Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 19 Jun 1997 11:51:03 -0500 (CDT)
From:      Craig Johnston <craig@gnofn.org>
To:        hardware@freebsd.org
Subject:   K5-PR166/IT5H FreeBSD overclock
Message-ID:  <Pine.GSO.3.95.970619111531.20332A-100000@sparkie.gnofn.org>

next in thread | raw e-mail | index | archive | help

Have successfully overclocked this chip from 117MHz to 131MHz under
FreeBSD by raising bus speed on my ABIT IT5H to 75MHz.  The mobo revision
is 1.5, the K5 is an ABR stepping, the memory is 32 megs of 50ns EDO DRAM,
and the cache is 512K 12 ns pipeline burst.  Now the mobo claims a
K5-PR200 is installed.

All the fastest chipset setting were used.  I tended to segfault on big
compiles until I raised the chip's core voltage to the 3.60V setting
included on this mobo.  Now it runs rock-steady so far, through a large
number of make worlds.  An intel heatsink/fan combo that seems halfway
decent was used with thermal compound, and becomes only slightly warm to
the touch.  Voltage regulator sinks on the mobo become hot to the touch
with case off during make world -- would be cooler with case on, I'm
getting no airflow. 

Kernel is 2.2-STABLE, cvsupped about a week or so ago.  It does a
make of GENERIC in 294 secs user + system time, with gcc -0 -pipe.
It makes my production kernel in just over 4 minutes.  :)

I'm extremely satisfied with the performance of this combination.
The CPU was $110 and the motherboard was $140, and I seem to have
something on the order of 4x the CPU performance of my old 486dx4-100.
The motherboard will accept the K6 processor, which I'll move up to
when they're $150 or so.  :)

Anyway, the secret to making it work for me was the voltage increase,
which I wanted to note here for those who may have failed to make this
overclock work.  3.52V bad, 3.60V good, so far. 
 
regards,
Craig




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.GSO.3.95.970619111531.20332A-100000>