similar to: [LLVMdev] GSOC Project Proposal: Profile-guided optimizations

1. Summary I will implement two optimizations in the LLVM compiler, both based on runtime profile information: an inlining mechanism that takes the call frequency into consideration, and a better basic block placement algorithm. 2. The project LLVM now includes an efficient framework [1] for instrumenting an application and collecting edge and path profile information. Profile-guided

Hi, Gratian, I did that Summer of Code. I used a different algorithm than Patterson's. It is a constraint system by Su and Wagner, which is more modern, and has some advantages over older works. In particular, it is non-iterative. I found it very hard to compare it with Patterson's analysis, because there is not much description in that paper. However, there is another paper, by

Hi! I'm a student who would like to participate on Google SOC for LLVM, and was thinking about what project to pick. I saw on the "Open projects" page that Value Range Propagation is not implemented and thought about doing it, based on a paper by Patterson (it's also used by GCC). But then I saw that last year someone did a Range Analysis pass that seems to do pretty much the

On Dec 28, 2010, at 9:39 AM, Lup Gratian wrote: > I recently downloaded LLVM 2.8 and started playing with the optimizations a bit. > I saw something curious while trying the following function: > > int g(unsigned int a) { > unsigned int c[100]; > c[10] = a; > c[11] = a; > unsigned int b = c[10] + c[11]; > > if(b > a*2) a = 4; > else a = 8; >

On Dec 28, 2010, at 12:48 PM, Chris Lattner wrote: > On Dec 28, 2010, at 9:39 AM, Lup Gratian wrote: >> I find it strange that it hasn't found that %add and %mul have the same value, %cmp would be then false, selecting and returning 15. If 'a' is replaced by a constant it works. > > You're right, that is a missed optimization. I added it to the missed optimization

Hello: I am following the example code in BasicBlockPlacement.cpp. That transformation relies on basic block execution frequencies that have been read from a profile file. This is how it loads the profile data: 1. In getAnalysisUsage, it calls addRequired<ProfileInfo>(), and 2. In the runOnFunction, it calls getAnalysis<ProfileInfo>() which returns a ProfileInfo object pointer.

BTW, have you look at MMTk (http://jikesrvm.org/MMTk) ? This is the garbage collection library that underlies JikesRVM. It is a 'research-oriented' implementation, meaning that it has lots of configurable settings and plugin interfaces for implementing a broad range of collection algorithms. I was amused by the fact that "building a hybrid copying/mark-sweep collector" is

Thanks a lot for replying Chris, I m trying to randomize the blocks in a program.I generate a random number( between the current "InsertPos" and the last block), and then iterate through the list of basicblocks , picking up block with position equal to that of the random number and place it into the current InsertPos and increment InsertPos. Running it like this:

I recently downloaded LLVM 2.8 and started playing with the optimizations a bit. I saw something curious while trying the following function: int g(unsigned int a) { unsigned int c[100]; c[10] = a; c[11] = a; unsigned int b = c[10] + c[11]; if(b > a*2) a = 4; else a = 8; return a + 7; } The generated code, with -O3 activated, is define i32 @g(i32 a) nounwind readnone { %add = shl i32

Hi I am modifying LLVM to enable runtime feedback directed optimizations. I have modified the instrumentation part of LLVM to keep instrumentation data in some pre-setup structures in memory and use the JIT to recompile some of the hot functions with profiling information. I need an analysis pass which loads the profiling data in memory and can be used by different optimization passes in LLVM,

Hi, I have been trying to randomize blocks in a program and modified "BasicBlockPlacement.cpp" for the purpose but getting segmentation fault.I am not able to determine the problem.Can anyone please decrypt these error messages or suggest what might be the possible cause of failure? -----------------------------------------------------------------------------------------------------

On Mon, 7 Feb 2005, Tanu Sharma wrote: > I have been trying to randomize blocks in a program and modified > "BasicBlockPlacement.cpp" for the purpose but getting segmentation > fault.I am not able to determine the problem.Can anyone please decrypt > these error messages or suggest what might be the possible cause of > failure? I'd be happy to fix this, but I need

Hi! I'm interested in profile-guided optimizations and was looking at the functionality LLVM provides. I have two questions: - can the current (optimal) edge profiling be used to determine the number of times a function calls another one? I need to know not only how many times a function was called, but also by whom. Or to say in in a different way, can a profile edge be formed from blocks

> 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? Target Data Layout Create Garbage Collector Module Metadata

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. >

> It *is* allowed to access ModulePass analysis information from an > FunctionPass? BasicBlockPlacement (a FunctionPass) also accesses the profile information and I assumed it worked (but haven't independently verified this). > Can you try to manually override the PI value in the > MyCodeGenPass::runOnMachineFunction() to the value seen in llc and then > access the

> On Sep 18, 2017, at 5:17 PM, Kyle Butt <iteratee at google.com> wrote: > > > > On Mon, Sep 18, 2017 at 1:16 PM, Andrew Trick <atrick at apple.com <mailto:atrick at apple.com>> wrote: > >> On Sep 14, 2017, at 6:53 PM, Kyle Butt via llvm-dev <llvm-dev at lists.llvm.org <mailto:llvm-dev at lists.llvm.org>> wrote: >> >> I plan on

Hi, > On Mon, 27 Sep 2004, Zhang Qiuyu wrote: > > > Is there a way to modify the string such as char a or char b? Could I > > use the way like "Replace an instruction with another Value" in Programm > > Manual? In fact, what I am interested in is string with visible > > expression, not all string, and I am trying to hide the orignal string > > by

Does someone know the rules by which 'coxph' returns 'frail', the predicted frailty terms? In my test function: ----------------------------------------------- fr <- function(){ #testing(frailty terms in 'survival' require(survival) dat <- data.frame(exit = 1:6, event = rep(1, 6), x = rep(c(0, 1), 3),

On Feb 26, 2009, at 12:25, Chris Lattner wrote: > On Feb 26, 2009, at 12:02 AM, Talin wrote: > >> With the increasing number of LLVM-based VMs and other projects, I >> suspect that the desire for more comprehensive garbage collection >> support in LLVM is only going to increase. > > What you see in LLVM right now is really only the second step of the >