similar to: [LLVMdev] Make LoopBase inherit from "RegionBase"?

Why not use the "standard" algorithm for detecting SESE-regions and building a program structure tree? It should handle everything you want. It also becomes much simpler to specify a connected SESE-region by entry/exit edges, while a disconnected region is specified by entry/exit blocks. Only defining regions on blocks is not enough to be able to quickly determine how to replace/move a

hi all, On 2010-1-7 0:11, John Mosby wrote: > In LLVM we could add support for generalized CFG regions and > RegionPasses. A region is a part of the CFG. The only information we > have is, that it has one entry and one exit, this it can be optimized > separately. > I think this is the best way to add region analysis. I must admit > this approach > helps me on another,

On 01/08/10 14:20, ether wrote: > sorry that i forgot to change the subjuect > > > hi all, Hi ether, now a kind of more complete answer. > On 2010-1-7 0:11, John Mosby wrote: >> In LLVM we could add support for generalized CFG regions and >> RegionPasses. A region is a part of the CFG. The only information we >> have is, that it has one entry and one exit, this

On 01/20/10 18:16, Jan Sjodin wrote: >> I think this should be implemented as a RegionFilter, that checks if a > >> region contains a loop, and that can be asked for further information. >> In general I do not think this kind of analysis belongs to a region, but >> as you proposed some kind of filter could be applied. In the short term >> the passes who need this

On 01/15/10 03:03, ether wrote: > hi, Hi, > about region class, i think provide some method to extract the top level > loop from a given region is useful in polly(our polyhedral optimization > framework), because the we are only doing optimization on the region > with loop. Hi ether, I think this should be implemented as a RegionFilter, that checks if a region contains a loop,

On 01/21/10 20:42, Jan Sjodin wrote: >>> Imo, a loop is simply a special kind of region, so a "filter" is perhaps the way to >>> go if you are interested in loops. Regions containing loops will have to be inspected >>> using the PST. >> >> Except loops that have multiple exits. they are not necessarily (single >> entry single exit) region, if

On 01/13/10 00:56, Jan Sjodin wrote: > Why not use the "standard" algorithm for detecting SESE-regions and building a program structure tree? > It should handle everything you want. It also becomes much simpler to specify a connected SESE-region > by entry/exit edges, while a disconnected region is specified by entry/exit blocks. Only defining regions on > blocks is not

On 01/20/10 18:05, Jan Sjodin wrote: >>>> bbs, to generate these edges. As we do not have to modify the CFG other >>>> passes like dominance information are still valid, and we do not have to >>>> create a lot of auxiliary bbs, to be able to detect all regions. This >>>> saves memory and runtime. In general it is probably not too easy to

Hi, all. I'm using void LoopBase::getExitBlocks (SmallVectorImpl< BlockT * > &ExitBlocks) const to get all exit blocks for a loop. The problem I find with this API is that it returns repetitive basic blocks in certain situations. Should repetitive basic blocks be removed? I have an example to show the problem. Following is the source code and the CFG is enclosed. int main() { int

On 01/13/10 05:09, Jan Sjodin wrote: > Hi Tobias > >> In general this approach saves a preliminary pass that has to insert new > >> bbs, to generate these edges. As we do not have to modify the CFG other >> passes like dominance information are still valid, and we do not have to >> create a lot of auxiliary bbs, to be able to detect all regions. This >>

On Apr 19, 2012, at 10:43 PM, xiaoming gu <xiaoming.gu at gmail.com> wrote: > Hi, all. I'm using void LoopBase::getExitBlocks (SmallVectorImpl< BlockT * > &ExitBlocks) const to get all > exit blocks for a loop. The problem I find with this API is that it returns repetitive basic blocks in certain > situations. Should repetitive basic blocks be removed? Users

Thanks, Also, every method inherited by LoopBase causes the same error, while Loop methods go smooth. On 9 May 2013 01:05, Andrew Trick <atrick at apple.com> wrote: > > On May 8, 2013, at 7:43 PM, Giacomo Tagliabue <giacomo.tag at gmail.com> > wrote: > > Hello, > I am building a loop pass following these instructions: >

Wow, commenting those two lines worked out fine for me, thanks! On 9 May 2013 09:34, Giacomo Tagliabue <giacomo.tag at gmail.com> wrote: > Thanks, > Also, every method inherited by LoopBase causes the same error, while Loop > methods go smooth. > > > On 9 May 2013 01:05, Andrew Trick <atrick at apple.com> wrote: > >> >> On May 8, 2013, at 7:43 PM,

On May 8, 2013, at 7:43 PM, Giacomo Tagliabue <giacomo.tag at gmail.com> wrote: > Hello, > I am building a loop pass following these instructions: http://llvm.org/docs/WritingAnLLVMPass.html > Everything works fine, I did it many times for Function Passes, but in the runOnLoopmethod, whenever I call a method of the loop L passed as argument, for example L->begin(), I get the

Antony Blakey wrote: > SVN head LLVM and Clang built out of the box for me a week ago on > MSYS/MINGW, using the following files: Ok, I tried to do the same. The compilations stops with the following error message (actually, I got to the same point in my attempt to build LLVM with the MinGW compiler shipped with Cygwin): llvm[1]: Compiling MachineLoopInfo.cpp for Debug build

On Apr 23, 2012, at 10:31 PM, Andrew Trick <atrick at apple.com> wrote: > > On Apr 19, 2012, at 10:43 PM, xiaoming gu <xiaoming.gu at gmail.com> wrote: > >> Hi, all. I'm using void LoopBase::getExitBlocks (SmallVectorImpl< BlockT * > &ExitBlocks) const to get all >> exit blocks for a loop. The problem I find with this API is that it returns

Hello, I am building a loop pass following these instructions: http://llvm.org/docs/WritingAnLLVMPass.html Everything works fine, I did it many times for Function Passes, but in the runOnLoopmethod, whenever I call a method of the loop L passed as argument, for example L->begin(), I get the following error: opt: symbol lookup error: /home/giacomo/llvmcsfv/Debug+Asserts/lib/Acsl.so: >

Hi, LoopBase::getExitBlocks in LoopInfo.h returns all exit basic blocks by iterating over all blocks of a loop and storing all branch targets that are outside the loop. This method allows for duplicates in the result vector, e.g.: Loop at depth 2 containing: %bb2<header><exit>,%bb1<latch>,%bb3<exit> where both bb2 and bb3 have bb4 as their exit block. This loop

hi, > Users generally expect a unique set of exit blocks, but don't make any > strong assumption. The worst that can happen is missed optimization or > redundant analysis. In most cases, the loop is in LoopSimplifyForm, so it's > probably not a problem in practice. > > > Another thing I should mention. The iteration order of ExitBlocks is > important. In llvm,

Hi, I was reading some loop related code and I don’t quite understand an assertion in LoopBase<BlockT, LoopT>::getLoopPredecessor(). /// getLoopPredecessor - If the given loop's header has exactly one unique /// predecessor outside the loop, return it. Otherwise return null. /// This is less strict that the loop "preheader" concept, which requires /// the predecessor to have