[rc5] Overcolocking, just once more

jpearl at qsi.com jpearl at qsi.com
Fri Jul 18 23:54:58 EDT 1997





>>Date: Mon, 14 Jul 1997 20:56:09 -0400
>>From: jpearl at qsi.com
>>Subject: [rc5] Dangers of overclocking
>>
>>In response to:
>>
>>Date: Sat, 12 Jul 1997 02:14:25 +0100
>>From: Rebecca and Rowland <rebecca at astrid.u-net.com>
>>Subject: [rc5] Dangers of overclocking
>>
>>>>The problem is that the CPU runs "just fine" for weeks or even months
>>>>straight, but someday the room temperature will rise up just enough
>>>>to make it fetch a bad byte from memory.
>>>>
>>>*If* the problem is due simply to overheating the CPU, you can get round
>>it
>>>with a bigger heatsink and fan.

>>>I suspect it's more likely due to running the entire system beyond its
>>>design limits, thus increasing the chance of errors - not necessarily
due
>>>to overheating.
>>
>>The whole system huh? Last I checked most any PC MB being built was made
to
>>do at least 66mhz on the BUSS and I don't believe Intel's 200mhz machines
>>receive a higher clock than that - yet. In my case the board was designed
>>to go as high as 83mhz, a shame my EDO memory can't go that high (grin).
My
>>P150 has been running at 187mhz for months now just fine and crunches
keys
>>with the rest.
>As I understand it, you increase the bus speed.  If you have, say, a 50MHz
>bus and a 150MHz clock-tripled processor, you might increase the processor
>speed to 200MHz by increasing the bus speed to 67MHz.
Yep, completely true! A 150mhz chip won't go much -past 187mhz though. The
best way to go with a 150mhz chip for those that are nervous about
overclocking is to simply run it at 2X75mhz. It still runs 150mhz but the
rest of the system is MUCH faster.

>As I said, the whole system (meaning: everything that takes its timing
from
>the system bus clock).
>Rowland.

Just once more - yes I do increase my entire system speed, so what? You
said ">>>I suspect it's more likely due to running the entire system beyond
its design limit". You implied that by increasing my BUSS speed I was
stressing my system beyond it's limits not simpy that everything takes it's
cue from the BUSS.

 In the case of an X86 type MB, the system is DESIGNED from the word go
with 66MHZ in mind and in some cases as much as 83MHZ. My stepping up to
those speeds in no way stresses any part of the system (with the possible
exception of the CPU) beyond it's "design limits". It is true that 75mhz
and 83mhz are not what Intel intended when they laid down the specs but
that doesn't mean the engineer who designed my boards didn't design it to
go that fast. The only part of the system that might be stressed is the
CPU, thus far out opf about 10 overclocked systems I've yet to have a CPU
go south. Of all those systems, 3 are currently being used for this contest
and none have locked up due to a processor "failure". A good heatsink, a
good fan, and some common sense are all it takes to overclock a system
without fear of failure.

If this doesn't make sense to you, point your browser to sysdoc.pair.com,
read a bit, and perhaps you'll understand where we're coming from on this.
I grow weary....

jvp at qsi.com


----
To unsubscribe, send email to majordomo at llamas.net with 'unsubscribe rc5' in the body.



More information about the rc5 mailing list