[llvm] r214630 - [x86] Remove the FIXME that was implemented in r214628. Managed to
Chandler Carruth
chandlerc at gmail.com
Sat Aug 2 04:34:24 PDT 2014
Author: chandlerc
Date: Sat Aug 2 06:34:23 2014
New Revision: 214630
URL: http://llvm.org/viewvc/llvm-project?rev=214630&view=rev
Log:
[x86] Remove the FIXME that was implemented in r214628. Managed to
forget to update the comment here... =/
Modified:
llvm/trunk/lib/Target/X86/X86ISelLowering.cpp
Modified: llvm/trunk/lib/Target/X86/X86ISelLowering.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/X86/X86ISelLowering.cpp?rev=214630&r1=214629&r2=214630&view=diff
==============================================================================
--- llvm/trunk/lib/Target/X86/X86ISelLowering.cpp (original)
+++ llvm/trunk/lib/Target/X86/X86ISelLowering.cpp Sat Aug 2 06:34:23 2014
@@ -18945,10 +18945,6 @@ static bool combineX86ShuffleChain(SDVal
/// This should never be an issue in practice as the shuffle lowering doesn't
/// produce sequences of more than 8 instructions.
///
-/// FIXME: Currently, we don't collapse instructions *into* PSHUFB. We should,
-/// and we should do so more aggressively than we form PSHUFB because once we
-/// have a PSHUFB, we might as well do as much shuffling as we can.
-///
/// FIXME: We will currently miss some cases where the redundant shuffling
/// would simplify under the threshold for PSHUFB formation because of
/// combine-ordering. To fix this, we should do the redundant instruction
More information about the llvm-commits
mailing list