slow forall

classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|

slow forall

Steve Kargl
All,

PR fortran/42118 concerns slow performance of FORALL.

Fortran 2018 has declared FORALL to be an obsolescent feature.
I doubt that anyone on this mailing list will ever improve
the performance of FORALL.

Does anyone care if I close the PR as WONTFIX?  If someone
answers "Yes, I care", be prepared to present your patch.

--
Steve
Reply | Threaded
Open this post in threaded view
|

Re: slow forall

Thomas Koenig-6
Steve,


> PR fortran/42118 concerns slow performance of FORALL.

> Does anyone care if I close the PR as WONTFIX?  If someone
> answers "Yes, I care", be prepared to present your patch.

Comment #6 is already fixed with the loop interchange.

The original test case depends on extending alias analysis in
a quite complex manner - I'm not about to do this.

So I have no objection to you closing this PR.

Regards

        Thomas
Reply | Threaded
Open this post in threaded view
|

Re: slow forall

Steve Kargl
On Sat, Oct 05, 2019 at 12:54:13PM +0200, Thomas Koenig wrote:

> Steve,
>
> > PR fortran/42118 concerns slow performance of FORALL.
>
> > Does anyone care if I close the PR as WONTFIX?  If someone
> > answers "Yes, I care", be prepared to present your patch.
>
> Comment #6 is already fixed with the loop interchange.
>
> The original test case depends on extending alias analysis in
> a quite complex manner - I'm not about to do this.
>
> So I have no objection to you closing this PR.
>

Thanks for taking a look at the PR.  The original test case
was submitted 10 years ago.  I assume, that after 10 years,
no one on this list cares about the performance of FORALL.
With FORALL moved to the obsolescent feature list, it will
get pushed even farther down the list of bugs.  It is a
simple matter of too many bugs, too little time, and
too few active patch contributors.  FORALL works (short
of any actual bugs).

--
Steve