[gomp4] Redesign oacc_parallel launch API

[Thomas Schwinge]

Hi Nathan!

On Wed, 29 Jul 2015 17:45:32 -0400, Nathan Sidwell <nathan@acm.org> wrote:
> On 07/29/15 08:24, Nathan Sidwell wrote:
> > On 07/29/15 05:22, Thomas Schwinge wrote:
> 
> >> Likewise for the other torture testing flags.
> >
> >
> > Investigating ...  (I've seen those failures be intermittent)

(For me, they were persistent, within the handful of testsuite runs I'd
done.)

> Interestingly the fails go away with an unoptimized libgomp.  I've observed 
> something vaguely like that before.  The observed failure mode was getting stuck 
> inside the driver library opening the device.  Which is very strange.

Uh...  :-/

> Anyway, I've committed the attached to gomp4 branch, which separates the ASYNC 
> and WAIT tags, for a slightly better interface.

Thanks!

At first, I saw a number of spurious FAILs and timeouts with different
test cases each time I ran the testsuite, which disappeared once I
manually triggered a full rebuild -- is there some build dependency
missing somewhere...

> It doesn't fixup the failure 
> thought.  Still thinking about that.

Hmm, for me the failures now seem to be done; again, within the handful
of testsuite runs I've done.


Grüße,
 Thomas
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 472 bytes
Desc: not available
URL: <http://gcc.gnu.org/pipermail/gcc-patches/attachments/20150730/0399c593/attachment.sig>