- Joined
- Apr 15, 2012
- Location
- Fort Smith, AR
Hi, guys,
I have my eye on a new Sandy Bridge (or Ivy Bridge) system, but it's going to be a while. In the meantime, I sure could use some help understanding my present system a little better. I've always had problem overclocking this thing and it wound up in the basement due to instability. Now I really need this computer, and after reading all the old news, I believe the problems were due to my ignorance of how to properly configure a smoothly tuned overclock.
A few years back, I assembled a system for the Pentium Extreme Presler. My initial attempts were riddled with problems, and I soon discovered I had underrated my power supply, after it failed on me. I moved up to a 650 W supply (purchased from Best Buy), and was able to create a pseudo-stable system, which I ran for about 2 years. I mean "pseudo" because it never would run Prime 95 for very long without crashing or freezing, but seemed fine on all other tests. Then it started kicking off again. I assumed it was a problem with the cooling block contact (dried out thermal grease), but my attempts to fix it failed. So in the basement it went.
I pulled it out 2 weeks ago, disassembled it, cleaned out everything and reassembled, w/ new BIOS chip, etc. I discovered my primary problems were, once again, the power supply - a BFI. So I ordered a Rosewill Hive & now have it running very stable on default clock settings, but when I attempt to overclock above 11%, I get the same problems running Prime95 - the screen locks up after 5-10 minutes.
I have backed up and come at this slowly - trying to get the FSB tanked up first, before increasing the multiplier. I get a stable system at FSB clocked to 298 MHz (default setting is 267), but anything over causes this problem, even after bumping the memory voltage up 3 places and the vcore up twice.
Can anyone please shed some light on this?
Thank you very much.
Here are my detailed settings:
BASIC CPU SETTINGS:
Multiplier = 13 (default value)
FSB setting = 302 (default is 267)
FSB resultant speed = 1209 MHz
CPU speed is therefore = 3926 MHz, only 13.11% OC
memory voltage = 2.05 (has been bumped twice from 1.95, default was 1.80)
vcore = 1.30 (default was 1.2875v)
MCH chipset = 1.50, default (movement to 1.55 had no effect)
ICH chipset = auto
CPU temp while running Prime95 = 38C
FSB termination voltage = auto
DRAM frequency = auto
Performance mode = auto
PCI ex freq = auto
PCI clk sync mode = auto
Configure DRAM timing by SPD = enabled
Hyper path 3 = auto
DRAM throttling threshold = auto
OTHER:
Power supply = Rosewill 650W (brand new)
Cooling system is Swiftech water cooling, CPU and N bridge wtrblks & 2-fan radiator
Highest stable overclock thus far is at FSB = 298 - runs Prime95 at length:
RAM voltage unknown (auto),
VCore set to Auto, but reads 1.23v on AiBooster
Chipset = 1.50
The highest stable OC thus far is at FSB=298:
RAM voltage set to auto (unknown)
Vcore set to auto (reads 1.26v in AiBooster)
Chipset set to 1.50v (default)
I have my eye on a new Sandy Bridge (or Ivy Bridge) system, but it's going to be a while. In the meantime, I sure could use some help understanding my present system a little better. I've always had problem overclocking this thing and it wound up in the basement due to instability. Now I really need this computer, and after reading all the old news, I believe the problems were due to my ignorance of how to properly configure a smoothly tuned overclock.
A few years back, I assembled a system for the Pentium Extreme Presler. My initial attempts were riddled with problems, and I soon discovered I had underrated my power supply, after it failed on me. I moved up to a 650 W supply (purchased from Best Buy), and was able to create a pseudo-stable system, which I ran for about 2 years. I mean "pseudo" because it never would run Prime 95 for very long without crashing or freezing, but seemed fine on all other tests. Then it started kicking off again. I assumed it was a problem with the cooling block contact (dried out thermal grease), but my attempts to fix it failed. So in the basement it went.
I pulled it out 2 weeks ago, disassembled it, cleaned out everything and reassembled, w/ new BIOS chip, etc. I discovered my primary problems were, once again, the power supply - a BFI. So I ordered a Rosewill Hive & now have it running very stable on default clock settings, but when I attempt to overclock above 11%, I get the same problems running Prime95 - the screen locks up after 5-10 minutes.
I have backed up and come at this slowly - trying to get the FSB tanked up first, before increasing the multiplier. I get a stable system at FSB clocked to 298 MHz (default setting is 267), but anything over causes this problem, even after bumping the memory voltage up 3 places and the vcore up twice.
Can anyone please shed some light on this?
Thank you very much.
Here are my detailed settings:
BASIC CPU SETTINGS:
Multiplier = 13 (default value)
FSB setting = 302 (default is 267)
FSB resultant speed = 1209 MHz
CPU speed is therefore = 3926 MHz, only 13.11% OC
memory voltage = 2.05 (has been bumped twice from 1.95, default was 1.80)
vcore = 1.30 (default was 1.2875v)
MCH chipset = 1.50, default (movement to 1.55 had no effect)
ICH chipset = auto
CPU temp while running Prime95 = 38C
FSB termination voltage = auto
DRAM frequency = auto
Performance mode = auto
PCI ex freq = auto
PCI clk sync mode = auto
Configure DRAM timing by SPD = enabled
Hyper path 3 = auto
DRAM throttling threshold = auto
OTHER:
Power supply = Rosewill 650W (brand new)
Cooling system is Swiftech water cooling, CPU and N bridge wtrblks & 2-fan radiator
Highest stable overclock thus far is at FSB = 298 - runs Prime95 at length:
RAM voltage unknown (auto),
VCore set to Auto, but reads 1.23v on AiBooster
Chipset = 1.50
The highest stable OC thus far is at FSB=298:
RAM voltage set to auto (unknown)
Vcore set to auto (reads 1.26v in AiBooster)
Chipset set to 1.50v (default)
Last edited: