PDA

View Full Version : Please kindly help with P4 3.0C OC on IC7-MAX3


L33T
05-15-2005, 08:55 PM
Hey,

I'd like as much help as possible with the last stretch of OC'g on my IC7-MAX3 system running a P4 3.0C. This is a rather long post as it contains a lot of empirical data, but I would ask that you please read through it and offer whatever it is that you may have to offer. Thank you in advance.

Thus far, I have a relatively stable checkpoint at fsb 247 (interpreted in CPU-Z as fsb 247.6 MHz). With this, I get a nominal 3.7 GHz.

This validation page should help: http://valid.x86-secret.com/show_oc?id=17890

Summarised here:
P4 3.0C @ 3.71
VCC = 1.600 or 1.616 or 1.58
RAM: Kingston 2700 ValueRAM 512mb @ 198.1 MHz (5:4) @ 2.5/3/3/7
PSU: Antec TrueBlue 480W (about new for two months)

Temperatures
(By load, that refers to the temperature when running F@h; by idle, that refers to the temperature monitored in WinXP)

CPU
T[load] = 48.5*C
T[idle] = 40.5*C

System
T[load] = 30.0*C
T[idle] = 30.0*C

PWM
T[load] = 45.5*C
T[idle] = 37.0*C


I tried performing the following OC's (rather systematically). I began and performed the OC's in the order which I have listed here. Please let me explain the nomenclature with a simple table:

Nomenclature:
P - Posts
V - Verifies (Just before you boot into the OS, with the Windows flag, you get a message that says, ...verified...)
WF - Windows flag (loading screen for WinXP)
L - Login screen for WinXP
ENV - booted into the WinXP environment successfully

fsb 267

VCC; tests performed @ AGP=1.55V unless otherwise specified
1.7250 - Freeze
1.7000 - P,V; fails just after verification but before display of Winflag
1.6750 - P,V, WF; freezes @ WF; @AGP=1.65, almost gets the login screen but not quite; @AGP=1.60, freezes on WF
1.6500 - P,V,WF,L; freezes @ L; @AGP=1.65, almost boots; @AGP=1.60, crashes & restarts instantly
1.6250 - P,V,WF; freezes @ WF; @AGP=1.65, L screen displays
1.6000 - P,V; restarts at V
1.5750 - P,V; restarts at V
1.5500 - P,V; restarts at V

fsb 266
1.6750 - P,V,WF, restarts @ WF
1.6500 - P,V,WF, restarts @ WF; @AGP=1.65, mouse pointer for login screen is seen, but then dies
1.6250 - P,V,WF, same as "1.6500" case
1.6000 - P,V,WF, restarts @ WF

fsb 268
1.7000 - P,V; restarts @ V
1.6750 - P,V; freezes @ WF
1.6500 - P,V, restarts @ WF
1.6250 - P,V, restarts @ WF
1.6000 - P,V, freezes @ WF; @AGP=1.60, restarts at WF; @AGP=1.65, restarts, but no WF

fsb 265
Note: Here I decided that the higher AGP seemed to make the OC's more stable. This seems to be backed up by members over at Xtreme.
(*) 1.6500 - Mouse ptr is shown, login screen seen, can type PWD, but it doesn't actually login
1.6250 - restarts @ WF
1.6000 - restarts

(* Best try; logged in successfully; then restarted after ~11 secs)

fsb 264
(Same things as above...)
1.6500 - restarts after a while, but does boot into WinXP

fsb 263
1.7000 - fails before WF
1.6750 - logged in; lasted 5 secs and then there was this error msg dialog box with a big red circle w/ a white 'x'; restarted the computer and tried again with same setting, and this time it lasted 13 secs in WinXP until it froze
1.6500 - Froze after 43 secs upon entering ENV
1.6250 - restart @ WF

fsb 262
1.7000 - fails before WF
1.6750 - WF; login for 4 secs, then restarted
1.6500 - login for 15 secs, then froze
1.6250 - login for 6 secs, then froze
1.6000 - strange error msg; second try and I get login for 4 secs then restart
1.5750 - restart instantly

fsb 261
1.6750 - login for 42 secs then froze
1.6500 - login for 26 secs then restart
1.6250 - login for over 1 min!!! but then died in disgrace
1.6000 - login for 40 secs then restart

fsb 260
1.6500 - semi-stable! it lasted an entire day, but with nothing using CPU...Trillian3pro still worked, but then we turned on F@h, and it died in disgrace

At this point I also tested some other RAM combinations...
3:2 Divider, fsb 260 --> 3.9GHz CPU & 173.3MHz RAM
5:4 Divider, fsb 260 --> 3.9GHz CPU & 208MHz RAM (perhaps stable?)

It restarted. So I set timings to 2.5/3/3/7

After like 15 mins it restart.

So I set Vdimm = 1.9V... still didn't work

I decreased Vdimm = 1.8V, but then increased all timing elements +1... didn't work

Kept that last result, but returned Vdimm to 1.9V.



That is all the information I have. If you have read this far, for that alone I thank you for taking your time. I know this is quite a lot to go through, and there are simply so many things, but I hope you'll take a few minutes and give your thoughts.

Thank you.

saaya
05-15-2005, 09:29 PM
1. welcome to XtremeSystems! :D :toast:

2. please post in the right section in future, this section is for xtreme overclocks and results only :)

3. "relatively stable" ? so is the cpu stable at stock at all? i only see settings at wich its unstable :confused: so at what sped are you stable and what is the problem? it doesnt oc any higher no matter what vcore bump you give it?

L33T
05-15-2005, 09:45 PM
Yep thanks for the welcome! Sry about the placement...I wasn't sure where to put requests for help.

There are several completely stable checkpoints (presumeably Prime95 would have ran indefinitely in either case):
fsb 243 (CPU: 3.65 GHz) -- Prime95, 13 hrs stable
fsb 234 -- Prime95, 24 hrs stable

saaya
05-15-2005, 10:08 PM
so what exactly is the issue?

L33T
05-15-2005, 10:25 PM
At those fsb freq's the computer is totally stable, but at these higher fsb's, the computer isn't stable. I'm trying to do everything in my power to make the computer hit 4.0GHz.

Regarding the VCORE question on higher voltages: There seems to be a pattern emerging (or several possible patterns, depending on how you view the whole thing):

One possible pattern is that the VCORE doesn't do well at 1.725V & above (No Post, regardless of fsb).

Another possible pattern is that the VCORE doesn't do well when it is too much above the "required" or "minimum" VCORE for a certain fsb.