From: Richard Biener <richard.guenther@gmail.com>
To: Tom de Vries <Tom_deVries@mentor.com>
Cc: "gcc-patches@gnu.org" <gcc-patches@gnu.org>
Subject: Re: [PING^2][PATCH, PR46193] Handle mix/max pointer reductions in parloops
Date: Wed, 26 Aug 2015 08:20:00 -0000 [thread overview]
Message-ID: <CAFiYyc1HaEUL+ou7=-ubyGskAXuk1uSAVzgs8xQ_fDc55Pwyeg@mail.gmail.com> (raw)
In-Reply-To: <55DB33E8.4060907@mentor.com>
On Mon, Aug 24, 2015 at 5:10 PM, Tom de Vries <Tom_deVries@mentor.com> wrote:
> On 22-07-15 20:15, Tom de Vries wrote:
>>
>> On 13/07/15 13:02, Tom de Vries wrote:
>>>
>>> Hi,
>>>
>>> this patch fixes PR46193.
>>>
>>> It handles min and max reductions of pointer type in parloops.
>>>
>>> Bootstrapped and reg-tested on x86_64.
>>>
>>> OK for trunk?
>>>
>>
>
> Ping^2.
>
> Original submission at
> https://gcc.gnu.org/ml/gcc-patches/2015-07/msg01018.html .
Please don't use lower_bound_in_type with two identical types.
Instead use wi::max_value and wide_int_to_tree.
Ok with that change.
Thanks,
Richard.
>
> Thanks,
> - Tom
>
>>
>>> 0001-Handle-mix-max-pointer-reductions-in-parloops.patch
>>>
>>>
>>> Handle mix/max pointer reductions in parloops
>>>
>>> 2015-07-13 Tom de Vries<tom@codesourcery.com>
>>>
>>> PR tree-optimization/46193
>>> * omp-low.c (omp_reduction_init): Handle pointer type for min or max
>>> clause.
>>>
>>> * gcc.dg/autopar/pr46193.c: New test.
>>>
>>> * testsuite/libgomp.c/pr46193.c: New test.
>>> ---
>>> gcc/omp-low.c | 4 ++
>>> gcc/testsuite/gcc.dg/autopar/pr46193.c | 38 +++++++++++++++++++
>>> libgomp/testsuite/libgomp.c/pr46193.c | 67
>>> ++++++++++++++++++++++++++++++++++
>>> 3 files changed, 109 insertions(+)
>>> create mode 100644 gcc/testsuite/gcc.dg/autopar/pr46193.c
>>> create mode 100644 libgomp/testsuite/libgomp.c/pr46193.c
>>>
>>> diff --git a/gcc/omp-low.c b/gcc/omp-low.c
>>> index 2e2070a..20d0010 100644
>>> --- a/gcc/omp-low.c
>>> +++ b/gcc/omp-low.c
>>> @@ -3423,6 +3423,8 @@ omp_reduction_init (tree clause, tree type)
>>> real_maxval (&min, 1, TYPE_MODE (type));
>>> return build_real (type, min);
>>> }
>>> + else if (POINTER_TYPE_P (type))
>>> + return lower_bound_in_type (type, type);
>>> else
>>> {
>>> gcc_assert (INTEGRAL_TYPE_P (type));
>>> @@ -3439,6 +3441,8 @@ omp_reduction_init (tree clause, tree type)
>>> real_maxval (&max, 0, TYPE_MODE (type));
>>> return build_real (type, max);
>>> }
>>> + else if (POINTER_TYPE_P (type))
>>> + return upper_bound_in_type (type, type);
>>> else
>>> {
>>> gcc_assert (INTEGRAL_TYPE_P (type));
>>> diff --git a/gcc/testsuite/gcc.dg/autopar/pr46193.c
>>> b/gcc/testsuite/gcc.dg/autopar/pr46193.c
>>> new file mode 100644
>>> index 0000000..544a5da
>>> --- /dev/null
>>> +++ b/gcc/testsuite/gcc.dg/autopar/pr46193.c
>>> @@ -0,0 +1,38 @@
>>> +/* { dg-do compile } */
>>> +/* { dg-options "-O2 -ftree-parallelize-loops=2
>>> -fdump-tree-parloops-details"
>>> } */
>>> +
>>> +extern void abort (void);
>>> +
>>> +char *
>>> +foo (int count, char **list)
>>> +{
>>> + char *minaddr = list[0];
>>> + int i;
>>> +
>>> + for (i = 0; i < count; i++)
>>> + {
>>> + char *addr = list[i];
>>> + if (addr < minaddr)
>>> + minaddr = addr;
>>> + }
>>> +
>>> + return minaddr;
>>> +}
>>> +
>>> +char *
>>> +foo2 (int count, char **list)
>>> +{
>>> + char *maxaddr = list[0];
>>> + int i;
>>> +
>>> + for (i = 0; i < count; i++)
>>> + {
>>> + char *addr = list[i];
>>> + if (addr > maxaddr)
>>> + maxaddr = addr;
>>> + }
>>> +
>>> + return maxaddr;
>>> +}
>>> +
>>> +/* { dg-final { scan-tree-dump-times "parallelizing inner loop" 2
>>> "parloops"
>>> } } */
>>> diff --git a/libgomp/testsuite/libgomp.c/pr46193.c
>>> b/libgomp/testsuite/libgomp.c/pr46193.c
>>> new file mode 100644
>>> index 0000000..1e27faf
>>> --- /dev/null
>>> +++ b/libgomp/testsuite/libgomp.c/pr46193.c
>>> @@ -0,0 +1,67 @@
>>> +/* { dg-do run } */
>>> +/* { dg-additional-options "-ftree-parallelize-loops=2" } */
>>> +
>>> +extern void abort (void);
>>> +
>>> +char *
>>> +foo (int count, char **list)
>>> +{
>>> + char *minaddr = list[0];
>>> + int i;
>>> +
>>> + for (i = 0; i < count; i++)
>>> + {
>>> + char *addr = list[i];
>>> + if (addr < minaddr)
>>> + minaddr = addr;
>>> + }
>>> +
>>> + return minaddr;
>>> +}
>>> +
>>> +char *
>>> +foo2 (int count, char **list)
>>> +{
>>> + char *maxaddr = list[0];
>>> + int i;
>>> +
>>> + for (i = 0; i < count; i++)
>>> + {
>>> + char *addr = list[i];
>>> + if (addr > maxaddr)
>>> + maxaddr = addr;
>>> + }
>>> +
>>> + return maxaddr;
>>> +}
>>> +
>>> +#define N 5
>>> +
>>> +static void
>>> +init (char **list)
>>> +{
>>> + int i;
>>> + for (i = 0; i < N; ++i)
>>> + list[i] = (char *)&list[i];
>>> +}
>>> +
>>> +int
>>> +main (void)
>>> +{
>>> + char *list[N];
>>> + char * res;
>>> +
>>> + init (list);
>>> +
>>> + res = foo (N, list);
>>> +
>>> + if (res != (char *)&list[0])
>>> + abort ();
>>> +
>>> + res = foo2 (N, list);
>>> +
>>> + if (res != (char *)&list[N-1])
>>> + abort ();
>>> +
>>> + return 0;
>>> +}
>>> -- 1.9.1
>>>
>>
>
next prev parent reply other threads:[~2015-08-26 7:44 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-07-13 11:03 [PATCH, " Tom de Vries
2015-07-22 18:44 ` [PING][PATCH, " Tom de Vries
2015-07-28 7:27 ` [committed, gomp4, " Tom de Vries
2015-08-24 15:11 ` [PING^2][PATCH, " Tom de Vries
2015-08-26 8:20 ` Richard Biener [this message]
2015-08-29 8:03 ` Tom de Vries
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to='CAFiYyc1HaEUL+ou7=-ubyGskAXuk1uSAVzgs8xQ_fDc55Pwyeg@mail.gmail.com' \
--to=richard.guenther@gmail.com \
--cc=Tom_deVries@mentor.com \
--cc=gcc-patches@gnu.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
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).