Reply With QuoteWell it doesn't help every tutorial and example on the web is DX centered (yes I know there are a few *old* openGL ones).Originally Posted by Pawlerson
I programmed in assembly because single core low IPC/frequency processors needed applications to be as speedy as possible back then, GUI programming sucked because you relied on slow API's that could destroy any benefit from LLP. Now you can program in whatever filthy abstract scripting lang and still have a decent performing app/game because GPU's (and processors) are insane now. Abstracting layers like SDL maybe too slow 10 years ago but now, there is absolutely no excuse to not abstract game/program layers to be cross platform except because M$ paid them to do so... or they're lazy... or they don't care about cross platform... or they think Linux is commie crap... oh well, maybe there is an excuse or two,but that's about what they are is excuses.
Yes, yes I have. The Win32 api had to be one of the most performance robbing hack jobs I have ever used. To it's credit, it was fairly standard and results were consistent (if not because most of the irrational behavior had been discovered and worked around). I was always trying to find ways to speed up Windows apps, especially GUI ones. The sad thing is no matter how awesome of a jump table window handler loop you made, no matter how quick hashed for clicks and closes, the best loops barely did crap compared to the msecs lost by the Win api... Yeah all gui's suck but still I won't praise that crap.
Just so you know.. A hyper visor with running on a processor that launches apps that have full control of the workhorse processor and the workhorse video card is by far the best way to approach gaming and critical applications. In case you think I'm defending X or quartz or whatever I think they are all pretty bad.
