    * gcc.dg/graphite/block-3.c: Ditto.
    * gcc.dg/graphite/block-4.c: Ditto.
    * gcc.dg/graphite/block-5.c: Ditto.
    * gcc.dg/graphite/block-6.c: Ditto.
    * gcc.dg/graphite/interchange-0.c: Ditto.
    * gcc.dg/graphite/interchange-1.c: Ditto.
    * gcc.dg/graphite/interchange-2.c: Ditto.
    * gcc.dg/graphite/interchange-3.c: Ditto.
    * gcc.dg/graphite/interchange-4.c: Ditto.
    * gcc.dg/graphite/interchange-5.c: Ditto.
    * gcc.dg/graphite/interchange-6.c: Ditto.
    * gcc.dg/graphite/run-id-1.c: Ditto.
    * gcc.dg/graphite/scop-16.c: Ditto.
    * gcc.dg/graphite/scop-17.c: Ditto.
    * gcc.dg/graphite/scop-18.c: Ditto.
    * gcc.dg/graphite/scop-matmult.c: Ditto.
    * gcc.dg/graphite/pr40157.c: Require 32bit or larger int.
    * gcc.dg/graphite/id-15.c: Ditto.

Ok for 4.5 head?





[-- Attachment #2: graphiteclean.patch --]
[-- Type: text/plain, Size: 5418 bytes --]

Index: block-1.c
===================================================================
--- block-1.c	(revision 154399)
+++ block-1.c	(working copy)
@@ -1,3 +1,5 @@
+/* { dg-require-effective-target size32plus } */
+
 #define MAX 8192
 
 void bar (void);
Index: block-3.c
===================================================================
--- block-3.c	(revision 154399)
+++ block-3.c	(working copy)
@@ -1,3 +1,5 @@
+/* { dg-require-effective-target size32plus } */
+
 #define N 24
 #define M 100
 
Index: block-4.c
===================================================================
--- block-4.c	(revision 154399)
+++ block-4.c	(working copy)
@@ -1,3 +1,5 @@
+/* { dg-require-effective-target size32plus } */
+
 #define N 24
 #define M 1000
 
Index: block-5.c
===================================================================
--- block-5.c	(revision 154399)
+++ block-5.c	(working copy)
@@ -1,3 +1,5 @@
+/* { dg-require-effective-target size32plus } */
+
 #define N 10000
 void foo (int);
 int test ()
Index: block-6.c
===================================================================
--- block-6.c	(revision 154399)
+++ block-6.c	(working copy)
@@ -1,3 +1,5 @@
+/* { dg-require-effective-target size32plus } */
+
 #define N 10000
 void foo (int);
 int test ()
Index: id-15.c
===================================================================
--- id-15.c	(revision 154399)
+++ id-15.c	(working copy)
@@ -1,3 +1,5 @@
+/* { dg-require-effective-target int32plus } */
+
 typedef long unsigned int size_t;
 extern void *memset (void *__s, int __c, size_t __n) __attribute__ ((__nothrow__)) __attribute__ ((__nonnull__ (1)));
 
Index: interchange-0.c
===================================================================
--- interchange-0.c	(revision 154399)
+++ interchange-0.c	(working copy)
@@ -1,3 +1,5 @@
+/* { dg-require-effective-target size32plus } */
+
 int a[1000][1000];
 
 int
Index: interchange-1.c
===================================================================
--- interchange-1.c	(revision 154399)
+++ interchange-1.c	(working copy)
@@ -1,3 +1,5 @@
+/* { dg-require-effective-target size32plus } */
+
 /* Formerly known as ltrans-1.c */
 
 double u[1782225];
Index: interchange-2.c
===================================================================
--- interchange-2.c	(revision 154399)
+++ interchange-2.c	(working copy)
@@ -1,3 +1,5 @@
+/* { dg-require-effective-target size32plus } */
+
 /* Formerly known as ltrans-2.c */
 
 double u[1782225];
Index: interchange-3.c
===================================================================
--- interchange-3.c	(revision 154399)
+++ interchange-3.c	(working copy)
@@ -1,3 +1,5 @@
+/* { dg-require-effective-target size32plus } */
+
 /* Formerly known as ltrans-3.c */
 
 double u[1782225];
Index: interchange-4.c
===================================================================
--- interchange-4.c	(revision 154399)
+++ interchange-4.c	(working copy)
@@ -1,3 +1,5 @@
+/* { dg-require-effective-target size32plus } */
+
 /* Formerly known as ltrans-4.c */
 
 double u[1782225];
Index: interchange-5.c
===================================================================
--- interchange-5.c	(revision 154399)
+++ interchange-5.c	(working copy)
@@ -1,3 +1,5 @@
+/* { dg-require-effective-target size32plus } */
+
 /* Formerly known as ltrans-5.c */
 
 int foo ()
Index: interchange-6.c
===================================================================
--- interchange-6.c	(revision 154399)
+++ interchange-6.c	(working copy)
@@ -1,3 +1,5 @@
+/* { dg-require-effective-target size32plus } */
+
 /* Formerly known as ltrans-6.c */
 
 int medium_loop_interchange(int A[100][200])
Index: pr40157.c
===================================================================
--- pr40157.c	(revision 154399)
+++ pr40157.c	(working copy)
@@ -1,4 +1,5 @@
 /* { dg-options "-O2 -fgraphite-identity" } */
+/* { dg-require-effective-target int32plus } */
 
 int buffer[256*256];
 int main(void)
Index: run-id-1.c
===================================================================
--- run-id-1.c	(revision 154399)
+++ run-id-1.c	(working copy)
@@ -1,3 +1,5 @@
+/* { dg-require-effective-target size32plus } */
+
 void abort (void);
 
 void foo (int N)
Index: scop-16.c
===================================================================
--- scop-16.c	(revision 154399)
+++ scop-16.c	(working copy)
@@ -1,3 +1,5 @@
+/* { dg-require-effective-target size32plus } */
+
 #define N 10000
 void foo (int);
 int test ()
Index: scop-17.c
===================================================================
--- scop-17.c	(revision 154399)
+++ scop-17.c	(working copy)
@@ -1,3 +1,5 @@
+/* { dg-require-effective-target size32plus } */
+
 #define N 10000
 void foo (int);
 int test ()
Index: scop-18.c
===================================================================
--- scop-18.c	(revision 154399)
+++ scop-18.c	(working copy)
@@ -1,3 +1,5 @@
+/* { dg-require-effective-target size32plus } */
+
 #define N 24
 #define M 1000
 
Index: scop-matmult.c
===================================================================
--- scop-matmult.c	(revision 154399)
+++ scop-matmult.c	(working copy)
@@ -1,3 +1,5 @@
+/* { dg-require-effective-target size32plus } */
+
 float A[1000][1000], B[1000][1000], C[1000][1000];
 
 /* Multiply two n x n matrices A and B and store the result in C.  */

^ permalink raw reply	[flat|nested] 5+ messages in thread

end of thread, other threads:[~2009-11-25  3:28 UTC | newest]

Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
     [not found] <25362_1259066617_4B0BD4F9_25362_23_1_20091124074335.TJEI7.1094331.imail@eastrmwml36>
     [not found] ` <1259068666.1518.2.camel@tobilaptop.fritz.box>
     [not found]   ` <cb9d34b20911241927i205a5067l31f7215f72108285@mail.gmail.com>
2009-11-25  4:09     ` [PATCH,TESTSUITE]Require 32 bit support for gcc.dg/graphite tests Sebastian Pop
2009-11-21 16:37 Andrew Hutchinson
2009-11-23 17:50 ` Janis Johnson
2009-11-24  8:20   ` Sebastian Pop
2009-11-24  9:49     ` Tobias Grosser

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 read-only IMAP folder(s) and NNTP newsgroup(s).