Newbie overclock

[Garyj881]

Ok i've had ago of over clocking my 3.2gh ht skt 478 today.

I got it up to 3936 on 1.52 volts and my memory running at ddr 320 as it wouldn't boot on standard ddr400.

Unfortunately it wasn't stable so i knocked the fsb down from 246 to 238.

Now it booted and run windows stable i decided to try orthos with the cpu at 3808 and two minutes into it i got an error.

Any ideas what I'm doing wrong or am i just hitting my limit?

Gary

 

[mrthingyx]

You're overclocking in reverse.

You should start with a modest increase in FSB (not multiplier) and run Prime/Orthos for eight hours at each overclock. Keep on going until the test fails, then slightly increase the CPU voltage.

There's nothing new to add that isn't in the guide at the top of this forum.


And remember, just because it posts/boots into Windows doesn't mean it will even last through opening IE or whatever.

 

[Garyj881]

You're overclocking in reverse.

You should start with a modest increase in FSB (not multiplier) and run Prime/Orthos for eight hours at each overclock. Keep on going until the test fails, then slightly increase the CPU voltage.

There's nothing new to add that isn't in the guide at the top of this forum.


And remember, just because it posts/boots into Windows doesn't mean it will even last through opening IE or whatever.

Sorry i meant FSB not multiplier, ive corrected it now.

Whats the max voltage i should increase to?

 

[furnace]

Sorry I know this is a little off topic; but when I get my new rig, I wanted to overclock that day and finish that day, if not then the next day... but after each increment I'm going to have to wait 8 hours for testing? I thought you just did an 8 hour test when a 10 minute one was OK and you had reached a speed you're happy with..? Like, just to ensure that it's stable.

 

[Garyj881]

Sorry I know this is a little off topic; but when I get my new rig, I wanted to overclock that day and finish that day, if not then the next day... but after each increment I'm going to have to wait 8 hours for testing? I thought you just did an 8 hour test when a 10 minute one was OK and you had reached a speed you're happy with..? Like, just to ensure that it's stable.

Well im going to give it a shot, i'm starting at 3600 mhz and im 10 mins in with no errors.

No pain no gain

 

[mrthingyx]

Sorry I know this is a little off topic; but when I get my new rig, I wanted to overclock that day and finish that day, if not then the next day... but after each increment I'm going to have to wait 8 hours for testing? I thought you just did an 8 hour test when a 10 minute one was OK and you had reached a speed you're happy with..? Like, just to ensure that it's stable.

Of course - stability is entirely up to the individual user and Orthos is only one tool of many.

BUT, Orthos will not just fail before ten minutes or after eight hours and going on the time interval between, that is a HUGE amount of time for an error to occur. At what point has the overclock become unstable in those 7hrs 50mins of testing? How far back should the CPU be wound as a result? What should be increased?

Your kind of test will work, but it in the end there is no guarantee that the same stable overclock will be achieved as by using the 8hr/each time.

I'm not sure what the maximum voltage for the P4Cs was/is... basically, keep going until the load temperature goes above 65oC or becomes unstable. And you nkow what the only way to make sure your PC reaches load equilibrium? With extended testing!

 

[Garyj881]

Right I've done 8hrs stable at 3.6ghz.

I will try it at 3.7 now

Will stress testing my cpu improve its performance?

 

[furnace]

It won't improve its performance, but it will tell you whether or not your computer is stable So it's an important thing to do if you don't want your computer crashing due to over clocking

 

[Garyj881]

Update: I got an error after 10 mins at 3800 mhz so i upped the v core to 1.5 but on full load i was reaching temps up to 62deg is that too high?