similar to: [LLVMdev] counting branch frequencies

I tried getting profile data from LLVM 3.1, using the method mentioned below. I tried it out on a simple matrix multiplication program. However, I noticed the following problems: 1. There is a warning message: "WARNING: profile information is inconsistent with the current program!" 2. The basic block counts (obtained from ProfileInfo::getExecutionCount(const BasicBlock*)) are

Hi Apala, Dibyendu is correct that this is likely due to pass order, but things get a bit complicated with -O[1-9] or -std-compile-opts as they insert early passes *before* the profiling code. I recommend that you use identical optimizations to insert instrumentation and to load the profiling data. E.g.: opt -insert-edge-profiling -O3 foo.bc -o foo.2.bc opt -profile-loader -O3 foo.bc

Hi Apala, On 11/09/12 11:20, apala guha wrote: > Is it possible to associate the branch frequency counts with the basic > blocks > in the intermediate representation? (e.g. Can I access basic block > frequencies in runOnFunction()?) Profile data really needs to be loaded at a module level, but once this has been done it can be accessed at any level (including function). In LLVM

Thanks everyone for the replies. After some experimentation, I found that the order in which the passes are specified matters: opt -O3 -profile-loader matmult.bc -o matmult.opt.bc (works) opt -profile-loader -O3 matmult.bc -o matmult.opt.bc (does not work) Also, I am able to avoid the inconsistency warning only for optimization levels -O3 and -O2. I get that warning when using -O1 and

Another issue is with ProfileInfo::getExecutionCount(Function* F). Looking at the source code and results, I am seeing that it always returns the execution count of the entry basic block of the function. If the entry basic block is part of a loop, its execution count does not match the function invocation count. Is my assumption wrong, that ProfileInfo::getExecutionCount(Function* F) is

Can we not run the -insert-edge-profiling and -profile-loader passes at the beginning of the opt? Orthogonal point is, is it worth doing any optimizations when -insert-edge-profiling is specified on command line? -Prashantha -----Original Message----- From: llvmdev-bounces at cs.uiuc.edu [mailto:llvmdev-bounces at cs.uiuc.edu] On Behalf Of Alastair Murray Sent: Wednesday, September 19, 2012

Hi, Is there a way to count branch frequencies using LLVM infrastructure? Thanks. -Apala Postdoctoral Scholar Department of Computer Science, University of Chicago Computation Institute, Argonne National Laboratory http://sites.google.com/site/apalaguha/home/ -------------- next part -------------- An HTML attachment was scrubbed... URL:

Hi all, In light of the expected removal of ProfileInfo this is a request for comments on the next few items that I now plan to work on for GSoC. Planned tasks: #0 Add support for determining branch weight metadata by profiling At the absolute minimum this will require writing a new profile loader which will set branch weight metadata based on profiling data. #1 Optionally use profiling

On Sun, Jul 15, 2012 at 8:32 AM, Alastair Murray <alastairmurray42 at gmail.com > wrote: > Hi Chandler, > > I'm a GSoC student working on profiling support (mentor CC'ed). I'm no > stranger to the issues with the current system: my original proposal was > written without knowledge of the limitations. This is why this list > hasn't heard much from me yet.

Hi Chandler, I'm a GSoC student working on profiling support (mentor CC'ed). I'm no stranger to the issues with the current system: my original proposal was written without knowledge of the limitations. This is why this list hasn't heard much from me yet. I would like to continue working on profiling support but I'm not attached to ProfileInfo and wouldn't be

Thanks for such a rapid response! > Don't know about Passes in the backend, but this could be a problem of > an FunctionPassManager trying to use a ModulePass. I manually applied the patch you provided for llc (I'm using the 2.5 release of LLVM not ToT) and it fixed my compilation error. When your patch replaced the FunctionPassManager used by llc with a PassManager the error went

Hi all, as proposed in http://lists.cs.uiuc.edu/pipermail/llvmdev/2009-February/020396.html I implemented the algorithm presented in [Ball94]. It only instruments the minimal number of edges necessary for edge profiling. The main changes introduced by this patch are: *) a interface compatible rewrite of ProfileInfo *) a cleanup of ProfileInfoLoader (some functionality in ProfileInfoLoader

Hi Andreas, First, thanks again for undertaking this work and submitting it back. There is a lot of good stuff here and it would be great to see it get back into the tree. I have a few major high-level comments on the patch. First off, the patch is quite large and should be broken down into separate incremental changes which are easier to review and apply. I think the patches should more or less

Hello folks, I'd like to remove all of the old and defunct profile info passes from LLVM. These have been almost entirely supplanted by the BranchProbability and BlockFrequency systems, which are actually on by default, and in use in optimization passes. The old system is not on, and hasn't been touched in years except to do minor build fixes and updates. As far as I'm aware, the

Hi Alastair, In addition to your planned tasks, you might want to put in some work to ensure branch probabilities are not lost during optimization. One known issue is LLVM optimizer can turn branchy code into switch statements and it would completely discard probability. Here is a simple example: static void func2(int N, const int *a, const int *b, int *c) __attribute__((always_inline)); void

Shuguang Feng wrote: >> What does "llc -debug-pass=Structure" say? Is the ProfileLoaderPass >> really the last pass to touch the ProfileInfo before you are using it? > > Below is the sequence of passes that I see. Although the > NoProfileInfo pass is being run, it should be subsequently overridden > by ProfileInfoLoaderPass (LoaderPass) correct? Yes. >

Hi Daniel, Daniel Dunbar wrote: > Hi Andreas, > > First, thanks again for undertaking this work and submitting it back. There is a > lot of good stuff here and it would be great to see it get back into the tree. Thanks for taking the time to review this, I know its a huge patch. I still have a few questions on how you would like this patch to be re-factored and split up. > [...]

Hi Alastair, Your patch looks great. I am new on this part of llvm. Why do we want to get rid of ProfileInfo? Is there any email thread related to this? I understand that we want to setBranchWeightMetadata from the profiling data generated by -insert-edge-profiling. I thought the existing loader can parse the profiling data and map the data to the IR, and only the setting of metadata is missing.

Hi, How can I use profile information into my pass for performing some analysis. I tried something like: PassManager PassMgr = PassManager(); cl::opt<std::string> ProfileDataFile(cl::Positional, cl::desc("<llvmprof.out file>"), cl::Optional, cl::init("llvmprof.out")); PassMgr.add(createProfileLoaderPass(ProfileDataFile));

The PTX back-end is robust (it's based on the sources used by nvcc), but I'm not sure about the OpenMP representation in LLVM IR. I believe the OpenMP constructs are already lowered into libgomp calls before leaving DragonEgg. It's been awhile since I've loooked at it though. If you use the PTX back-end and have any issues, please don't hesitate to post to the list and cc: