Can C++ Compete with Fortran?
Recent benchmarks show C++ encroaching steadily on Fortran’s high-performance monopoly. For some benchmarks, C++ is even faster than Fortran! No, this is not a result of better optimizing compilers, preprocessors, or even language extensions. It’s achieved simply through the use of template techniques. Through the magic of template metaprogramming, optimizations such as loop fusion, unrolling, tiling, and algorithm specialization can be performed automatically at compile time. Another successful package that uses this technique is, of course, the Boost C++ Libraries. Another goal of Blitz++ is to extend the conventional dense array model (in other words, not sparse) with new and useful features. Some examples of such extensions are flexible storage formats, tensor notation, and index placeholders. Blitz++ is more than just cool arrays; you also get matrix math, random number generation, and a variety of supporting functionality.
