llvm/Transforms at cbf676b3ba907f72405a02938f5cd85fb3e6a46c - llvm - Gitea: Git with a cup of tea

RPCS3/llvm

mirror of https://github.com/RPCS3/llvm.git synced 2025-03-05 00:59:19 +00:00

History

Jim Grosbach cbf676b3ba float comparison to double 'zero' constant can just be a float 'zero.'

InstCombine was incorrectly considering the conversion of the constant
zero to be unsafe.

We want to transform:
define float @bar(float %x) nounwind readnone optsize ssp {
  %conv = fpext float %x to double
  %cmp = fcmp olt double %conv, 0.000000e+00
  %conv1 = zext i1 %cmp to i32
  %conv2 = sitofp i32 %conv1 to float
  ret float %conv2
}

Into:
define float @bar(float %x) nounwind readnone optsize ssp {
  %cmp = fcmp olt float %x, 0.000000e+00   ; <---- This
  %conv1 = zext i1 %cmp to i32
  %conv2 = sitofp i32 %conv1 to float
  ret float %conv2
}


rdar://10215914


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@140869 91177308-0d34-0410-b5e6-96231b3b80d8

2011-09-30 18:45:50 +00:00

..

Try to eliminate the use of the 'unwind' instruction.

2011-09-02 22:41:11 +00:00

ArgumentPromotion

fix PR8932, a case where arg promotion could infinitely promote.

2011-01-16 08:09:24 +00:00

Change tests from "opt %s" to "opt < %s" so that opt doesn't see the

2009-09-11 18:01:28 +00:00

Change tests from "opt %s" to "opt < %s" so that opt doesn't see the

2009-09-11 18:01:28 +00:00

Update this test to the new EH model.

2011-09-20 22:29:43 +00:00

Roll r127459 back in:

2011-03-11 21:52:04 +00:00

Teach the ConstantMerge pass about alignment. Fixes PR10514!

2011-07-27 19:47:34 +00:00

Land the long talked about "type system rewrite" patch. This

2011-07-09 17:41:24 +00:00

CorrelatedValuePropagation

Test for lazy value info's ability to prove the absense of NULLs in pointers.

2011-01-16 21:57:20 +00:00

Update some tests to the new EH scheme.

2011-09-01 00:58:03 +00:00

DeadStoreElimination

Try again at r138809 (make DSE more aggressive in removing dead stores at the end of a function), now with less deleting stores before memcpy's.

2011-09-06 18:14:09 +00:00

Move "atomic" and "volatile" designations on instructions after the opcode

2011-08-12 22:50:01 +00:00

Update some tests to the new EH scheme.

2011-09-01 00:58:03 +00:00

Change tests from "opt %s" to "opt < %s" so that opt doesn't see the

2009-09-11 18:01:28 +00:00

Move "atomic" and "volatile" designations on instructions after the opcode

2011-08-12 22:50:01 +00:00

Mark the eh.typeid.for intrinsic as being 'const', which it is inside

2011-09-09 07:50:37 +00:00

Test case for r140670: indvars should hoist sext.

2011-09-28 02:13:32 +00:00

Replace more uses of 'unwind' in the tests with calls to landingpad and

2011-09-19 22:11:35 +00:00

float comparison to double 'zero' constant can just be a float 'zero.'

2011-09-30 18:45:50 +00:00

InstSimplify: Don't try to replace an extractvalue/insertvalue pair with the original value if types don't match.

2011-09-05 18:16:19 +00:00

Add test for r127138.

2011-03-07 21:28:14 +00:00

Update some tests to the new EH scheme.

2011-09-01 00:58:03 +00:00

Move "atomic" and "volatile" designations on instructions after the opcode

2011-08-12 22:50:01 +00:00

Try to eliminate the use of the 'unwind' instruction.

2011-09-02 22:41:11 +00:00

Atomic load/store support in LICM.

2011-08-15 20:52:09 +00:00

Remove the old tail duplication pass. It is not used and is unable to update

2011-08-30 23:03:45 +00:00

Stop emitting instructions with the name "tmp" they eat up memory and have to be uniqued, without any benefit.

2011-09-27 20:39:19 +00:00

remove asmparser support for the old getresult instruction, which has been subsumed by extractvalue.

2011-06-17 06:57:15 +00:00

Update to the new EH syntax.

2011-09-19 23:50:34 +00:00

LoopStrengthReduce

Stop emitting instructions with the name "tmp" they eat up memory and have to be uniqued, without any benefit.

2011-09-27 20:39:19 +00:00

Reapply r139759. Disable IV rewriting by default. See PR10916.

2011-09-15 20:58:37 +00:00

Split the landing pad basic block with the correct function. Also merge the

2011-09-27 00:59:31 +00:00

Convert more tests over to the new atomic instructions.

2011-09-26 20:27:49 +00:00

LowerExpectIntrinsic

Introduce "expect" intrinsic instructions.

2011-07-06 18:22:43 +00:00

Change tests from "opt %s" to "opt < %s" so that opt doesn't see the

2009-09-11 18:01:28 +00:00

Change tests from "opt %s" to "opt < %s" so that opt doesn't see the

2009-09-11 18:01:28 +00:00

Update more tests to the new EH scheme.

2011-08-31 21:39:05 +00:00

Atomic load/store handling for the passes using memdep (GVN, DSE, memcpyopt).

2011-08-17 22:22:24 +00:00

When removing a function from the function set and adding it to deferred, we

2011-02-09 06:32:02 +00:00

When eliminating unnecessary retain+autorelease on return values,

2011-09-29 22:27:34 +00:00

Add file.

2011-08-19 21:21:20 +00:00

Update more tests to the new EH scheme.

2011-08-31 21:04:11 +00:00

Remove empty test.

2011-08-02 02:47:45 +00:00

Stop emitting instructions with the name "tmp" they eat up memory and have to be uniqued, without any benefit.

2011-09-27 20:39:19 +00:00

Make sure IPSCCP never marks a tracked call as overdefined in SCCPSolver::ResolvedUndefsIn. If we do, we can end up in a situation where a function is resolved to return a constant, but the caller is marked overdefined, which confuses the code later.

2011-09-20 23:28:51 +00:00

Replace more uses of 'unwind' in the tests with calls to landingpad and

2011-09-19 22:11:35 +00:00

SimplifyLibCalls

Stop emitting instructions with the name "tmp" they eat up memory and have to be uniqued, without any benefit.

2011-09-27 20:39:19 +00:00

Fix test; sorry for any inconvenience.

2011-09-01 21:25:42 +00:00

fix PR10286, a problem with the .ll printer handling block addresses that are out-of-scope.

2011-08-03 06:15:41 +00:00

Replace uses of unwind with unreachable for the same effect.

2011-09-19 22:16:15 +00:00

Remove the old tail duplication pass. It is not used and is unable to update

2011-08-30 23:03:45 +00:00