Displaying 4 results from an estimated 4 matches for "processbasicblock".
2010 Jul 13
0
[LLVMdev] Bitcode from build-self-4-mingw32
...at 03:22
Subject: [LLVMdev] Bitcode from build-self-4-mingw32
To: LLVM Developers Mailing List <llvmdev at cs.uiuc.edu>
The buildbot build-self-4-mingw32 is failing with an assertion:
Assertion failed: (isStackEmpty() && "Stack not empty at end of basic
block?"), function processBasicBlock, file
/Users/buildslave/zorg/buildbot/smooshlab/slave/build-self-4-mingw32/llvm.src/lib/Target/X86/X86FloatingPoint.cpp,
line 305.
Could anyone produce a bitcode file for that failure,
please?---------- Forwarded message ----------
From: Jakob Stoklund Olesen <stoklund at 2pi.dk>
Date: Sun,...
2010 Jul 15
1
[LLVMdev] Splitting floating point intervals.
...ies. I have a
situation where a floating point value is split over a loop boundary, and
the register allocator choses to spill the value outside the loop. That
leaves me with a floating point load in the loop preheader, which is
crashing LLC with
X86FloatingPoint.cpp:305:
bool<unnamed>::FPS::processBasicBlock(llvm::MachineFunction&,
llvm::MachineBasicBlock&): Assertion `isStackEmpty() && "Stack not empty at
end of basic block?"'
What's going on here? Where do floating point values which are live over BB
boundaries live, if not on the FP stack?
Cheers,
Lang.
----------...
2010 Jul 10
0
[LLVMdev] Bitcode from build-self-4-mingw32
The buildbot build-self-4-mingw32 is failing with an assertion:
Assertion failed: (isStackEmpty() && "Stack not empty at end of basic block?"), function processBasicBlock, file /Users/buildslave/zorg/buildbot/smooshlab/slave/build-self-4-mingw32/llvm.src/lib/Target/X86/X86FloatingPoint.cpp, line 305.
Could anyone produce a bitcode file for that failure, please?
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: app...
2012 Nov 07
1
[LLVMdev] AVX support
We have been using LLVM 3.1 to support JITing of AVX. From dumping the MC generating by the MCJIT I noticed it always emits 'VZEROUPPER' to clear the high 128 bit before calling another function. In some cases I know the function called either only use AVX or does not use SSE. I will like to inform the backend it is safe not to emit that instruction.
Have not been able to figure out how