Vector processing on nehelam?

[nn_step]

I'm not confused mate. Did a thesis project on Molecular Modeling of Protein structures and worked on the SW coding for it. We completely bypassed the driver layer to use the HW directly and trust me, the math for vector-based graphics is almost identical to the math for vector processing (they are both matrix calculations). Heck, the app I wrote would LOVE this feature in a processor. I wouldn't even have to recode it (unlike SSEn, which we would have to recode for)!

what instruction set did you code it for?

[nn_step]

Plain vanilla C baby. Runs on Mac OS 9, OS X, Win95 - Vista, Linux, and several Unix variants.

All I would need is a compiler update to take advantage of these features.

Then you could have step the compiler to optimize it for SSE without much recoding

[BlueBiker]

If you went to the trouble to optimize the performance of this math intensive app, it seems counterproductive to ignore the huge benefits from SSE/SSE2 that the compiler can give almost for free.

Presumably it was only a few inner loops that really needed turboing, and a runtime selection of different code paths wouldn't make much of a dent in the 650KB. Not as if it would require significant Q/A either.

Such is the PHB.

[BlueBiker]

Heckuva project to cut your teeth on. I can understand how real-world requirements would affect your choices, but if this was all in C then I don't see the boss's reluctance to compile with various optimization flags.

And really, why wasn't it written in FORTRAN?