q6600 Small FFT's Epic Fail

justinbkerr

Distinguished
May 11, 2008
5
0
18,510
Okay I have a Q6600 OC'ed from 2.4 ghz to 3.4ghz using a Golden Orb II air cooler on a XFX 680i LT motherboard.
My FSB = 378 x 9 / Voltage = 1.425 in Bios (CPUz reads 1.360?)

1.) First question is why does CPUz read a voltage different from the setting in my bios?
2.) Second Question using Realtemp, the processor idles at 45-50c and goes up to 68-72c under prime95 running a blend or large FFT torture test.
HOWEVER
When I run the small FFT's torture test I often get errors in core 4 and temps skyrocket to 80c after only a short period.
I have seen this at lower overclocks as well but fixed it by bumping up the voltage. But with the small FFT's test already pushing the processor to 80c I don't wanna add any more juice.

The systems is stable both in games and normal use and the blend and large FFT tests are no problem.
Would this be considered a safe overclock and safe temps?
Or, should I back it off? I had a stable overclock @ 3.2ghz with 1.225v (almost .2v less power!) and processor ran in the 65c range under full load.
I want performance, but I don't want to burnout my CPU.

 

Lupiron

Distinguished
Feb 9, 2008
1,711
0
19,780
Okay, first, use Core Temp, or Real Temp to list the VID of your chip.

CPUz reflects VDrop. That is the initial drop from Bios VCore to Idle windows.

Then there is VDroop. The drop while running the Small fft test!

You prolly need to re seat the cooler. Because your temps are way high, even for that speed!

That would be my number 1 recommendation!

Make sure the paste is even!

Make sure it makes full contact with the IHS of the CPU.

You can test this by cleaning them both, and ONLY placing the paste on one or the other. Then setting it, mounting it for 1 min, then removing it straight up to see that the heat paste is now on both items evenly!

See ya soon!

--Lupi
 

kad

Distinguished
Feb 29, 2008
524
0
18,980
Temps under full load must not be more than 65
That Orb HS is rubbish, get a decent one with good fan
Get good TIM such as OCZ Freeze Extreme
Make sure you have full contact between HS and proc
Your OC is considered stable when you run Prime95 for 24 hours with no error and Max. temp 65 using RealTemp 2.6
Run your system with case open and see if temps are better which means you have poor air flow inside your caseand must be improved
 

mtc1

Distinguished
Jul 5, 2008
14
0
18,510
Robust Graphics Booster___________ [Auto]
CPU Clock Ratio (Note)____________ [9x]
CPU Host Clock Control_ [Enabled]
CPU Host Frequency (MHz)__________ [370]
PCI Express Frequency (Mhz)_______ [102]
C.I.A. 2__________________________ [Disabled]
Performance Enhance__________ [Standard]
System Memory Multiplier (SPD)____ [2.00]
Memory Frequency (Mhz) 740
DRAM Timing Selectable_______ SPD __ [Manual]

CAS Latency Time______________ [6]
Dram RAS# to CAS# Delay_______ [7] <--use your rams values (tRCD)
Dram RAS# Precharge Delay______ [7] <--use your rams values (tRP)
Precharge Delay (tRAS)__________[24] <--use your rams values (tRAS)
ACT to ACT Delay (tRRD)_________[auto]
Rank Write to READ Delay_________[auto]
Write to Precharge Delay_________[auto]
Refresh to ACT Delay____________[auto]
Read to Precharge Delay_________[auto]
Static tRead value______________[auto]
Static tRead adjust______________[auto]

System Voltage Control____ [Manual]
DDR2 OverVoltage Control__ [+0.5V]
PCI-E OverVoltage Control_ [+0.1V]
FSB OverVoltage Control___ [+0.2V]
(G)MCH OverVoltage Control [+0.2V]
CPU Voltage Control_______ [1.49375v]


vcore

bios: 1.460v
idle: 1.456v
load: 1.408v


Temps

idle: 34c
load:63c (small fft)
load 55c (Large FFT)




I too am getting crazy results with SMALL FFT. Right now I am holding steady at 3.33GHz, but if I go any higher, I crash on small fft. I can run Blend or Large FFT without a prob. Small FFT kills me. Seems like it is stable under normal circumstances. Small FFT doesn't seem like normal circumstances. Am I wrong?
 

Evilonigiri

Splendid
Jun 8, 2007
4,381
0
22,780
Small FFT stresses the CPU, so normally an error here would indicate CPU instability. However, looking at your high vcore, your CPU should be stable. Since it's not, it could be an FSB hole or ram instability.

If it's an FSB hole, just bump the FSB up to 400MHz and see what happens.
 

mtc1

Distinguished
Jul 5, 2008
14
0
18,510





Good call.


I will try that out.


Thanks for the input. I will let you know how it goes.
 

mtc1

Distinguished
Jul 5, 2008
14
0
18,510
400 x 9 with the above settings the same will not boot into windows.




:-(




I will try coming down some and seeing if I can get past 3.33
 

mtc1

Distinguished
Jul 5, 2008
14
0
18,510




Trying to find something stable above 3.33GHz



Any suggestions? My settings are listed above.



Thanks ahead of time.
 

mtc1

Distinguished
Jul 5, 2008
14
0
18,510
Sorry about that. I forgot this forum doesn't have it in my signature.




Q6600 (G0)
Gigabyte GA-P35-DS3L Rev2
4GB (4 x 1GB) 240-Pin DDR2 800 (PC2 6400)
8600GTS 256MB
2x SAMSUNG Spinpoint F1 HD502IJ 500GB Power Supply
ANTEC AE430 PSU
Antec Three Hundred Case
Tuniq Tower 120
Windows XP Pro SP3 (nLite)
Acer X193w Monitor
 

mtc1

Distinguished
Jul 5, 2008
14
0
18,510






Core Temp 0.99 VID = 1.3250v
 

mtc1

Distinguished
Jul 5, 2008
14
0
18,510
If you think you can make it stable at 3.6, let's have it. I have been trying everything over 3.33. I haven't been able to keep it stable yet.


What you got?



Thanks
 

Lupiron

Distinguished
Feb 9, 2008
1,711
0
19,780
Look like they dont!

And I am busy, I hate BG boards anyways!

Names are the same, or randomly selected.

Not like other boards!

Maybe tomorrow!

--Lupi