[hsa merge 07/10] IPA-HSA pass

[Alexander Monakov]

On Fri, 15 Jan 2016, Jakub Jelinek wrote:
> On Fri, Jan 15, 2016 at 07:38:14PM +0300, Ilya Verbin wrote:
> > On Fri, Jan 15, 2016 at 17:09:54 +0100, Jakub Jelinek wrote:
> > > On Fri, Jan 15, 2016 at 05:02:34PM +0100, Martin Jambor wrote:
> > > > How do other accelerators cope with the situation when half of the
> > > > application is compiled with the accelerator disabled?  (Would some of
> > > > their calls to GOMP_target_ext lead to abort?)
> > > 
> > > GOMP_target_ext should never abort (unless internal error), worst case it
> > > just falls back into the host fallback.

Agreed -- the way it aborts today rather than using host fallback looks rather
surprising to me.

> > Wouldn't that lead to hard-to-find problems in case of nonshared memory?
> > I mean when someone expects that all target regions are executed on the device,
> > but in fact some of them are silently executed on the host with different data
> > environment.
> 
> E.g. for HSA it really shouldn't matter, as it is shared memory accelerator.
> For XeonPhi we hopefully can offload anything.  NVPTX is problematic,
> because it can't offload all the code, 

Sorry, can you clarify -- what do you mean by "can't offload"?

> but if it can be e.g. compile time detected that it will not be possible, it
> can just provide offloaded code for the target.

(as a result of previous confusion I can't follow this part either)

Thanks.
Alexander