Philip Reames via llvm-dev
2021-May-25 15:41 UTC
[llvm-dev] llvm.assume after CodeGenPrepare
I think there's two interacting pieces here:
* I don't believe we have a way to represent an assume at the MI
layer. If we let them flow through codegen, we'd need to add such a
thing.
* The tradeoffs between information preservation and lowering may be
different at different points in the pipeline. Johannes frames this
as a canonicalization problem. That sounds reasonable, but it's
also reasonable that we may need to give up on preserving assumes at
some point. (As a silly example, we probably don't want them at
MC.) Where exactly that point is unclear, and is mostly a matter of
practical engineering tradeoffs.
If you felt like exploring alternate lowering points, that would seem
entirely reasonable. It might not work out, or it might require a bunch
of work to make happen, but the basic idea seems entirely worth exploring.
Philip
On 5/25/21 3:11 AM, Markus Lavin via llvm-dev wrote:>
> With recent changes in BasicAA (mostly by Nikita Popov I believe)
> llvm.assumes can now guide in the AA decision making. Which is of
> course great.
>
> For example for C input (or IR equivalent) as follows it can make a
> huge difference if the variable ‘x’ is known to be non-zero when AA is
> queried during scheduling
>
> __builtin_assume(x != 0);
>
> for (int i = 0; i < 64; i += 4) {
>
> v[(i + 0) * x] = v[(i + 0) * x] >> 2;
>
> v[(i + 1) * x] = v[(i + 1) * x] >> 2;
>
> v[(i + 2) * x] = v[(i + 2) * x] >> 2;
>
> v[(i + 3) * x] = v[(i + 3) * x] >> 2;
>
> }
>
> Unfortunately it appears that the CodeGenPrepare pass removes
> llvm.assume so that they never reach the code generator. Currently
> commit 91c9dee3fb6d89ab3 (and before that commit 6d20937c29a1a1d67)
> eliminate assumptions in CodeGenPrepare for reasons that appear to be
> optimization (avoiding blocks that would be empty if it was not for
> the llvm.assume and its predecessors).
>
> It seems these two efforts are quite contradictory. Is there any
> deeper thinking behind this? I for one would be in favor of not
> eliminating assumes.
>
> -Markus
>
>
> _______________________________________________
> LLVM Developers mailing list
> llvm-dev at lists.llvm.org
> https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20210525/dc241bb1/attachment.html>
Björn Pettersson A via llvm-dev
2021-May-25 18:39 UTC
[llvm-dev] llvm.assume after CodeGenPrepare
I think Markus use case is that MachineInstr::mayAlias may use AA (and even
possibly TBAA), and AA could make use of llvm.assume to make decisions.
So the alias analysis in cadegen is partly based on analysing the
"final" IR representation before ISel, so the llvm.assume does not
need to be present in the MI layer for that to work.
PS. It might seem a bit strange and dangerous to use the IR and IR analyses
during the codegen (e.g. afte PHI elimination), but afaik it works as long as
MachineMemOperands are handled correctly (being dropped or updated correctly
when doing certain transforms such as store merging).
/Björn
From: llvm-dev <llvm-dev-bounces at lists.llvm.org> On Behalf Of Philip
Reames via llvm-dev
Sent: den 25 maj 2021 17:42
To: Markus Lavin <markus.lavin at ericsson.com>; llvm-dev <llvm-dev at
lists.llvm.org>; jannh at google.com; nikita.ppv at gmail.com
Subject: Re: [llvm-dev] llvm.assume after CodeGenPrepare
I think there's two interacting pieces here:
* I don't believe we have a way to represent an assume at the MI layer.
If we let them flow through codegen, we'd need to add such a thing.
* The tradeoffs between information preservation and lowering may be
different at different points in the pipeline. Johannes frames this as a
canonicalization problem. That sounds reasonable, but it's also reasonable
that we may need to give up on preserving assumes at some point. (As a silly
example, we probably don't want them at MC.) Where exactly that point is
unclear, and is mostly a matter of practical engineering tradeoffs.
If you felt like exploring alternate lowering points, that would seem entirely
reasonable. It might not work out, or it might require a bunch of work to make
happen, but the basic idea seems entirely worth exploring.
Philip
On 5/25/21 3:11 AM, Markus Lavin via llvm-dev wrote:
With recent changes in BasicAA (mostly by Nikita Popov I believe) llvm.assumes
can now guide in the AA decision making. Which is of course great.
For example for C input (or IR equivalent) as follows it can make a huge
difference if the variable 'x' is known to be non-zero when AA is
queried during scheduling
__builtin_assume(x != 0);
for (int i = 0; i < 64; i += 4) {
v[(i + 0) * x] = v[(i + 0) * x] >> 2;
v[(i + 1) * x] = v[(i + 1) * x] >> 2;
v[(i + 2) * x] = v[(i + 2) * x] >> 2;
v[(i + 3) * x] = v[(i + 3) * x] >> 2;
}
Unfortunately it appears that the CodeGenPrepare pass removes llvm.assume so
that they never reach the code generator. Currently commit 91c9dee3fb6d89ab3
(and before that commit 6d20937c29a1a1d67) eliminate assumptions in
CodeGenPrepare for reasons that appear to be optimization (avoiding blocks that
would be empty if it was not for the llvm.assume and its predecessors).
It seems these two efforts are quite contradictory. Is there any deeper thinking
behind this? I for one would be in favor of not eliminating assumes.
-Markus
_______________________________________________
LLVM Developers mailing list
llvm-dev at lists.llvm.org<mailto:llvm-dev at lists.llvm.org>
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev<https://protect2.fireeye.com/v1/url?k=efdd418c-b0467bc9-efdd0117-867b36d1634c-7a75399069504f9e&q=1&e=17991efb-205e-4418-bc44-152a2fed67d4&u=https%3A%2F%2Flists.llvm.org%2Fcgi-bin%2Fmailman%2Flistinfo%2Fllvm-dev>
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20210525/2838ed52/attachment.html>