Q6600 + Ultra-120 extreme @ 3.4 = 70c!?

[touser]

Hi everyone, i've been out of the overclocking scene since my watercooled dual 2.4 xeon days. I decided to give it a shot again last night with my q6600. The problem now seems to be it is running much warmer than it should be? At 3.4ghz, 69-70c on core 2, while the other cores hover around 63-67c while running OCCT. The ambient room temp is 73, with the vcore reporting 1.31-1.32v under full load. I applied as5 several days ago following the instructions on the arctic silver website for quad cores precisely. The HSF is tightened down on all four corners until the springs bottomed out.
Any insight would be fantastic.

Bios Settings:
Vcore = 1.385
VTT = 1.23
MCH = 1.29
ICH = 1.05
ICHIO = 1.5
DRAM = 2.2v @ 1:1
fsb = 378


System specs:
Q6660 G0, vid: 1.25v
Abit ip35 pro
8800gtx + HR-03+ @ 650/1000
4gb Crucial Ballistix BL2KIT12864AA804
(2x) 120mm intake fans, (2x) 120mm exhaust
120mm fans on cpu and gpu HSF

 

[Hardass]

Did you remove the hsf and check contact imprint? You may have a concave cpu or hs that needs to be lapped.

 

[benbaked]

How does the base of your heatsink and your quad's IHS look? If they're not lapped they can surely benefit from it, when lapped you may see a decrease in temps of up to 10c. Additionally you may want to use a more powerful fan on your heatsink, or use two of them in a push/pull method. Is this a B3 or G0 stepping quad?

 

[Opteron 1100]

those temps are a little high for that cooler but , really at that speed it is not unusual for quads.
Do you have the Fan EQ enable to ramp up the fan as temps go up ? I would
also try and re-seat the HS just to make sure you got it mounted correct
with enough paste , then post back.

 

[touser]

Thanks for the help. I will dismount the hsf a little later and check the contact imprint. Also, its is a G0 stepping. The fan is blowing full speed, i forget how many CFM but its fairly high.

 

[touser]

I pulled the HSF off to check the contact patch and everything seemed alright. I remembered microcenter had a dangerden mc-tdx waterblock in stock so i bit the bullet and picked that up. Pulled the old radiator and swiftech pump out and just finished the install. 55-56c full load now with OCCT running, hopefully it will drop slightly after the as5 sets up, but needless to say i'm pleased.

 

[Opteron 1100]

wait you have water cooling ?
I get those temps on air

 

[BLOOP!]

He's in southern california. Ambient temps in here are really high right now

 

[leftheaded]

I pulled the HSF off to check the contact patch and everything seemed alright. I remembered microcenter had a dangerden mc-tdx waterblock in stock so i bit the bullet and picked that up. Pulled the old radiator and swiftech pump out and just finished the install. 55-56c full load now with OCCT running, hopefully it will drop slightly after the as5 sets up, but needless to say i'm pleased.

looks like the water brought the temps down. i had similar results with my TRUE, so I lapped it. i've got it running on a different q6600 (that isn't lapped) and it's performing way better.

also, it was only after my second or third install that i realized there is a divot in the base of the HSF where the center of the x-bracket must sit. if it isn't centered right, you won't get even-pressured contact

 

[baditude_df]

Yea, those load temps with an mc-tdx are very close to what I get @ 3.6 with my quad. Factor out the high ambient you are getting right now, and we would be running virtually the same temps. I've never did much testing at 3.4GHz, but our voltages also seem very close. I run a stable 3.6 @ 1.375V. Have you left a little headroom on the voltage? Or is that the lowest stable voltage you found to run 3.4? As it happened, I bumped the voltage to 1.4 when I made the jump from 3.2 to 3.4 with the intention of incrementally bumping it down to find the lowest stable voltage. When I made the jump to 3.6GHz I forgot to bump up again, and to my surprise it booted and primed stable. I was quite pleased.

Your temps with the TRU120E remind me of what I used to get on a REALLY hot day. Since I haven't got AC, I clocked down to 3.0GHz on those days. I wouldn't say there was anything significantly wrong with your air setup other than the high ambient temps.

 

[touser]

Yea, those load temps with an mc-tdx are very close to what I get @ 3.6 with my quad. Factor out the high ambient you are getting right now, and we would be running virtually the same temps. I've never did much testing at 3.4GHz, but our voltages also seem very close. I run a stable 3.6 @ 1.375V. Have you left a little headroom on the voltage? Or is that the lowest stable voltage you found to run 3.4? As it happened, I bumped the voltage to 1.4 when I made the jump from 3.2 to 3.4 with the intention of incrementally bumping it down to find the lowest stable voltage. When I made the jump to 3.6GHz I forgot to bump up again, and to my surprise it booted and primed stable. I was quite pleased.

Your temps with the TRU120E remind me of what I used to get on a REALLY hot day. Since I haven't got AC, I clocked down to 3.0GHz on those days. I wouldn't say there was anything significantly wrong with your air setup other than the high ambient temps.

What VTT and MCH voltages are you running? Also, what DRAM ratio? Right now i still have it clocked at 3.4ghz, bios vcore of 1.385, full load vcore reporting 1.33. I did bump it up to 3.5ghz and started OCCT, all seemed to be going well so i went out to lunch and came back to a blank screen and a hard-locked system. Dropped it back to 3.4 and i'm going to prime it for 24 hours, if that holds up i guess i'll bump the voltage up to 1.395 and see where it goes?

 

[baditude_df]

Most of it's in my sig. DRAM Ratio 8:11
Reported Vcore is 1.288 idle/load when set to 1.3125 in BIOS(I have my system clocked back to 3.2V for the hot weather).

Platform Voltages are as follows:
FSB - 1.5v
DRAM - 1.35
SPP(nb) - 1.45
MCP(sb) - 1.57
HT(hypertransport) - 1.35
Settings mnemonics are different due to NVidia chipset, but they all mean the same thing.

 

[touser]

Ahhh, ok. Your nb and sb voltages are way higher than mine, maybe i should try bumping those up.