On Thu, 06 Apr 2000, you wrote:
> >One of the more notable examples is for the Alpha. The Digital
> >c89 compiler produces code through an optimizer. Then the
> >assembler does peephole optimization. And then the linker
> >does global dead code removal. All it means is that the days
> >of writing tight hand coded assembler is long gone :-)
>
> When I worked with my first RISC chip (an early MIPS
> processor, circa 1990) I was briefly skeptical, but MIPS
> used (possibly even invented) that same strategy: translate
> your C with a very good compiler and a very good assembler
> and feed the results through their magic linker that chewed
> up EVERYTHING and spit it out sometimes totally rearranged.
> The resultant machine code could be so subtle and elegant,
> showing almost human insight and trickery, that on at least
> one occasion I can recall my hair literally standing on end.
> Of course, it was a real joybuzzer any time we had to
> disassemble/debug such ultra-optimized code since it
> seemed to be entirely unrelated to the original C sources...
I was making my living porting a Unix database engine when MIPS got started.
They claimed that their technology was 1/2 hardware and 1/2 software. They
were selling the combination of their chip and their compiler. It blew
everything else around completely away. The assembler manual was a really
strange read. It talked about concepts that sounded like the things quantum
physicists talk about. "Violation of causality" comes to mind. I'm pretty sure
they were doing speculative execution, and playing games with the instruction
pipeline. It was clear that if you weren't writing a compiler for that thing,
you didn't want to know.
>
>
> **********************************************************
> To unsubscribe from this list, send mail to
> [EMAIL PROTECTED] with the following text in the
> *body* (*not* the subject line) of the letter:
> unsubscribe gnhlug
> **********************************************************
--
email me at [EMAIL PROTECTED]
"Somebody once told me that in basketball you can't hold the ball and run.
I got a basketball and tried it and it worked just fine. He obviously didn't
understand basketball." Attributed to Roger Miller.
**********************************************************
To unsubscribe from this list, send mail to
[EMAIL PROTECTED] with the following text in the
*body* (*not* the subject line) of the letter:
unsubscribe gnhlug
**********************************************************
