On Wed, Jan 18, 2017 at 08:02:14PM +0300, Alexander Monakov wrote:
> On Wed, 18 Jan 2017, Jakub Jelinek wrote:
> > > It is, but I think my approach is compatible with inlining too (and has a
> > > more
> > > localized impact on the compiler).
> >
> > But your 2/5 patch disables inlining into the SIMT regions. Or do you mean
> > the approach with some new IFN for the pointers to privatized vars?
>
> I think it can be done either in the current framework, or in that more
> involved
> approach with a new IFN.
>
> > How would that work with the inliner (all copies of addressable vars/params
> > from functions inlined into the SIMT region would need to gain something
> > similar)?
>
> Inlining needs to do just like omp-low; if we take the current framework, it
> would need to collect addressable locals into one struct, replace references
> to
> those locals by field references in the inlined body. Then it needs to
> appropriately increase allocation size/alignment in SIMT_ENTER() call
> arguments.
> And finally it would need to initialize the pointer to structure, either
> immediately after SIMT_ENTER, or in a more fine-grained manner by a
> __builtin_alloca_with_align-like function (__b_a_w_a is not usable for that
> itself, because currently for known sizes gcc can make it a local variable).
One of the problems with that is that it means that you can't easily turn
addressable private variables into non-addressable ones once you force them
into such struct that can't be easily SRA split.
In contrast, if you can get the variable flags/attributes work, if they
become non-addressable (which is especially important to get rid of C++
abstraction penalties), you simply don't add them into the specially
allocated block.
Jakub
