llvm dev - Feb 2020 - RFC: Mark BasicAA as a CFG-only pass.

Hi,

I'd like to understand if it makes sense to keep BasicAA as a not CFG-only
pass, or if it can be updated to CFG-only. The change was made in D44564
<https://reviews.llvm.org/D44564>.
>From what I gathered the motivation was PhiValuesAnalysis not being
properly updated, and BasicAA having an instance of it.
PhiValuesAnalysis now uses callback values to invalidate deleted values (
r340613 <https://reviews.llvm.org/rL340613>), PhiValuesAnalysis is also
being updated in MemDepAnalysis (D48489 <https://reviews.llvm.org/D48489>)
and
BasicAA is invalidated if PhiValuesAnalysis gets invalidated.

I may not have the full context here, so I'd like some feedback: does it
make sense to make BasicAA a CFG-only pass again?

Thank you,
Alina
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20200210/1f57ad9e/attachment.html>

Hi,

Here's a tentative patch of the changes for this: D74353
<https://reviews.llvm.org/D74353>.

Thank you,
Alina


On Mon, Feb 10, 2020 at 11:34 AM Alina Sbirlea <alina.sbirlea at
gmail.com>
wrote:
> Hi,
>
> I'd like to understand if it makes sense to keep BasicAA as a not
CFG-only
> pass, or if it can be updated to CFG-only. The change was made in D44564
> <https://reviews.llvm.org/D44564>.
>
> From what I gathered the motivation was PhiValuesAnalysis not being
> properly updated, and BasicAA having an instance of it.
> PhiValuesAnalysis now uses callback values to invalidate deleted values (
> r340613 <https://reviews.llvm.org/rL340613>), PhiValuesAnalysis is
also
> being updated in MemDepAnalysis (D48489
<https://reviews.llvm.org/D48489>
> ) and BasicAA is invalidated if PhiValuesAnalysis gets invalidated.
>
> I may not have the full context here, so I'd like some feedback: does
it
> make sense to make BasicAA a CFG-only pass again?
>
> Thank you,
> Alina
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20200210/7c01ced4/attachment.html>

On 2/10/20 2:35 PM, Alina Sbirlea wrote:
> Hi,
>
> Here's a tentative patch of the changes for this: D74353 
> <https://reviews.llvm.org/D74353>.

I suppose that, as expected, it's invalidated less often this way. Given 
that it's generally stateless, does this really represent a cost savings?

  -Hal

>
> Thank you,
> Alina
>
>
> On Mon, Feb 10, 2020 at 11:34 AM Alina Sbirlea 
> <alina.sbirlea at gmail.com <mailto:alina.sbirlea at
gmail.com>> wrote:
>
>     Hi,
>
>     I'd like to understand if it makes sense to keep BasicAA as a not
>     CFG-only pass, or if it can be updated to CFG-only. The change was
>     made in D44564 <https://reviews.llvm.org/D44564>.
>
>     From what I gathered the motivation was PhiValuesAnalysis not
>     being properly updated, and BasicAA having an instance of it.
>     PhiValuesAnalysis now uses callback values to invalidate deleted
>     values (r340613
>     <https://reviews.llvm.org/rL340613>),PhiValuesAnalysis is also
>     being updated in MemDepAnalysis (D48489
>     <https://reviews.llvm.org/D48489>) and BasicAA is invalidated if
>     PhiValuesAnalysis gets invalidated.
>
>     I may not have the full context here, so I'd like some feedback:
>     does it make sense to make BasicAA a CFG-only pass again?
>
>     Thank you,
>     Alina
>
-- 
Hal Finkel
Lead, Compiler Technology and Programming Languages
Leadership Computing Facility
Argonne National Laboratory

-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20200210/c01de515/attachment.html>