> The other problem in this area is that the rules for instruction
> sequencing are different on each processor revision, and can make up to
> a factor of two difference. Also the "CISC" instructions tend to be
> disproportionately slow on 486s but not on the others as far as I know.
> This means that you can write assembly code which runs well on one
> processor but not another. By contrast, gcc can be told which processor
> to optimise for so hopefully it should give more consistent performance.
>
> (I am not taking one side or the other here. I think it is evident that
> hand coded assembly language should be faster if done carefully by
> someone with a sound knowledge of the processor. On the other hand,
> taking full advantage of a modern processor is extremely difficult.)
Yes. One of the reasons why there is an 'arch' directory is so that
some platform-specific stuff can be done.
Cheers,
Dick Johnson
***** FILE SYSTEM WAS MODIFIED *****
Penguin : Linux version 2.1.127 on an i686 machine (400.59 BogoMips).
Warning : It's hard to remain at the trailing edge of technology.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/