Overclocking mit nForce 2 Overclocking in the Paxis - whow to do [Back to OC Guide Pacge 1] [Back to OC Guide Page 2] [Back to OC Guide Page 3]

Continuation - Multiple CPU Overclocking 

Area 1 had the practice CPU FSB Overclocking, now here it is about pure Overcklocking
about the processor CPU Multiplier. Bios name often "Clock Ratio"


All drafts are demanded here where itself the rise CPU FSB is not advisable because,
e.g., memory is not suitable for overclocking. 

Also a combination FSB and multiple rise is possible. 

Some Mainboards support only a maximum multipler rise on 12.5 
Why ?
 


With the multiplier CPU rise we also go forward in small steps. 
That is as a next step Multi 0.5 more.  
With a multiplier 10 CPU like an AMD XP3000+ , e.g. 10.5

We carry out the setting in the Bios, together with the rise of the CPU vcore on the
next possible value. 
Bios save and exit, then return on Windows level. 

Now our Processor CPU is first time overclocked. If we test immediately whether the 
thing also system stable runs. In addition we start the program "Super PI " and 
execute one 1M calculation as a test.
Choose Panel ''Calculate''  ''1M'' and click on ''ok''  to start a calculation

In 19 Loops are tested calculation which all should wear through faultlessly.
If the test breaks before the 19ten Loop from our system does not run faultlessly.
All Loops run through then our picture looks as follows:

Click on the picture to enlarge this Message: ''PI calculation is done'' is fine Avoids work of background programs in run are. This does the calculation only more slowly.

If the test was successfully one does in the Bios the next step.
Raise CPU Multiplier again by 0.5.
Once more Bios save and exit, then return on Windows level.

Afterwards again Super PI calculation run, ...
So long the overclocking runs perfectly simply the interrelation Bios Multi rise and
system test continued.

What is if we have Super PI Calculation Errors ? 
The CPU does not run faultlessly.
Here must be raised up in the Bios the processor CPU vcore.
At first time the rise of CPU vcore is still no problem. 
In the course of overclocking one is confronted perhaps vcore tension dubious
and for the CPU can be also unhealthy.
I publish here no clear rules which CPU vcore is still considered as relatively safe!
The empirical value says: 1.825 to max. 1.85 volts vcore, with a fine CPU cooling.
One should not raise a mobile XP CPU with default 1.45 vcore so far.

With the gradual overclocking named at the top one gets when the CPU more 
vcore tension needs.
In the course of overclocking one will get the CPU already for another Multi 0.5
step more vcore requires than before. 
A clear signal itself the top performance CPU to the end approaches.
With the pure multiplier CPU overclocking it can happen surely quickly.

TheSuper PI 1M calculattion is a surely fast calculation and no extreme test
on system stability. We should not rely 100% on this test !
What should follow a longer more intense stress is a test with Super PI or 
with Prime. Download of Prime Prime 2004 v.025 
With Super PI do an 32M calculation for a final test.

Super PI or Prime, both needs time. Does not do the test if to her the 
PC needs.
One starts the program Prime as follows:  [for screenshot]

An alternative or additional system stess test is the program SI Sandra. [Link] 

If the programs determine mistake, you take back the Overclocking minimal
and tests again. At the end you get a solidly running system ! 

* caught in the Frameset. Please click on: [Home] *

[Back to OC Guide Pacge 1] [Back to OC Guide Page 2] [Back to OC Guide Page 3] [Back to Main Page] [Home]