Date: Mon, 24 Nov 1997 14:26:26 +0800 (WST) From: Terry Dwyer 61 8 9491 5161 <tdwyer@io.telstra.com.au> To: Burton Sampley <bsampley@bsampley.vip.best.com> Cc: sporkl@dti.net, freebsd questions <freebsd-questions@freefall.FreeBSD.org> Subject: Re: Upclocking Message-ID: <Pine.BSF.3.91.971124131906.19645A-100000@io.telecom.com.au> In-Reply-To: <Pine.BSF.3.96.971123193005.16166D-100000@bsampley>
next in thread | previous in thread | raw e-mail | index | archive | help
Spike and Burton,
I have to disagree with some of what Burton says about overclocking
A couple of definitions:
I'm not trying to be pedantic here, but I've never heard the term
"upclocking". I'm just trying to clarify what I'm talking about.
Processor overclocking By increasing the processor multiplier it
is possible (with some P5's and 6's) to
run the processor at a clock speed higher
than it's box label says it should go.
Bus overclocking Increasing the bus clock above the setting
normal for that Processor has the effect of
overclocking the bus _and_ the processor
at the same time. Big win!
Tom's hardware page explaines it so much better than the couple of
paragraphs above.
Intel produced a hell of a lot of Pentium class processors that
were not able to be Processor overclocked. Bus overclocking is
the only avenue open to people with that sort of processor.
Generally these were "Boxed Set" processors - the after market
type.
On Sun, 23 Nov 1997, Burton Sampley wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
>
> Spike,
>
> Check out Tom's Hardware Guide at: http://www.sysdoc.pair.com/
>
>
> You'll find most the info you need about overclocking your CPU. I'm
> currently overclocking my P5-233MMX to 266 MHz on an Asus P/I P55T2P4
> motherboard by setting the bus speed to 75 MHz instead of 66 MHz that
> Intel recommends. This motherboard has an undocumented (well, at least by
> the manufacturer) bus speed of 83 MHz.
I've seen at least one manual for the P55T2P4 that _does_ document the
83MHz bus clock settings, but thats by the way. 8-)
>
> The major draw back to overclocking is heat and the accelerated bus speed
> causes some PCI cards to not work. I was able to bring my system up at 83
> MHz, but my SCSI controller (an old 2940) would work correctly. The above
> link gives you a ton of info.
I have two machines at home - both P55T2P4's, both Hardware Rev 3.0 - not
the recommended 3.1 Hardware revision of the motherboard.
In the FreeBSD box I have a P133 processor bus overclocked to 166MHz and
an Adaptec 2940UW that goes like a bat out of hell at this bus speed.
Thank you Intel!
This machine has 64MB of RAM - 70ns fast page with _no_ memory timing
backoffs. 8-)
In the Win95 box, (no flames please! 8-) ), I have a P166MMX processor bus
overclocked to 210MHz and an Adaptec 2940 BIOS Rev 1.21.
The T2P4's BIOS tells me it's only found a 200MHz processor, but that's
because the BIOS rev (2.5 I think) doesn't support processors above 200MHz.
In fact it _is_ running at 210MHz. BIOS rev 2.6 fixes the display
problem but I haven't applied it.
This machine has 32MB of 60ns EDO RAM with no memory timing backoffs.
Heat is not a problem for either of these processors, although I don't
know what a P233MMX behaves at 83MHz.
I've been running these machines for > 6 months now with only one
noticable problem. Some sound programs don't work correctly at the higher
bus speed. This may be because there is a software timing issue that is
sensitive to bus speed. It only happens on the FreeBSD box in "amp" (of
winamp fame). From memory, most other audio apps are fine on the unix and
Win95 machines.
Processor temperature
If I touch either of the processors - I was concerned about heat too -
they are slightly warm to the touch, but certainly _not_ hot. My
definition of a hot processor is "unpleasantly warm", I am conservative!
A couple of tips
Cache RAM
If you have less than 64MB of system memory and you want to bus overclock
your machine to 83 MHz, make _sure_ you pull out any aftermarket add-in
cache RAM module. Only trust the cache RAM on the Motherboard or you will
more than likely only achieve 75MHz bus speed. The reason? ASUS, (the
Good Guys) put _fast_ cache ram (6 or 8ns I think) on the motherboard,
they had to for the 83MHz bus speed didn't they 8-) Any Cache module you
plugged in afterwards probably uses 12-15ns RAM which is far too slow for
83MHz.
System RAM
If you have good 70ns Fast Page RAM and you're currently running a 66MHz
bus machine, there's an _extremely_ good chance you can overclock the bus
to 83MHz. Simply set the memory speed timings to defaults in the
appropriate PCI config page and give it a try. The BIOS has a feature to
reset all of a page's parameters to default. I've bus overclocked
probably a dozen or more 'T2P4's with a mixture of RAM types, mostly 70ns
Fast Page RAM and I've been sucsessful _every_ time.
SCSI Controllers
I can't say anything about controllers other than Adaptec, except that NCR
controllers seem to just work without problems from everything I've read.
Adaptec controllers however, _must_ have a BIOS revision of 1.21 or later
or you're out of luck. I had one of the early cards, so I upgraded to a
new 2940UW (best value for a server with lots of disk) and the old 2940
BIOS rev 1.16 is now in my 14YO son's P100 FreeBSD box.
Card Problems
I don't have any problems with Video cards (1 X S3 Virge and 1 S3 Trio 64V+)
ISA cards behave well in both machines 1 X AST 4 port serial card, I X
GUS audio card, 1 X SB16 non-PNP audio card.
In general, bus overclock and you should be very happy with the results.
Sorry for the long posting. I did try to be brief and still provide
enough info to help people.
>
> Hope this helps.
>
> - - burton -
>
> On Sun, 23 Nov 1997, sporkl wrote:
>
> > Hello.
> >
> > I have heard mention pf "upclocking" a chip to a higher speed. Is
> > this what it seems to be, trying to trun a foo speed chip att foo+bar MHZ?
> > How is this accomplished, via hardware or software? What are the drawbacks
> > or side effects? thank you.
> >
> >
> >
> > -Spike Gronim
> > sporkl@dti.net
> >
> > "Tradition is the chastity belt of the mind"
> >
> >
> >
> >
>
> - ---------------
>
> Burton Sampley
> bsampley@best.com or bsampley@haywire.csuhayward.edu
> PGP key available at http://www.best.com/~bsampley/pgp.html
>
>
> -----BEGIN PGP SIGNATURE-----
> Version: 2.6.2
>
> iQCVAwUBNHj2nXt2O8KJtMdBAQFETAQAu0LrG62W9VkaEesXUKnIu678w1UDbuOc
> Sg7BKD4DTT5oigL02QYReP8Zq0MHKNO6yWAvAYKMpJg8Q25J63W+Ko8aDZEvbREO
> gC0zUzfOeNQIlQCDr8Lh51JdjncAn4vt4alqZzjerRhZ8+blBISZKUNg6njgUzux
> DWXjfB7Pa+I=
> =VOdC
> -----END PGP SIGNATURE-----
>
>
_-_|\ Terry Dwyer E-Mail Work: tdwyer@io.telecom.com.au
/ \ System Administrator Private: tdwyer@omen.com.au
*_.^\_/ Telstra Corporation Phone: +61 8 9491 5161 Fax: +61 8 9221 2631
v Perth Western Australia (I do not speak for Telstra or Telecom)
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.3.91.971124131906.19645A-100000>