New rig is built...oc question

[Dunadan]

OK...the new 2.4C rig is built. Went pretty smooth. A couple questions -

1) When I am ocing my FSB....changing it also changes (slightly) the AGP/PCI frequency. I have the AGP/PCI frequency fixed, but up above where it calcs the CPU/AGP/PCI I see the AGP/PCI change. For instance...I ocd my FSB to 250, and I believe the AGP/PCI went to something like 70/35. Is this ok? I thought that was supposed to stay fixed at 66/33? Or is it just the ratio that is locked?

2) When running Prime95...do you just start it up and let it run? Is there a problem starting it and then just leaving your pc?

Thanks for the help.

 

[Actaeon]

Originally posted by: Dunadan
OK...the new 2.4C rig is built. Went pretty smooth. A couple questions -

1) When I am ocing my FSB....changing it also changes (slightly) the AGP/PCI frequency. I have the AGP/PCI frequency fixed, but up above where it calcs the CPU/AGP/PCI I see the AGP/PCI change. For instance...I ocd my FSB to 250, and I believe the AGP/PCI went to something like 70/35. Is this ok? I thought that was supposed to stay fixed at 66/33? Or is it just the ratio that is locked?

2) When running Prime95...do you just start it up and let it run? Is there a problem starting it and then just leaving your pc?

Thanks for the help.

1.) the AGP/PCI should be locked. I don't know about the BIOs settings in P4 boards, but look for some BIOs setting.

2.) No problem with leaving Prime95 on overnight or whatever.

 

[Actaeon]

doublepost, sorry.

 

[Dunadan]

Oh well...tried to run 3dmark2001 after ocing my FSB to 250 and it crashed. I changed my ration to 5:4 so ram was running at stock. Thought 250 would be a slam dunk. Any pointers for first time ocer?

 

[dunkster]

When using only Prime95 for stability testing, failure does not resolve whether the problem is CPU or memory instability.

When pushing FSB, I test first with Memtest86 Test#5, adjusting timings/Vdimm as necessary for several error-free passes. Test#5 (block moves) seems to be the most critical test, and it runs quickly. Several passes takes only a few minutes.

After establishing memory stability, move on to Prime95 testing, adjusting Vcore as necessary for CPU stability.

Hope this helps!

 

[Mav3N]

seeing that u have a nearly identicle setup as myself

just manufacturer of ram and such
the memory settings are critical to the overclock

relaxing them some and making sure ur runing in duel channel mode is also essential

u should be able to hit 250 with 5:4 multiplier np
but u may need to kick up the ram voltage as well as the cpu voltage

1.575 should be fine for cpu and there is no need ever to run it over 1.6
2.7 or 2.8 volts on the memory will help you as well

there seems to be a max limit on the abit is7 board for the overclock i havent seen anyone get over 265 fsb on it succesfully no matter the voltage or the divider settings

 

[joe2004]

Since 2.4C is HT processor, Windows will open only one thread and use only 50% of utilization in case of Prime95. You need to start two instances of Prime95 to get the full load on the processor. To do that you need to start the second instance of Prime95 from a command line with prime95 -a1.
The newest "unstable" version 2.37 of Prime95 might crash in case of the third group of tests while using "blend" option. Meaning that application crashes rather than the CPU. Use older stable version of Prime95.
The changes on AGP/PCI clock are minor and will not affect the performance in any noticable way, don't worry about that.