Subject: Relative performance figures, FYI...
To: None <port-i386@NetBSD.ORG>
From: Michael L. VanLoon -- HeadCandy.com <michaelv@MindBender.serv.net>
List: port-i386
Date: 10/06/1996 15:34:16
FYI...  I finally got a working Pentium Pro motherboard and CPU.  I
had been having some problems with a SuperMicro motherboard.  I don't
know if it was the motherboard or CPU.  On Friday I received an Asus
motherboard and 200MHz CPU to replace the faulty ones.  It worked
perfectly.


And here are some relative performance figures:

AMD 5x86 133MHz, NICE Super-EISA 486 motherboard, 24MB 60ns RAM, 512K
L2 writeback cache, BusLogic BT747s EISA SCSI controller (on a 2GB
Barracuda and a 1GB Hawk SCSI drives):

    Make world(*): ~6 hours, depending on various factors.

Intel Pentium 120MHz (**), Asus P/I-P55TP4N (Triton-1) motherboard,
64MB 60ns EDO RAM, 512K L2 pipeline-burst cache, BusLogic BT956c PCI
SCSI controller (same drives as above):

    Make world(*): 3 hours, 10 minutes, totally clean (no pre-built
                   man pages, dependencies, etc.).

Intel Pentium Pro 200MHz, Asus P/I-P6NP5 (Natoma) motherboard, 64MB
60ns EDO RAM, 256K L2 P6 cache, BusLogic BT956c PCI SCSI controller
(same drives as above):

    Make world(*): 1 hour, 21 minutes, totally clean (same as above).


Wow.


A complete clean build of my very bloated kernel (including make
clean, config, make depend, make) takes a little over fifteen minutes
on my P5/120, and just over seven minutes on my P6/200.


Wow.


* -  My own custom "make world" script that does things in a specific
     order, and does a few things redundantly.  Building the
     NetBSD-1.2 sources.

** - I have tested it at both 100MHz (33/66MHz bus), and at 120MHz
     (30/60MHz bus) (20% faster CPU, 10% slower bus).  It is just
     barely faster at 120MHz than at 100MHz, for all the tests I ran,
     in spite of the slower bus.  However, the difference is so small
     that I can conclude (for my use anyway), that a 150MHz Pentium
     would be slower than a 133MHz (12.8% faster CPU, 10% slower bus),
     and a 180MHz Pentium would definitely be slower than a 166MHz
     (8.4% faster CPU, 10% slower bus).

-----------------------------------------------------------------------------
  Michael L. VanLoon                           michaelv@MindBender.serv.net
        --<  Free your mind and your machine -- NetBSD free un*x  >--
    NetBSD working ports: 386+PC, Mac 68k, Amiga, Atari 68k, HP300, Sun3,
        Sun4/4c/4m, DEC MIPS, DEC Alpha, PC532, VAX, MVME68k, arm32...
    NetBSD ports in progress: PICA, others...
-----------------------------------------------------------------------------