Addressed comments.

Addressed comments.

In D86262#2227676, @lebedev.ri wrote:

Why is this the correct course of action?
For example, if the IR already had memcpy/memset, the DA will also be pessimized,
so it really seems like a workaround, not a fix.
Why not just enhance DA?

Please clang-format it.

Check for hasMadMacF32Insts() added.

Change updated with addressed comments.

Hi @cebowleratibm, it seems test is located in wrong place: test/CodeGen/PowerPC instead of llvm/test/CodeGen/PowerPC.

removed extra new lines

Ping.

Test updated

I was completely wrong

In D82650#2116849, @yaxunl wrote:

I think the point of -ffp-contract=fast is that it allows emitting fma across statements if there are such opportunities whereas -ffp-contract=on only allows fma in one statement

https://llvm.org/docs/CompileCudaWithLLVM.html#flags-that-control-numerical-code

Yes, you're right. I was confused by current FE behavior and a number of different commits, reverts and re-commits.

ping...

Added comment, fixed typo

updated comments

Actually it seems all of these dozens of "CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i32" may be removed from most of tests not related to getCFInstrCost

LGTM, although minor change in test' comment

I've just got the bug, will take a look

fixed spelling

There are the tests in llvm/test/CodeGen/AMDGPU/simplify-libcalls.ll

diff reuploaded

added test

test added

test updated as requested

In D81260#2077371, @rampitec wrote:

But how did it work?! It should have skipped any call which is not a lifetime intrinsic.

Phabricator somehow shows wrong diff after patch update. Correct diff is

-       if (!CI) continue;
+       if (!CI || isa<DbgInfoIntrinsic>(CI) || CI->isLifetimeStartOrEnd())
+         continue;

AMDGPU part LGTM,
let's wait a couple of days to allow others to take a look

fixed typo

fixing lint message