[rfc] improving 32bit user performance/experience...

[Christopher Chan]

> I challenge anyone to find someone using Ubuntu 8.10/9.04 on a
> processor which doesn't support the full i586 instruction set (eg
> i386/i486 or something with incomplete i586 support).
>   
i586 binaries should be only installed on actual Pentium computers. 
Every other subsequent cpu's architecture all handle i486 or i686 and 
above better than they would handle i586 tuned binaries. The Mandrake 
business of releasing i586 is plain dumb and only relevant to actual 
Pentium processors.

> All older VIA processors, AMD Geode procs and so on support the full
> i586 instruction set, which including MMX instructions and registers,
> which itself can provide a good win.
>   
You might find that i586 specific optimization will actually be slower 
on those processors than i486 optimization. Have you done a test at all?

> Also, even if 1% of the users use i586, we can allow instruction
> scheduling for deeper pipelines (with eg -mtune=i686) for the other
> 99%, generating fewer stalls on more modern processors - important
> with far higher core vs memory latency. This gains more for i686 than
> it loses for i586.
>
>   
Pentiums are so rare. I still have one but I sure am not going waste my 
time getting Pentium specific tuning.

>
> I think we have opportunity; Gentoo users tackled this problem in
> their way 10 years ago.
>   

I wonder how many had actual Pentiums and how many months it took to 
compile.