public inbox for gcc-patches@gcc.gnu.org
 help / color / mirror / Atom feed
From: Cesar Philippidis <cesar@codesourcery.com>
To: Jakub Jelinek <jakub@redhat.com>,
	Chung-Lin Tang <cltang@codesourcery.com>
Cc: "gcc-patches@gcc.gnu.org" <gcc-patches@gcc.gnu.org>
Subject: Re: [OpenACC] Update OpenACC data clause semantics to the 2.5 behavior - runtime
Date: Wed, 20 Jun 2018 16:59:00 -0000	[thread overview]
Message-ID: <06069b0d-46dd-ec3d-d010-6cca6701bec8@codesourcery.com> (raw)
In-Reply-To: <20180620164506.GO7166@tucnak>

On 06/20/2018 09:45 AM, Jakub Jelinek wrote:
> On Tue, Jun 19, 2018 at 10:01:20AM -0700, Cesar Philippidis wrote:

>> >From 53ee03231c5e6e4747b4ef01335079a2d4a98480 Mon Sep 17 00:00:00 2001
>> From: Cesar Philippidis <cesar@codesourcery.com>
>> Date: Tue, 19 Jun 2018 09:33:04 -0700
>> Subject: [PATCH 7/7] runtime changes
>>
>> ---
>>  libgomp/libgomp.h       |   7 +-
>>  libgomp/libgomp.map     |  12 +++
>>  libgomp/oacc-mem.c      | 196 ++++++++++++++++++++++++++++++++-------
>>  libgomp/oacc-parallel.c | 198 ++++++++++++++++++++++++++++++++++------
>>  libgomp/openacc.f90     | 112 +++++++++++++++++++++++
>>  libgomp/openacc.h       |   6 ++
>>  libgomp/openacc_lib.h   |  40 ++++++++
>>  libgomp/target.c        |  41 ++++-----
>>  8 files changed, 528 insertions(+), 84 deletions(-)
>>
>> diff --git a/libgomp/libgomp.h b/libgomp/libgomp.h
>> index 10ea8940c96..3a8cc2bd7d6 100644
>> --- a/libgomp/libgomp.h
>> +++ b/libgomp/libgomp.h
>> @@ -853,6 +853,8 @@ struct splay_tree_key_s {
>>    uintptr_t tgt_offset;
>>    /* Reference count.  */
>>    uintptr_t refcount;
>> +  /* Dynamic reference count.  */
>> +  uintptr_t dynamic_refcount;
>>    /* Pointer to the original mapping of "omp declare target link" object.  */
>>    splay_tree_key link_key;
>>  };
> 
> I'm not entirely happy about this, it grows the structure for not just
> OpenACC, but also OpenMP which will never use it.  Are there any fields
> not used by OpenACC?  E.g. is link_key used?
> Or could the dynamic refcounts be an array allocated (for OpenACC mappings
> only) after the tgt->array array, accessed using
> key->tgt->dynamic_refcounts[key - key->tgt->array] ?
Sorry, I mistakenly committed this patch with the front end changes. Can
I address this issue in a follow up patch?

If it means anything, we have a significant async change that removes
the async_refcount field in that struct.

Cesar

  reply	other threads:[~2018-06-20 16:59 UTC|newest]

Thread overview: 31+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-05-25 20:09 [OpenACC] Update OpenACC data clause semantics to the 2.5 behavior Cesar Philippidis
2018-06-19 16:56 ` Cesar Philippidis
2018-06-19 16:58   ` [OpenACC] Update OpenACC data clause semantics to the 2.5 behavior - C++ Cesar Philippidis
2018-06-20 16:16     ` Jakub Jelinek
2018-06-19 16:59   ` [OpenACC] Update OpenACC data clause semantics to the 2.5 behavior - C Cesar Philippidis
2018-06-20 16:14     ` Jakub Jelinek
2018-06-19 17:00   ` [OpenACC] Update OpenACC data clause semantics to the 2.5 behavior - Fortran Cesar Philippidis
2018-06-20 16:17     ` Jakub Jelinek
2018-06-19 17:01   ` [OpenACC] Update OpenACC data clause semantics to the 2.5 behavior - middle end Cesar Philippidis
2018-06-20 16:20     ` Jakub Jelinek
2018-06-19 17:01   ` [OpenACC] Update OpenACC data clause semantics to the 2.5 behavior - runtime Cesar Philippidis
2018-06-20 16:45     ` Jakub Jelinek
2018-06-20 16:59       ` Cesar Philippidis [this message]
2018-06-20 17:03         ` Jakub Jelinek
2018-06-20 17:07           ` Cesar Philippidis
2019-05-02 14:03     ` Thomas Schwinge
2019-05-29 14:32       ` Thomas Schwinge
2019-12-09 11:44     ` In 'libgomp/target.c:gomp_exit_data', remove open-coded 'gomp_remove_var' (was: [OpenACC] Update OpenACC data clause semantics to the 2.5 behavior - runtime) Thomas Schwinge
2019-12-11 16:59     ` [OpenACC] Update OpenACC data clause semantics to the 2.5 behavior - runtime Thomas Schwinge
2020-05-19 13:58     ` 'gomp_map_vars' locking protocol (was: [OpenACC] Update OpenACC data clause semantics to the 2.5 behavior - runtime) Thomas Schwinge
2018-06-19 17:02   ` [OpenACC] Update OpenACC data clause semantics to the 2.5 behavior - compiler tests Cesar Philippidis
2018-06-19 17:03   ` [OpenACC] Update OpenACC data clause semantics to the 2.5 behavior - runtime tests Cesar Philippidis
2018-12-14 21:25 ` [OpenACC] Update OpenACC data clause semantics to the 2.5 behavior Thomas Schwinge
2019-02-22 11:02   ` Thomas Schwinge
2019-04-29 10:48 ` [GCC 9 RM attention] " Thomas Schwinge
2019-04-29 11:05   ` Jakub Jelinek
2019-04-30  8:28     ` Thomas Schwinge
2019-12-13 14:14 ` [OpenACC] Elaborate/simplify 'exit data' 'finalize' handling (was: [OpenACC] Update OpenACC data clause semantics to the 2.5 behavior) Thomas Schwinge
2019-12-13 23:34   ` Julian Brown
2019-12-18 17:12     ` Thomas Schwinge
2019-12-18 17:13 ` [OpenACC] Update OpenACC data clause semantics to the 2.5 behavior Thomas Schwinge

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=06069b0d-46dd-ec3d-d010-6cca6701bec8@codesourcery.com \
    --to=cesar@codesourcery.com \
    --cc=cltang@codesourcery.com \
    --cc=gcc-patches@gcc.gnu.org \
    --cc=jakub@redhat.com \
    /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).