Nelson H. F. Beebe wrote: > Robert Seacord of CERT posted a note on another mailing list about > some work done colleagues of mine at the University of Utah on > generating random code sequences that can be used to uncover compiler > bugs: > > Volatiles Are Miscompiled, and What to Do about It > Eric Eide & John Regehr > Proceedings of the Eighth ACM and IEEE International > Conference on Embedded Software (EMSOFT), > Atlanta, Georgia, USA, Oct. 2008 > http://www.cs.utah.edu/~regehr/papers/emsoft08-preprint.pdf > http://www.cs.utah.edu/~eeide/emsoft08/ > > Finding and Understanding Bugs in C Compilers > Xuejun Yang, Yang Chen, Eric Eide, & John Regehr > http://www.cs.utah.edu/~regehr/papers/pldi11-preprint.pdf > http://embed.cs.utah.edu/csmith/ > > The papers' authors have reported hundreds of such bugs to the > developers of gcc and llvmc, and generally got prompt fixes. > > The software from both papers is intended to be generally available. > I successfully built that from the first paper yesterday. The Web > site for the second package still says ``Csmith is coming soon''. > > I believe that it would be worthwhile for Open64 developers to > coordinate use of these valuable tools in routine testing of the > Open64 C and C++ compilers. > > The basic idea is that each of the (possibly 250,000 different) random > programs generated by the tools are compiled and run with multiple > compilers and optimization levels. Their output contains a checksum > of global variables that are computed by legal and standard-conforming > C code, and any discrepancy in the output of the same program built > with different compilers or optimization choices indicates a > code-generation bug. > afaik the generated programs are already available and there's really no need to wait for csmith http://embed.cs.utah.edu/embarrassing/jan_10/harvest/all.tar.bz2
While it's "fun" to find bugs in the compiler there's generally enough of those going around already to sufficiently keep people busy. While they may overlap with things from real programs I'd think focusing on user reported bugs is more important.. (Someone does in fact use Open64 for something other than benchmarks, right?) ------------------------------------------------------------------------------ Enable your software for Intel(R) Active Management Technology to meet the growing manageability and security demands of your customers. Businesses are taking advantage of Intel(R) vPro (TM) technology - will your software be a part of the solution? Download the Intel(R) Manageability Checker today! http://p.sf.net/sfu/intel-dev2devmar _______________________________________________ Open64-devel mailing list Open64-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/open64-devel
