[llvm-dev] Inlining + CSE + restrict pointers == funtimes

[Jeroen Dobbelaere via llvm-dev]

That's indeed a good example that shows that dropping the metadata (A) is the only correct solution.

Greetings,

Jeroen Dobbelaere


From: Finkel, Hal J. <hfinkel at anl.gov>
Sent: Wednesday, January 22, 2020 17:33
To: Jeroen Dobbelaere <dobbel at synopsys.com>; Neil Henning <neil.henning at unity3d.com>; llvm-dev at lists.llvm.org
Subject: Re: [llvm-dev] Inlining + CSE + restrict pointers == funtimes


At a high level, EarlyCSE should be intersecting the metadata of instructions that it combines. If it doesn't, and also doesn't drop the metadata, that seems like a bug, regardless of anything else.
On 1/22/20 9:30 AM, Jeroen Dobbelaere wrote:
[..]
-- (A) dropping the noalias information will result in less optimization opportunities
-- (B) preferring the load with the noalias annotation can allow more reorderings and a better schedule.
  (especially when there would be a 'store' in the context where the restrict is valid; aka in 'called').

My preference would go to (B) as that opens up more optimization opportunities.



I'm concerned that this isn't sound. So, imagine if I had something like this:
int * restrict r = a;
...
int x = noaliasing ? *r : *a;

then we need the aliasing load to retain its dependencies relative to other things in the block. If we combine them into one load, it needs to be the aliasing one.

Thanks again,

Hal

[...]
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20200122/33550ceb/attachment.html>