I see a couple companies here throwing around the term "high performance", and as someone who has worked with actual high performance computing (HPC) technology, the perversion of this term is annoying.<p>What is HPC? HPC is writing software for supercomputers. HPC is implementing parallel algorithms. HPC is Intel's lab on a chip, a Linux cluster with Infiniband networking, etc.<p>> SpinPunch makes high performance browser games to rival PC and console titles.<p>No, SpinPunch, my browser is not a Linux cluster with 1024 nodes, each with 8 CPUs. JavaScript is not HPC. C, C++, and Fortran are high performance. MPI and the Intel Math Kernel Library are high performance. Using MPI-IO to interact with a parallel virtual file system to read a 1 GB file efficiently is high performance.<p>> SpaceBase is a server-side, in-memory, high-performance, concurrent and distributed spatial data-store...SpaceBase is implemented in Java, and provides a Java and C++ APIs.<p>That sounds great, SpaceBase, but I didn't see any white papers or mentions of academic journal articles on your site. If your revolutionary system is high performance, you have to prove it. At the very least, you need a white paper to show how fantastic your HPC software is. Here are a few examples of that:<p>Google MapReduce: <a href="http://static.googleusercontent.com/external_content/untrusted_dlcp/research.google.com/en/us/archive/mapreduce-osdi04.pdf" rel="nofollow">http://static.googleusercontent.com/external_content/untrust...</a><p>Message Passing Interface (MPI): <a href="http://www.sciencedirect.com/science/article/pii/0167819196000245" rel="nofollow">http://www.sciencedirect.com/science/article/pii/01678191960...</a><p>Intel Math Kernel Library: <a href="http://software.intel.com/en-us/articles/intel-math-kernel-library-white-papers/" rel="nofollow">http://software.intel.com/en-us/articles/intel-math-kernel-l...</a>
ScaLAPACK: <a href="http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=234898&tag=1" rel="nofollow">http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=234898&...</a><p>Elemental: <a href="http://www.cs.utexas.edu/~flame/pubs/FLAWN44_revised.pdf" rel="nofollow">http://www.cs.utexas.edu/~flame/pubs/FLAWN44_revised.pdf</a><p>You're walking a very fine line when you claim that your system is high performance, and I am not impressed by the total lack of any paper describing how you came to this conclusion. I found this snippet from the company's blog:<p>> It scales gracefully across CPUs and across a computing grid. It scales so well that it allows building large shardless MMO games. In-fact, shardless MMO games and virtual worlds were one of the use-cases SpaceBase was specifically built to handle. (Distributed SpaceBase, or SpaceBase-on-a-grid is currently in the advanced stages of testing, and will be available for evaluation and purchase within a couple of months. The single-node, multi- and many- core deployment is available for evaluation right now!)<p>Please, just show us the numbers!<p>tl;dr HPC is all about scaling across computing clusters. Show us an experiment with speedup numbers.