> Pymol at least *compiles* under a vast amount of processor > architectures, if only under Linux: > http://buildd.debian.org/build.php?arch=&pkg=pymol > Nothing wrong with testing your code on as much platforms as possible of > course :)
It's harder than you think. I have a *really* old SGI, which cannot run the distributed binaries. It took me hours to get PyMOL to compile (not counting the hours it took for the compiler to actually run) with gcc. I think I had to leave some parts out anyway. Aside from the difficulty compiling, virtually every benchmark I've seen has shown binaries compiled with gcc to be far slower than ones compiled with SGI's compilers. The real problem is with OS support, not architecture support. An ideal situation would be for support for more arcane platforms to be farmed out to people who have a vested interest in seeing them supported. We do not personally use SGIs for visualization, but our experimental colleagues do, and I've seen quite a few labs at various universities that have continued to buy SGIs out of habit (and/or because they're rolling in cash, and prefer computers that work consistently). In addition, I've found that PyMOL shines when dealing with *huge* molecules, which is exactly where PC architecture's limitations show up. Our lab owns two Origins with compiler licenses, and there are a couple of us who would be happy to help with PyMOL porting. We have made extensive use of PyMOL and have integrated it into our in-house software, and I would be happy to contribute some of my time. However, my impression has been that future PyMOL development will be almost entirely centralized. ---------------------------------------------------------------------------- Nathaniel Echols Programmer n...@bioinfo.mbb.yale.edu Gerstein Lab 203-589-6765 Yale University ----------------------------------------------------------------------------
