public inbox for gcc-patches@gcc.gnu.org
 help / color / mirror / Atom feed
From: Jakub Jelinek <jakub@redhat.com>
To: Julian Brown <julian@codesourcery.com>
Cc: James Norris <jnorris@codesourcery.com>,
	       GCC Patches <gcc-patches@gcc.gnu.org>,
	       "Joseph S. Myers" <joseph@codesourcery.com>,
	       Nathan Sidwell <Nathan_Sidwell@mentor.com>
Subject: Re: [Bulk] [OpenACC 0/7] host_data construct
Date: Fri, 13 Nov 2015 15:31:00 -0000	[thread overview]
Message-ID: <20151113153118.GR5675@tucnak.redhat.com> (raw)
In-Reply-To: <20151102183339.365c3d33@octopus>

On Mon, Nov 02, 2015 at 06:33:39PM +0000, Julian Brown wrote:
> Firstly, on trunk at least, use_device_ptr variables are restricted to
> pointer or array types: that restriction doesn't exist in OpenACC, nor
> actually could I find it in the OpenMP 4.1 document (my guess is the
> standards are supposed to match in this regard). I think that a program
> such as this should work:

So, after talking about this on omp-lang, it seems there is agreement
that only arrays and pointer types (or reference to arrays or pointers)
should be allowed in use_device_ptr clause and that for pointers/reference
to pointers it should probably act the way I've coded it up, i.e. that
for them it translates the pointer to point to corresponding object to the
one to which it points on the host.  It is too late to change the standard
now, but will be changed soon, and hopefully clarified in examples.

> void target_fn (int *targ_data);
> 
> int
> main (int argc, char *argv[])
> {
>   char out;
>   int myvar;
> #pragma omp target enter data map(to: myvar)
> 
> #pragma omp target data use_device_ptr(myvar) map(from:out)
>   {
>     target_fn (&myvar);
>     out = 5;
>   }
> 
>   return 0;
> }

That would make the above non-conforming for OpenMP.

> Secondly, attempts to use use_device_ptr on (e.g.
> dynamically-allocated) arrays accessed through a pointer cause an ICE
> with the existing trunk OpenMP code:
> 
> #include <stdlib.h>
> 
> void target_fn (char *targ_data);
> 
> int
> main (int argc, char *argv[])
> {
>   char *myarr, out;
> 
>   myarr = malloc (1024);
> 
> #pragma omp target data map(to: myarr[0:1024])
>   {
> #pragma omp target data use_device_ptr(myarr) map(from:out)
>     {
>       target_fn (myarr);
>       out = 5;
>     }
>   }
> 
>   return 0;
> }

Can't reproduce this ICE (at least not on gomp-4_5-branch, but there
aren't significant changes from the trunk there).

> Furthermore, this looks strange to me (006t.omplower):
> 
>   .omp_data_arr.5.out = &out;
>   myarr.8 = myarr;
>   .omp_data_arr.5.myarr = myarr.8;
>   #pragma omp target data map(from:out [len: 1]) use_device_ptr(myarr)
>     {
>       D.2436 = .omp_data_arr.5.myarr;
>       myarr = D.2436;
> 
> That's clobbering the original myarr variable, right?

Just use -fdump-tree-omplower-uid to see that it is a different variable.
Basically, for OpenMP use_device_ptr creates a private copy of the
pointer for the body of the target data construct, and that pointer is
assigned the target device's address.  For arrays the implementation
creates an artificial pointer variable (holding the start of the array
initially) and replaces all references to the array in the target data
body with dereference of the pointer.

	Jakub

  parent reply	other threads:[~2015-11-13 15:31 UTC|newest]

Thread overview: 33+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-10-22 19:14 James Norris
2015-10-22 19:15 ` [OpenACC 1/7] host_data construct (C/C++ common) James Norris
2015-10-22 19:15 ` [OpenACC 2/7] host_data construct (C FE) James Norris
2015-10-22 19:16 ` [OpenACC 3/7] host_data construct (C front-end) James Norris
2015-10-22 19:18 ` [OpenACC 4/7] host_data construct (middle end) James Norris
2015-10-22 19:19 ` [OpenACC 5/7] host_data construct (gcc tests) James Norris
2015-10-22 19:20 ` [OpenACC 6/7] host_data construct James Norris
2015-10-22 19:22 ` [OpenACC 7/7] host_data construct (runtime tests) James Norris
2015-10-22 20:42 ` [OpenACC 0/7] host_data construct Joseph Myers
2015-10-22 20:53   ` James Norris
2015-10-23 16:01 ` [Bulk] " James Norris
2015-10-26 18:36   ` Jakub Jelinek
2015-10-27 15:57     ` Cesar Philippidis
2015-11-02 18:33     ` Julian Brown
2015-11-02 19:29       ` Jakub Jelinek
2015-11-12 11:16       ` Julian Brown
2015-11-18 12:48         ` Julian Brown
2015-11-19 13:13           ` Jakub Jelinek
2015-11-19 14:29             ` Julian Brown
2015-11-19 15:57               ` Jakub Jelinek
2015-11-30 19:34                 ` Julian Brown
2015-12-01  8:30                   ` Jakub Jelinek
2015-12-02 15:27                   ` Tom de Vries
2015-12-02 15:59                   ` Thomas Schwinge
2015-12-02 19:16                     ` Cesar Philippidis
2015-12-02 19:28                       ` Steve Kargl
2015-12-02 19:35                       ` Jakub Jelinek
2015-12-02 19:54                         ` Cesar Philippidis
2015-12-02 22:14                     ` [gomp4] " Thomas Schwinge
2016-04-08 13:41                       ` Fortran OpenACC host_data construct ICE (was: [gomp4] Re: [OpenACC 0/7] host_data construct) Thomas Schwinge
2016-02-02 13:57                     ` [OpenACC 0/7] host_data construct Thomas Schwinge
2015-11-13 15:31       ` Jakub Jelinek [this message]
2015-12-23 11:02     ` [Bulk] " 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=20151113153118.GR5675@tucnak.redhat.com \
    --to=jakub@redhat.com \
    --cc=Nathan_Sidwell@mentor.com \
    --cc=gcc-patches@gcc.gnu.org \
    --cc=jnorris@codesourcery.com \
    --cc=joseph@codesourcery.com \
    --cc=julian@codesourcery.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).