similar to: [LLVMdev] Analysis of polly-detect overhead in oggenc

Tobi, it looks like this code is the problem: for (std::vector<Value *>::iterator PI = Pointers.begin(), PE = Pointers.end(); ;) { Value *V = *PI; if (V->getName().size() == 0) OS << "\"" << *V << "\""; else OS << "\"" << V->getName() <<

I have found that the extremely expensive compile-time overhead comes from the string buffer operation for "INVALID" MACRO in the polly-detect pass. Attached is a hack patch file that simply remove the string buffer operation. This patch file can significantly reduce compile-time overhead when compiling big source code. For example, for oggen*8.ll, the compile time is reduced from

Hi Sebastian, Yes, you have pointed an important reason. If we comment this source code you have listed, then the compile-time overhead for oggenc*8.ll can be reduced from 40.5261 ( 51.2%) to 20.3100 ( 35.7%). I just sent another mail to explain why polly-detect pass leads to significant compile-time overhead. Besides the reason you have pointed, another reason is resulted from those string

Hi, I think this should also go on the llvm-dev mailing list... On Sat, Jul 13, 2013 at 11:18 PM, Star Tan <tanmx_star at yeah.net> wrote: > > At 2013-07-14 02:30:07,"Tobias Grosser" <tobias at grosser.es> wrote: >>On 07/13/2013 10:13 AM, Star Tan wrote: >>> Hi Tobias, >> >>Hi Star, >> >>thanks for the update. I copied the polly

Chris, When asm-verbose=true, it is spending the extra time here: % cumulative self self total time seconds seconds calls s/call s/call name 33.67 17.19 17.19 849 0.02 0.04 AddModuleTypesToPrinter 17.51 26.13 8.94 265849169 0.00 0.00 llvm::Value::getType() 8.74 30.59 4.46 258948922 0.00 0.00

Hi Tobias, I have attached a patch file to optimize string operations in Polly-Detect pass. In this patch file, I put most of long string operations in the condition variable of "PollyViewMode" or in the DEBUG mode. Bests, Star Tan At 2013-07-22 22:27:48,"Tobias Grosser" <tobias at grosser.es> wrote: >On 07/22/2013 01:46 AM, Star Tan wrote: >> At

On 07/22/2013 01:46 AM, Star Tan wrote: > At 2013-07-22 12:16:53,"Tobias Grosser" <tobias at grosser.es> wrote: >> I propose two more patches: >> >> 1) Transform the INVALID macro into function calls, that format >> the text and that set LastFailure. > Translating the INVALID macro into function calls would complicate the operations for

On 07/14/2013 08:05 AM, Star Tan wrote: > I have found that the extremely expensive compile-time overhead comes from the string buffer operation for "INVALID" MACRO in the polly-detect pass. > Attached is a hack patch file that simply remove the string buffer operation. This patch file can significantly reduce compile-time overhead when compiling big source code. For example, for

On Dec 12, 2010, at 12:52 PM, Anton Korobeynikov wrote: >> Very true, but we do want the testsuite to run fast. Can it be sped up? > Can't we just run testsuite with asm-verbose=0 ? I meant llvm/test. I'm just surprised it is that slow, what is costing all the time? -Chris

On 07/22/2013 11:58 PM, Star Tan wrote: > Hi Tobias, > > > I have attached a patch file to optimize string operations in Polly-Detect pass. > In this patch file, I put most of long string operations in the condition variable of "PollyViewMode" or in the DEBUG mode. OK. > From 448482106e8d815afa40e4ce8543ba3f6f0237f1 Mon Sep 17 00:00:00 2001 > From: Star Tan

I am trying to parse LLVM IR from a bit code file. I went through the following steps. hello.cpp #include <iostream> int main() { std::cout << "Hello world!" << "\n"; return 0;} dump.cpp #include <llvm/IR/Module.h>#include <llvm/IRReader/IRReader.h>#include <llvm/IR/LLVMContext.h>#include <llvm/Support/SourceMgr.h> using

Sebastian Pop wrote: > Jack Howarth wrote: > > The attached patch eliminates http://llvm.org/bugs/show_bug.cgi?id=15817 by removing the remaining > > "; XFAIL:*" added in http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20130415/171812.html. > > The Isl/CodeGen/scevcodegen-1.ll testcase in polly appears as an XPASS in current llvm/polly 3.3 > >

I am new to LLVM and I've a question about the optimization passes and in particular about the loop transformation passes. I've setup the LLVM tool-chain with Polly installed. The following example is causing problems with me: http://llvm.org/docs/Passes.html#indvars The input is a .c file with the following loops: # 01 Standard, N=100 for (i=0; i<N; i++) b[i] = a[i] + 5; # 02

On 07/21/2013 07:33 PM, Star Tan wrote: > At 2013-07-22 01:40:31,"Tobias Grosser" <tobias at grosser.es> wrote: > >> On 07/21/2013 09:49 AM, Star Tan wrote: >>> Hi all, >>> >>> >>> I have attached a patch file to reduce the polly-detect overhead. >> >> Great. >> >>> My idea is to avoid calling TypeFinder in

On 05/31/2013 10:11 AM, Sebastian Pop wrote: > Sebastian Pop wrote: >> Sebastian Pop wrote: >>> Jack Howarth wrote: >>>> The attached patch eliminates http://llvm.org/bugs/show_bug.cgi?id=15817 by removing the remaining >>>> "; XFAIL:*" added in http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20130415/171812.html. >>>>

Star Tan wrote: > I have found that the extremely expensive compile-time overhead comes from the string buffer operation for "INVALID" MACRO in the polly-detect pass. > Attached is a hack patch file that simply remove the string buffer operation. This patch file can significantly reduce compile-time overhead when compiling big source code. For example, for oggen*8.ll, the compile

The attached patch eliminates http://llvm.org/bugs/show_bug.cgi?id=15817 by removing the remaining "; XFAIL:*" added in http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20130415/171812.html. The Isl/CodeGen/scevcodegen-1.ll testcase in polly appears as an XPASS in current llvm/polly 3.3 and trunk svn for both x86_64-apple-darwin* and x86_64 Fedora 15 when built against isl

On 05/31/2013 01:09 PM, Jack Howarth wrote: > On Fri, May 31, 2013 at 10:59:52AM -0700, Tobias Grosser wrote: >> On 05/31/2013 10:11 AM, Sebastian Pop wrote: >>> Sebastian Pop wrote: >>>> Sebastian Pop wrote: >>>>> Jack Howarth wrote: >>>>>> The attached patch eliminates http://llvm.org/bugs/show_bug.cgi?id=15817 by removing the

Mmm I found out a very strange behavior (to me) of the SCEV analysis of the loop bound of the external loop I posted. When in ScopDetection it gets the SCEV of the external loop bound in the "isValidLoop()" function with: const SCEV *LoopCount = SE->getBackedgeTakenCount(L); It returns a SCEVCouldNotCompute, but if I change the "if" block inside the loop from: if

On 01/16/2012 01:19 PM, Pieter Custers wrote: > I am new to LLVM and I've a question about the optimization passes and > in particular about the loop transformation passes. I've setup the LLVM > tool-chain with Polly installed. > > The following example is causing problems with me: > http://llvm.org/docs/Passes.html#indvars > > The input is a .c file with the