Hi Philip, et al,
Thank you for your excellent work on PCRE. I am working to advance the performance of regular expressions in the context of data mining and one thing that particularly piques my interest is offloading regular expressions onto commodity graphics processors. There are a few groups that have proven that DFAs can take advantage of the massively threaded nature of SIMD architectures, and it looks like NVIDIA's newest processor will actually use MIMD kernels. Does anyone have knowledge of PCRE working on CUDA and any particular best practices? So far the most concrete documents I've read revolve around the "gnort" team's use of PCRE on CUDA, though I do not believe this source code is publically available. Thanks! Kristopher -- ## List details at http://lists.exim.org/mailman/listinfo/pcre-dev
