Re: IBM 6x86MX233 blacklisted ?

Peter T. Waltenberg (peterw@karaka.chch.cri.nz)
Tue, 03 Feb 1998 08:44:47 +1300 (NZDT)


On 02-Feb-98 Joel Jaeggli wrote:
> Well... The IBM cpu is a Cyrix M2 which is not a bad thing in and of
> itself however, if your expecting it to outperform a Pentium-II, that's
> not going to happen. I have an M2 pr-166 (normally runs at 133)
> overclocked to 150mhz (75mhz bux x 2x clock double) and it still doesn't
> compare to a Pentium-166mmx on the same board (M-Technology 534f
> (sis-5582) with 512k cache), running at it's rated speed. The bovine rc5
> client performance page provides a reasonable metric for cpu performance
> on certain operations...
>
> http://alde.com:7070/x86speed.sql

> what it will tell you is that the m2 166 falls between the pentium-120 and
> the pentium-133.

Thats heavilly numeric intensive code. With more normal code it performs better
than that.

It probably won't outperform a P-II, then again, it's not necessarilly going
to be MUCH slower either.

> Also to build a kernel for one of these beasts, you need to apply one of
> the numerous Cyrix patches.

A standard kernel works fine. Using set686 will squeeze a bit more performance
out.

> On Mon, 2 Feb 1998, Steffen Grunewald wrote:
>
>> Hi list,
>>
>> recently I got an offer for an IBM6x86MX233 based computer. Looks
>> pretty good (even a bit faster than iPII-233 due to the larger cache)...
>> Are there any arguments against buying such a beast ?
>> Is there a FAQ on Linux' CPU support ? Perhaps a comparison of
>> performance and throughput ?
>> Are there (besides the ones that appeared in this mailing list past
>> weekend) any blacklisted single-processor main boards ?

Dual P133 does kernel builds (make -j2) in 5-6 minutes,
Cy686MMX (2*75M) PR200 3.5 minutes.

Check the Cyrix home pages. www.cyrix.com and
http://wauug.erols.com/~balsa/linux/cyrix/

Value for money I'd say they are an excellent buy.

As a general comment, if you plot CPU benchmarks with the axis starting at 0 :)
you'll see there's damn all difference in performance between the various high
end machines. Plotting performance against $ is even more interesting.
Work out how much you are prepared to spend for 5% extra performance ...

The only bad thing I can say about them is that they often get put into cheap
machines with cheap motherboards.

>> Every hint is welcome.

Peter
----------------------------------
E-Mail: P.Waltenberg@irl.cri.nz
Date: 03-Feb-98
Time: 08:32:16
----------------------------------