similar to: [LLVMdev] llvm interprocedural analysis and optimization

There is/are implicit dependency for the optimization on its analysis. So, if you run the optimization, the analysis will be turned on implicitly, through the PassManager. Chuck On 2/12/2012 10:10 AM, Xin Tong wrote: > If I turn on one of the llvm interprocedural optimizations without > turning on the analysis it uses. will the analysis be turned on > automatically ? > > Thanks

Hello I am trying to enable globalsmodref-aa by default. globalmodref-aa is a ModulePass and therefore can be invalidated and need to be rerun. I see globalsmodref-aa is enabled in LTO passmanager by adding the analysis pass explicitly. I wonder whether globalsmodref-aa can be enabled/run based on pass dependencies, i.e. those indicated in getAnalysisUsage(). Thanks, Xin -------------- next

Hi Xin, Thank you for your reply! I have tried the 3 alias analyses you have mentioned on LLVM 3.5: 1) $ opt -globalsmodref-aa -aa-eval < xxx.bc > /dev/null (May-alias response 100%) 2) $ opt -tbaa -aa-eval < xxx.bc > /dev/null (May-alias response 100%) 3) $ opt -cfl-aa -aa-eval < xxx.bc> /dev/null (Unknown command line argument '-cfl-aa') It seems that they are not

Dear all, I was wondering if there are some reliable alias analyses build on top of LLVM other than basicaa. Thank you! Zhiyuan -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20150407/db07dba3/attachment.html>

Hi I have this program in which fooBuf can only take on NULL or the address of local_fooBuf, and fooBuf and local_fooBuf have scope of the foo function. Therefore there is no way for the fooPtr argument to alias with fooBuf. However, LLVM basicaa and globalsmodref-aa say the 2 pointers may alias. I am thinking whether i should implement a limited form of point-to alias on the fooBuf pointer in

I still can't build LLBrowse on my Debian5-i386 machine today, The following is a full build console output. I am using LLVM-2.8 release build, with needed wxWidgets and CMake. Thank you Chuck sideshow.eecg>time cmake ../llbrowse -- The C compiler identification is GNU -- The CXX compiler identification is GNU -- Check for working C compiler: /steffan/a/a0/czhao/bin/bin32/gcc -- Check

OK try it now - I checked in a few more fixes. On Tue, Feb 22, 2011 at 8:29 AM, Chuck Zhao <czhao at eecg.toronto.edu> wrote: > I still can't build LLBrowse on my Debian5-i386 machine today, > The following is a full build console output. > I am using LLVM-2.8 release build, with needed wxWidgets and CMake. > > Thank you > > Chuck > > sideshow.eecg>time

On Sat, Feb 19, 2011 at 12:27 PM, Talin <viridia at gmail.com> wrote: > LLBrowse - a GUI tool which allows you to inspect the contents of LLVM > modules - now runs on Linux and OS X, and it works with both LLVM 2.8 and > current LLVM head. I've updated the docs to include instructions on checking > out and building the code under several different environments, which you

On 7/27/2010 12:40 PM, Devang Patel wrote: > On Tue, Jul 27, 2010 at 7:46 AM, Chuck Zhao<czhao at eecg.toronto.edu> wrote: >> LLVM (2.7 release version) provides 2 implementations for inlining >> function callsites: >> >> - InlineSimple.cpp (-inline): inline simple callsites >> according to its cost analysis >> - InlineAlways.cpp

LLBrowse - a GUI tool which allows you to inspect the contents of LLVM modules - now runs on Linux and OS X, and it works with both LLVM 2.8 and current LLVM head. I've updated the docs to include instructions on checking out and building the code under several different environments, which you can read here: http://llvm.org/svn/llvm-project/llbrowse/trunk/doc/LLBrowse.html (the doc also

LLVM (2.7 release version) provides 2 implementations for inlining function callsites: - InlineSimple.cpp (-inline): inline simple callsites according to its cost analysis - InlineAlways.cpp (-always-inline): inline all callsites that are marked with "always_inline" attribute. They are both subclasses of Inline.cpp that assumes the function's definition (body) is

Hi, My apologies for the trouble. I've disabled building DynamicTypeChecks for now (r138224) and now it builds cleanly on 32bit for me here. As for SAFECode support for 32bit vs 64bit, I believe 32bit should work just fine although I haven't personally tested this. Let me know if you have any further issues/questions. ~Will On Sun, Aug 21, 2011 at 9:26 AM, Chuck Zhao <czhao at

Should not the 4.0.0 release be March _2017_ in the "Release Emails" section on http://llvm.org/ ? -Xin

LLVMers, I need to intercept all LLVM IR instructions that will write into memory and start to do analysis on these instructions. In addition to StoreInst, what are all other IRs that will write into memory? E.g. char * ptr = malloc(...); ///... //with use(s) of ptr later The LLVM IR for the above code would be: %0 = tail call noalias i8* @malloc(i32 137) nounwind ///... //

I do not know too much about clang, is opencl publicly supported in clang ? how about cuda ? Thanks Xin

I have a question on the LLVM JIT I did some brief memory reading one day and I found that a call to a non-library function is resolved by the X86CompilationCallback, but the X86CompilationCallback is reached through a trampoline. why can not the generated code jump to the X86CompilationCallback function directly ? 0x2b0a6a4d103b: mov $0x2b0a6a561010,%rax 0x2b0a6a4d1045:

I understand that we need to push the address to a register then branch using the register. But i am asking why there is a trampoline there such that a call to foo is first branched to an snippet and the snippet branches to the X86CompilationCallback. is this snippet necessary ? Thanks Xin On Tue, Feb 22, 2011 at 12:39 PM, Reid Kleckner <reid.kleckner at gmail.com>wrote: > The

but does it have default values ? Thanks Xin On Tue, Mar 13, 2012 at 5:19 AM, Che-Liang Chiou <clchiou at gmail.com> wrote: > You specify shader model, bit size and etc. arch-specified parameters > though -march, -mattr and -mcpu, but AFAIK, PTX backend does not use > the GPU thread/block/grid size information in optimization yet. > > On Mon, Mar 12, 2012 at 8:17 PM, Xin

The address of the callee may be more than 2 GB away in memory, which cannot be encoded as an immediate offset in the call instruction. So, the value is first materialized with a mov instruction which can encode the immediate and then jumped to through a register. Reid On Tue, Feb 22, 2011 at 12:03 PM, Xin Tong Utoronto <x.tong at utoronto.ca> wrote: > I have a question on the LLVM JIT

John, The release source code (sc-main.tar) won't compile cleanly under Debian6-i386 (gcc/g++: 4.4.5). The compiler back trace is attached. Please fix it/them and repost. Or, 64b system is a requirement? Thank you Chuck llvm[4]: Compiling TypeRuntime.cpp for Release+Asserts build (PIC) cc1plus: warnings being treated as errors