i680 and Q6600 OC help - Temps?

Here's my results so far at attempting a 3 Ghz OC on a Q6600 with B2 stepping. I've heard these chips can get up to 3.6 on air. I'm getting the following temps/voltages under prime 95 load. I haven't done the vdroop mod yet. Which software is the most accurate for core temps and voltage? At these temps can I go higher in voltage? Is it OK to run prime 95 for 6 hours with these temps?

Speedfan4.44
Vcore 1.34v
System: 45C
CPU: 65C
Core 0: 62C
Core 1: 64C

Core Temp 1.0
VID=1.3250v
Core 0: 58C
Core 1: 59C
Core 2: 55C
Core 3: 51C

CPU Z (latest version)
Current Voltage: 1.344v

Everlast Ultimate
CPU core: 1.34v
CPU: 65C
CPU 1: 62C
CPU 2: 64C
CPU 3: 59C
CPU 4: 56C
MCP: 45



System specs:
EVGA 680i 122-CK-NF68-A1
Core 2 Quad Q6600 Kentsfield B2 Stepping
Arctic Cooling Freezer 7 Pro CPU fan
4 GB CORSAIR Dominator 1066Mhz CM2X1024-8500C5D
CMXAF1 Ram fan
Two GTX 260 Sonic 216SP in SLI mode
Antec Nine Hundred Black Steel ATX Mid Tower
Antec HGC900 PS
SyncMaster 216BW Monitor
Lite-On IHAS124-04 DVD
Seagate Barracuda 1TB HD

BIOS Settings
Flashed to v P33

Vcore Voltage: 1.41250
CPU FSB: 1.4
[nForce SPP: 1.45
nForce MCP: 1.525
HT nForce SPP <-> MCP: 1.34

HPET Function - Disabled
Nvidia GPU Ex: Disabled
LinkBoost: Disabled
System Bios Cacheable: Disabled
Cpu spread spectrum: Disabled
PCIe spread spectrum: Disabled
HT spread spectrum: Disabled
Sata spread spectrum: Disabled
Sli-Ready Memory: [FSB - Memory Clock Mode: Unlinked

PCIEX16_1 Frequency (Mhz): 100
PCIEX16_2 Frequency (Mhz): 100
PCIEX16_3 Frequency (Mhz): 100
SPP<->MCP Ref Clock, Mhz: 200
nForce SPP --> nForce MCP [5 x]
nForce SPP <-- nForce MCP [5 x]

CPU Multiplier: 9x
x FSB - Memory Ratio: Auto
FSB (QDR), Mhz: 1340 (means 335FSB)
Actual FSB (QDR), Mhz: 1066.7
MEM (DDR), Mhz: 800 (means 400mhz speeds)
Actual MEM (DDR), Mhz: 800

Memory Timing Setting: Optimal
tCL (CAS Latency):5
tRCD: 5
tRP: 5
tRAS: 15
Command Per Clock (CMD): 2T

Advance Memory Settings
tRRD: AUTO
tRC: AUTO
tWR: AUTO
tWTR: AUTO
tREF: AUTO

Limit CPUID MaxVal: Disabled
Intel SpeedStep: Disabled
PPM mode: Native Mode
CPU Thermal Control: Disabled
C1E Enhanced Halt State: Disabled
Execute Disable Bit: Disabled
Virtualization Technology: Disabled

CPU Core 0: Enabled
CPU Core 1: Enabled
CPU Core 2: Enabled
CPU Core 3: Enabled

That looks familiar my Q6600 would read the same temps when overclocked to 3.0- 3.2ghz no matter what i did they just stayed at those temps i switch up to a very basic water cooling kit by thermaltake it would keep it in the high 40's.... so it seem that these chips run hot when OC ......have you looked at corsair self contain liquid units ?? they seem to do a great job with temps

http://www.newegg.com/Product/Product.aspx?Item=N82E16835181017

http://www.newegg.com/Product/Product.aspx?Item=N82E16835181016

dont know if you want to go that rout but its a suggestion

Coretemp/Realtemp is what you should be using for temps. Keep that CPU under 80C while stress testing.

You have a B3 stepping too so those may not hit 3.6Ghz with great temps. G0 is the latest(best) stepping.

Picture is also too small for me to read anything on it. Attach the normal sized file to the forum.

Thanks muelo1000. I'll look into that option but I'd like to see if at least 3.4Ghz is possible on air with the 9x multi.

Tjunction Max value for Q6600 G0 is 100c, but for B3 I heard it was 90C so I'll just set Core Temp at that 90C for Tjmaxx. In the picture it was set at 85C. Tcase for B3 is 62.2C and G0 is 71C. I'll be sure to keep it under 80C fully stressed in Coretemp 1.0 with Tjunction Max set on 90C then.

I've attached the file in zip format so that it can be saved to desktop. The Windows photo viewer may be used to zoom in.

Also under load the fan speed of CPU0 is 1776 RPM as seen in the picture. I'm assuming that the fan is hooked up correctly on the mobo and that the RPM is within expected range?

The memory is rated for 5-5-5-15-2t @2.1v at 1066, so I'll manually set to those timings at 1.9v to not stress the north bridge too much even at the SSP under 1.5v.

After running some Prime 95 small FFT tests, temps on Core temp raised to:

66
68
62
58

That's at a low vcore of 1.3 and 1330 FSB

Since there's a 10 degree difference between 2 of the cores, is it safe to assume that the AS5 wasn't correctly applied?

Taking it up to 3.2 Ghz with a vcore of 1.35 when running the Prime 95 small FFT test brings the temps on Core Temp 1.0 to 75-80C.

I think a Thermalright Ultra 120 Extreme or similar is going to be needed on this air cooled setup to get 3.2-3.6Ghz with temps less than 75C on Core Temp 1.0

Manged to get Core Temp 1.0 to display the correct temps and match Real Temp. Temps in Core Temp 1.0 are too low by 5C for Q6600 B3 stepping. All temps match now at idle and under load, and cores 2/3 being lower than 0/1 is reported to be the norm.

Idle: 40 42 39 38

Blend Test: 59 60 56 51


Small FFTs: 72 73 67 64