Speedfan has Vcore2 @ 2.10v

[NinjaJedi]

I have my e6600 running on 1.4v speedfan (4.33) shows Vcore1 @ 1.41v and Vcore2 @ 2.10v Is this a speedfan misread? My setup is running 400x8 for a 3.2ghz clock. In coretemp it shows 450x8 for a 3.6 CPU-Z shows the correct clock. Is that a coretemp misread?

e6600 with a zalman 9500
ga-965p-s3 (rev 1.1)
patriot 2x 1gb @ 400mhz 4-4-4-12
antec NEO HE 550
x1800xt 512mb with a zalman vf900

Another question. I ran that setup on prime95 and coretemp and speedfan topped out at 62c but TAT never went about 58c what is the more accurate temp? Are those temps ok?

 

[nullpointerus]

Originally posted by: NinjaJedi
I have my e6600 running on 1.4v speedfan (4.33) shows Vcore1 @ 1.41v and Vcore2 @ 2.10v Is this a speedfan misread?

Vcore1 = CPU
Vcore2 = RAM

So you must be running your RAM at 2.10v, right?

My setup is running 400x8 for a 3.2ghz clock. In coretemp it shows 450x8 for a 3.6 CPU-Z shows the correct clock. Is that a coretemp misread?

Yep, a fairly common one.

e6600 with a zalman 9500
ga-965p-s3 (rev 1.1)
patriot 2x 1gb @ 400mhz 4-4-4-12
antec NEO HE 550
x1800xt 512mb with a zalman vf900

Another question. I ran that setup on prime95 and coretemp and speedfan topped out at 62c but TAT never went about 58c what is the more accurate temp? Are those temps ok?

Which test from Prime95?

Small FFT --> tons of heat
Large FFT --> moderate heat
Blend --> moderate heat

I run about 60c load in Small FFT, which is decent, considering my fans are very low speed.

LINPACK will kill those temps. I see 75c load in LINPACK, and my processor continually throttles.

In a game, however, I might only see 35c or 45c load temps, which are great.

 

[NinjaJedi]

Originally posted by: nullpointerus

Originally posted by: NinjaJedi
I have my e6600 running on 1.4v speedfan (4.33) shows Vcore1 @ 1.41v and Vcore2 @ 2.10v Is this a speedfan misread?

Vcore1 = CPU
Vcore2 = RAM

So you must be running your RAM at 2.10v, right?

My setup is running 400x8 for a 3.2ghz clock. In coretemp it shows 450x8 for a 3.6 CPU-Z shows the correct clock. Is that a coretemp misread?

Yep, a fairly common one.

e6600 with a zalman 9500
ga-965p-s3 (rev 1.1)
patriot 2x 1gb @ 400mhz 4-4-4-12
antec NEO HE 550
x1800xt 512mb with a zalman vf900

Another question. I ran that setup on prime95 and coretemp and speedfan topped out at 62c but TAT never went about 58c what is the more accurate temp? Are those temps ok?

Which test from Prime95?

Small FFT --> tons of heat
Large FFT --> moderate heat
Blend --> moderate heat

I run about 60c load in Small FFT, which is decent, considering my fans are very low speed.

LINPACK will kill those temps. I see 75c load in LINPACK, and my processor continually throttles.

In a game, however, I might only see 35c or 45c load temps, which are great.

aah ok yes RAM is at 2.1 Thanks. I was running a blend test.

 

[VirtualLarry]

Originally posted by: nullpointerus
Which test from Prime95?

Small FFT --> tons of heat
Large FFT --> moderate heat
Blend --> moderate heat

I run about 60c load in Small FFT, which is decent, considering my fans are very low speed.

LINPACK will kill those temps. I see 75c load in LINPACK, and my processor continually throttles.

In a game, however, I might only see 35c or 45c load temps, which are great.

Your processor throttles at 75C? Something's very wrong there. What CPU and stepping do you have? I think you've been mislead by the newest coretemp. I believe that your temps are actually closer to 90C, in which case yes you might be throttling.

I don't know why people don't believe me that CoreTemp 0.96.1 is wrong.

 

[nullpointerus]

Originally posted by: VirtualLarry

Originally posted by: nullpointerus
Which test from Prime95?

Small FFT --> tons of heat
Large FFT --> moderate heat
Blend --> moderate heat

I run about 60c load in Small FFT, which is decent, considering my fans are very low speed.

LINPACK will kill those temps. I see 75c load in LINPACK, and my processor continually throttles.

In a game, however, I might only see 35c or 45c load temps, which are great.

Your processor throttles at 75C? Something's very wrong there. What CPU and stepping do you have? ...

I have an older L2 stepping of E4400:
http://processorfinder.intel.c...tails.aspx?sSpec=SLA3F

... I think you've been mislead by the newest coretemp. I believe that your temps are actually closer to 90C, in which case yes you might be throttling.

I don't know why people don't believe me that CoreTemp 0.96.1 is wrong.

Actually, I was talking about SpeedFan's output, not CoreTemp's.

 

[redhook]

NinjaJedi:

Check out the Computronix's Temp Guide at Tom's (sticky in Overclocking-CPU) re need to offset SpeedFan by 15 degrees for 100 degree Tjunction max chips--like yours with L2 stepping. Meaning that if SpeedFan is showing 75, you're probably at 90!

 

[redhook]

Sorry, I meant to direct my previous post to Null.

 

[nullpointerus]

Originally posted by: redhook
[nullpointerus]:

Check out the Computronix's Temp Guide at Tom's (sticky in Overclocking-CPU) re need to offset SpeedFan by 15 degrees for 100 degree Tjunction max chips--like yours with L2 stepping. Meaning that if SpeedFan is showing 75, you're probably at 90!

Ouch.

Thank you, Larry & redhook, for telling me this!


I may back the chip down to 2.6-2.8 GHz, where its voltage requirements are much lower.

 

[VirtualLarry]

You need better cooling for sure.

 

[nullpointerus]

Originally posted by: VirtualLarry
You need better cooling for sure.

Well, those temps (+15c) are at 1.5v & 50% o/c at full load.

75c during Prime95 Small FFT does sound a little high regardless, but I'm not sure what could be improved.


CPU Coolers

I've tried the following coolers, with roughly identical results:

-- Tuniq Tower 120 (included fan @ 2100rpm)
-- GeminII (2 120mm Yate Loon fans @ 1350 rpm)

With the GeminII, I was able to add a 140mm intake over the CPU vent, but this produced no change in CPU temps. Nonetheless, the side panel fan did dramatically lower my NB idle temps (i.e. >50c to <40c) since the NB has only passive cooling (and no heatpipe).


Case Cooling

My case is a Cooler Master 690 w/ the following fans:

-- (2) 140mm <-- top exhausts
-- (1) 120mm <-- rear exhaust (rattles terribly when used as intake BTW)

-- (1) 140mm <-- bottom intake
-- (1) 140mm <-- side CPU intake
-- (1) 140mm <-- side GPU intake
-- (1) 120mm <-- front HDD intake

(The PSU is on a separate airflow path.)

I can run the computer with the side panel off in a room ~70F, but there's no real change.

However, even with the side panel closed, the air around the CPU case exhaust fans is *never* warm.

In case you're thinking all those fans near the CPU cooler created "dead spots," I saw no decrease in temps from running without the top exhaust fans.


TIM Application

When installing the HSF's, I applied the recommended amount of arctic silver (i.e. a pea-sized amount in the center of the CPU's heatspreader) and tightened the cooler mounts well.

When I removed the Tuniq to install the GeminII, the Tuniq left a nice, even coating of arctic silver on both the CPU and heatsink. So I'm guessing the TIM was applied correctly.

(Yes, I did clean off what the old cooler left behind before re-applying TIM and installing the new cooler.)


So... where should I start looking... ?

Does any of the above information generate some ideas?

 

[VirtualLarry]

Hmm. It sounds like you know what you're doing, and there seem to be plenty of case fans. I can't believe you are getting those high a temps with a Tuniq. Then again, you are at 1.5v. My E2140 M0 is at 3.2Ghz, 1.425v, with a HyperTX2 and loads at 85C with Prime95 small FFTs. But I still don't get throttling. How do you know you are throttling? Did you test with RMclock?

 

[nullpointerus]

I didn't verify the throttling with anything that can detect/display EIST settings.

The reason I think my processor throttles under LINPACK is that my core temperatures (in SpeedFan 4.33) continually go up and down like 20c every second, and the overall CPU temp (Tcase?) fluctuates wildly, too.

What bothers me is that the air coming off the CPU cooler isn't warm. One would expect, with decent heat transfer through the heatpipes, that the air coming off the HSF would at least be warm when the CPU is hitting 90c internally. Right?

I mean, I had a P4 box where the exhausted air was definitely warm.


I'll try RMClock tomorrow -- if it works with Vista x64 --
... and I have some ideas for more thorough testing with the side panel off. :thumbsup: