search for: createprintfunctionpass

...he callgraphnodes and call F->print(Out) on the functions for each node. If you'd prefer, you can just emit "not implemented for CGSCC nodes yet" if you're not interested in implementing the logic. + bool runOnLoop(Loop *L, LPPassManager &) { + FunctionPass *Printer = createPrintFunctionPass(Banner, &Out); + Printer->runOnFunction(*L->getHeader()->getParent()); + delete Printer; + return false; + } This can just call print on the function, no need to create the pass. It would be even better to just print the blocks in the loop or print "not implemented yet...

On Monday 15 March 2010 13:45:14 David Greene wrote: > On Sunday 14 March 2010 18:32:35 Chris Lattner wrote: > > This is much better than the first iteration but still has many issues. I believe I've addressed all your points with this patch except I didn't use StringRef. It doesn't appear to be useful since createPrinterPass will be sent a const std::string & and will

...ust emit "not implemented for CGSCC nodes yet" if you're not interested in > implementing the logic. Ok, I'll try to implement the logic. <Goes off to learn about CallGraph... :)> > + bool runOnLoop(Loop *L, LPPassManager &) { > + FunctionPass *Printer = createPrintFunctionPass(Banner, &Out); > + Printer->runOnFunction(*L->getHeader()->getParent()); > + delete Printer; > + return false; > + } > > This can just call print on the function, no need to create the pass. It > would be even better to just print the blocks in the loop...

On Monday 29 March 2010 13:20:37 Chris Lattner wrote: > > How should we handle the llvm-gcc and clang changes? > > What changes are required? Attached. They are trivial. > You have access to all llvm repositories, including clang and llvm-gcc. Ok. Is there are preferred order to commit things? I think I'm ready to go.

On Mar 30, 2010, at 12:27 PM, David Greene wrote: > On Monday 29 March 2010 13:20:37 Chris Lattner wrote: > >>> How should we handle the llvm-gcc and clang changes? >> >> What changes are required? > > Attached. They are trivial. Why not just change the order of the arguments to createPrintModulePass, and make the string default to ""? -Chris

...Adding a Module pass will change the ordering of passes in the passmanager, which will cause your option to completely change behavior of the optimization sequence. +Pass *LoopPass::getPrinterPass(raw_ostream &O, + const std::string &Banner) const { + return createPrintFunctionPass(Banner, &O); +} likewise. The printer needs to be a loop pass. +++ lib/CodeGen/MachineFunctionPrinterPass.cpp (revision 0) @@ -0,0 +1,51 @@ +//===-- MachineFunction.cpp -----------------------------------------------===// Please update the comment. +#include "llvm/CodeGen/MachineFunc...

On Friday 12 March 2010 08:13:05 Kalle Raiskila wrote: > David Greene wrote: > > Here's a rework using PassManager as Chris suggested. Comments? > > Tried this second patch with the svn version 97812 (the one the patch is > made against), but it doesn't compile: > "llvm/include/llvm/Pass.h:127: Error: expected unqualified-id before "&" >

David Greene wrote: > Here's a rework using PassManager as Chris suggested. Comments? Tried this second patch with the svn version 97812 (the one the patch is made against), but it doesn't compile: "llvm/include/llvm/Pass.h:127: Error: expected unqualified-id before "&" token" Seems raw_ostream is forward declared but not defined (adding a missing #include

...dn't want to mess with PassManager. All of this subtle stuff gets really tricky. So now I need another printer class? This is getting messy... > +Pass *LoopPass::getPrinterPass(raw_ostream &O, > + const std::string &Banner) const { > + return createPrintFunctionPass(Banner, &O); > +} > > likewise. The printer needs to be a loop pass. And another one... These extra printer passes are troublesome. I have to define them even though they are not used anywhere. After a LoopPass runs, wouldn't one want to look at the whole Function, not just th...

Hi Josh, This line works for me: opt file.ll -loop-vectorize -S -o - -mtriple=x86_64 -mcpu=corei7-avx -debug You need to specify the triple on the command line if it is not inside the module. Thanks, Nadav On May 9, 2013, at 5:53 PM, Joshua Klontz <josh.klontz at gmail.com> wrote: > Nadav, > > Please forgive my ignorance, but 'opt -mcpu=corei7 -loop-vectorize -S

Nadav, Please forgive my ignorance, but 'opt -mcpu=corei7 -loop-vectorize -S -debug double.ll' doesn't appear to make a difference. In fact it seems to be ignored as garbage values for -mcpu don't raise an error. Am I overlooking something else also? Many Thanks, Josh On Thu, May 9, 2013 at 6:06 PM, Nadav Rotem <nrotem at apple.com> wrote: > Hi Josh, > > Your

.../opt/Debug+Asserts/opt.o llvm::Pass::getAsPMDataManager() /n/fs/scratch/tpondich/ParallelAssert/llvm-objects/tools/opt/Debug+Asserts/AnalysisWrappers.o llvm::initializeFindUsedTypesPass(llvm::PassRegistry&) /n/fs/scratch/tpondich/ParallelAssert/llvm-objects/tools/opt/Debug+Asserts/opt.o llvm::createPrintFunctionPass(std::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, llvm::raw_ostream*, bool)/n/fs/scratch/tpondich/ParallelAssert/llvm-objects/tools/opt/Debug+Asserts/opt.o llvm::createIndVarSimplifyPass() /n/fs/scratch/tpondich/ParallelAssert/llvm-objects/tools/op...