🚀 go-pugleaf

In <3vivd6$jqo@nnrp2.primenet.com> english@primenet.com (Lawson English) writes:

>Dan Pop (Dan.Pop@mail.cern.ch) wrote:
>
>: For the same clock comparisons, here you are:
>
>: System            CPU        ClkMHz  Cache      SPECint SPECfp  Info  Source
>: Name              (NUMx)Type ext/in  Ext+I/D    92      92      Date  Obtained
>: ================= ========== ======= ========== ======= ======= ===== ==========
>: SNI/Pyr PC/E5S    Pentium    33/66   256+8/8      67.4    61.5  Sep93 SPEC newsl
>: Compaq DeskproXL  Pentium    33/66   256+8/8      65.1    63.6  Sep93 SPEC newsl
>: IBM 250           MPC601     66      32           62.6    72.2  Jul94 www.ibm
>
>: I avoided the Intel Xpress numbers and stuck to Siemens Nixdorf and Compaq,
>: for obvious reasons.
>
>: The SPEC apps don't seem to run faster on the 601/66 than on the P5/66.
>: Oh, sorry, I forgot that SPEC data are pure bullshit (because they show the
>: Pentiums faster than "everbody" knows they actually are :-)
>
>I don't know about the SPEC data being pure bs, but your numbers for the 
>Pentium systems certainly are: the 66mhz Pentium systems ALL have 66mhz 
>data buses, not 33mhz data buses -that's PCI speed.

It's not my data.  It's freely available from 
ftp.cdf.toronto.edu:/pub/spectable
>
>OTOH, the IBM 601 machine IS running with a 33mhz data bus, and has no L2 
>cache.
>
>Add cache to that machine and rerun the benchmarks. Given your emotional 
>position on the subject, you will be unpleasantly surprised.

OK.  Here are the numbers for a system designed for 2nd level cache
(the 250 apparently isn't, so I can't add any cache to it, sorry):

IBM 40P           MPC601     66      256+32       75.1    77.0  Mar95 www.ibm

Should I be "unpleasantly surprised"?

Dan
--
Dan Pop
CERN, CN Division
Email: Dan.Pop@mail.cern.ch 
Mail:  CERN - PPE, Bat. 31 R-004, CH-1211 Geneve 23, Switzerland