similar to: [LLVMdev] [PATCH] Before/After IR Dumps

On Thursday 04 March 2010 01:54:55 Chris Lattner wrote: > On Mar 2, 2010, at 1:45 PM, David Greene wrote: > > This set of patches adds support for dumping IR before or after specified > > Passes. It adds the following command-line options: > > > > -print-before=<pass-option> > > -print-after=<pass-option> > > -print-before-all > >

On Mar 2, 2010, at 1:45 PM, David Greene wrote: > This set of patches adds support for dumping IR before or after specified > Passes. It adds the following command-line options: > > -print-before=<pass-option> > -print-after=<pass-option> > -print-before-all > -print-after-all This patch looks very invasive for such a simple thing, isn't there a better

On Mar 4, 2010, at 7:28 AM, David Greene wrote: > On Thursday 04 March 2010 01:54:55 Chris Lattner wrote: >> On Mar 2, 2010, at 1:45 PM, David Greene wrote: >>> This set of patches adds support for dumping IR before or after specified >>> Passes. It adds the following command-line options: >>> >>> -print-before=<pass-option> >>>

Hello, does anyone knows this predict is not resulting? # lm predict dfTestes3sitesCriptic$Velocity_corrected <- ifelse (dfTestes3sitesCriptic$Season == "A" & dfTestes3sitesCriptic$BeforeAfter == "Before", (dfTestes3sitesCriptic$Velocity * mVel3ABefAfter), (ifelse (dfTestes3sitesCriptic$Season == "Sp" & dfTestes3sitesCriptic$BeforeAfter ==

On Thursday 04 March 2010 09:28:22 David Greene wrote: > I'm happy to revisit design decisions if you let me know what you want to > see changed. Here's a rework using PassManager as Chris suggested. Comments? -Dave -------------- next part -------------- A non-text attachment was scrubbed... Name: clang-beforeafter.patch Type:

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 Monday 29 March 2010 12:08:45 Chris Lattner wrote: > >> Your patch doesn't include the file that defines > >> createMachineFunctionPrinterPass, but I assume it's fine. > > > > Odd. This was a patch generated from a clean upstream. It built fine. > > I'll see if maybe I missed something. > > I think it's just missing from the patch

On Mar 29, 2010, at 11:14 AM, David Greene wrote: >>>> >>>> After making these changes, please test the resultant patch against >>>> mainline with just the patch applied and commit if it looks ok. Thanks >>>> David, >>> >>> Will do. >> >> Thanks! > > How should we handle the llvm-gcc and clang changes? What

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

On Sunday 14 March 2010 18:32:35 Chris Lattner wrote: > This is much better than the first iteration but still has many issues. > There is no need to keep cc'ing cfe-dev on these patches which have nothing > to do with clang. There's a clang patch in this set. > Please rename the getPrinterPass method to createPrinterPass to indicate > that the result returns a new pass,

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

On Oct 16, 2008, at 8:29 AM, Edward Lee wrote: > Is there a simple way to require a pass, e.g., Reg2Mem/Mem2Reg, to > run before/after my transformation pass? Or do I do something like: One simplest way is to handle this is to add these passes around your pass in the pass manager. pm.add(Reg2Mem) pm.add(MyPass) pm.add(Mem2Reg) > > > struct myOpt { > myOpt() { >

It's not being executed. I'm both trying to print via outs() and add info via STATISTIC, neither is occurring. On Mon, Jul 14, 2014 at 12:22 PM, Justin Holewinski <jholewinski at nvidia.com> wrote: > On Mon, 2014-07-14 at 08:31 -0700, Ryan Taylor wrote: > > Where is the documentation about registering a machine pass? I'm > > unable to find it. > > >

On Mar 12, 2010, at 8:10 AM, David Greene wrote: > 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: >>

Is there a simple way to require a pass, e.g., Reg2Mem/Mem2Reg, to run before/after my transformation pass? Or do I do something like: struct myOpt { myOpt() { mBefore = createDemoteRegisterToMemoryPass(); mAfter = createPromoteMemoryToRegisterPass(); } getAnalysisUsage(AU) { AU.addRequired(my stuff); mBefore.getAnalysisUsage(AU); mAfter.getAnalysisUsage(AU); }

Eli Friedman <eli.friedman at gmail.com> writes: >> Ah. So is the MachineFunction version expected to work correctly? > > It's part of the default set of CodeGen passes. It is? Was that true in 3.1? I can't see where it is initialized in llc. I probably missed something important. :) Thanks! -David

Hi Danny, // Add TypeBasedAliasAnalysis before BasicAliasAnalysis so that // BasicAliasAnalysis wins if they disagree. This is intended to help // support "obvious" type-punning idioms. - if (UseCFLAA) - addPass(createCFLAliasAnalysisPass()); addPass(createTypeBasedAliasAnalysisPass()); addPass(createScopedNoAliasAAPass()); + if (UseCFLAA) +

So, I can make all these testcases work, but it's a little tricky (it involves tracking some things, like GEP byte range, and then checking bases and using getObjectSize, much like BasicAA does). Because i really don't want to put that much "not well tested" code in a bugfix, and honestly, i'm not sure we will catch any cases here that BasicAA does not, i've attached a

Where is the documentation about registering a machine pass? I'm unable to find it. I have built a machine function pass similar to the one found in HexagonHardwareLoops.cpp. So I have generated a machine pass.cpp file, modified 'Target'.h and 'Target'TargetMachine.cpp (to add pass via addPass() in the addPreRegAlloc()). All this builds/compiles fine. When running llc the