search for: pofl

...i also add the library > build rule to MAKEFILEs, but not sure if it work under linux/cygwin/mac, > sorry. > hope this help > > best regards > > --ether [the complete mail] hi ether, great start. I pushed your work to our git repository at http://repo.or.cz/w/llvm-complete/pofl.git This repository is ment to track work on a first version of LLVM Poly that could be proposed for integration to LLVM. Feel free to register as a user and to commit to the repository. I just had some discussions on the LLVM IRC channel concerning the integration of Poly. The preferred way...

...under cmake+visual studio 2009, and i also add the library > build rule to MAKEFILEs, but not sure if it work under linux/cygwin/mac, > sorry. > hope this help > > best regards > > --ether hi ether, I pushed your work to our git repository at http://repo.or.cz/w/llvm-complete/pofl.git So we can work on first version that could be committed to the LLVM svn repository. I just had some discussions on the LLVM IRC channel concerning the integration of Poly.

...s. Polly is a polyhedral optimization framework for llvm, which similar to Graphite for gcc (http://gcc.gnu.org/wiki/Graphite). The main work flow of Polly is: 1. Transfer LLVM IR to polyhedral model represent (Polyhedral IR) with the LLVM Region framework ( http://repo.or.cz/w/llvm-complete/pofl.git/shortlog/refs/heads/regioninfo) and the Scalar Evolution infrastructure. 2. Perform transformation on Polyhedral IR, e.g. advanced automatic parallelization, optimize loops for cache locality. 3. Generate LLVM IR with necessary metadata annoted from Polyhedral IR. With Polly, we can g...

...only detect regions that fit in a schema that you have implemented. Is this correct? I would be interested how your pass handles unstructured control flow. Especially in the problematic cases I put in the RegionInfo test suite. They are available at this place: http://repo.or.cz/w/llvm-complete/pofl.git/tree/refs/heads/regioninfo:/test/Analysis/RegionInfo Ether and me are currently adding doxygen comments. If you are interested you could also have a look at our doxygen documentation http:://tobias.osaft.eu/regioninfo/doxygen To me it seems as if our approaches are close enough together to...

...interested to understand your pass and the differences between our approaches. Hopefully we could combine both of them and get LLVM a really great region detection framework. If you are interested the current code for our region pass is in this git repository: http://repo.or.cz/w/llvm-complete/pofl.git/shortlog/refs/heads/regioninfo Is there any repository for your code? Some information about our region pass can be found in this mail: http://lists.cs.uiuc.edu/pipermail/llvmdev/2010-March/029996.html I just copied the relevant part for you: -------------------------------------------------...

Hi folks, A few months back I finished writing and testing a pass which implements "structural analysis" as described originally by Sharir in 1980 ("Structural analysis: A new approach to flow analysis in optimizing compilers") and more recently by Muchnick in Advanced Compiler Design and Implementation. It analyses the CFG and recognises specific region schema, such as