Re: TI-H: risc
[Prev][Next][Index][Thread]
Re: TI-H: risc
>That isn't a very good example Grant. That is just one of the many commands
>that RISC and CISC have. If I wanted a RISC processor to accomplish one of
>the CISC 59 new "MMX" commands, then it would take the RISC processor more
>clocks to get it accomplished. To set a greater example, if I wanted to
>make a game based on MMX commands, then I wanted to port it to an IBM RISC,
>the game would ultimately be slower, unless I knew the IBM RISC commands (if
>any) that were similar to the MMX commands. Even pipelining wouldn't help
>the RISC accomplish it, since CISC can play that game to. On the contrary,
>if I wanted to program something in RISC, then it may be easily applicable
>to CISC, although CISC would probably be slower, for obvious reasons. Like
>I said the first time: It ALL depends upon the application.
No one writes games in ASM for windows computers. Writing programs in ASM,
even under linux, is very very hard and it would be faster to use a high
level C language...
The MMX engine on my PPC runs at about 234MHz. That means my 300MHz G3 is
able to do a 233MHz MMX... Pretty good.
And that is emulating. If the game is ported and recompiled, it would run
a ton faster.
Windows is also the worst OS to program anything under. All of the top
compilers aren't posix compatible. That means they don't use the standard
for code.
I can take a unix app, and port it to a posix target easily.
Once again, good points, but no base...
Follow-Ups:
References: