Normal temperature for Q6600 under load
 

Hello

I'm thinking there might be a problem with the cooling on my box.
When running 4 instances of llr.exe i'm running a temperature of 78-80C.
This is on a stock Q6600 (Quad-Core 2.4ghz) running at factory settings (no overclock).

Anybody else running llr on Q6600 who might shed some light as to wether this is normal or not?

Any help is much appreciated!

My Q6600 is running @ 3.2GHz @ 1.28V real... under stress I get 55-60C depending on core. This is a G0 stepping CPU cooled by Thermalright Ultra120 Extreme.

[quote=mcduck;144462]Hello

I'm thinking there might be a problem with the cooling on my box.
When running 4 instances of llr.exe i'm running a temperature of 78-80C.
This is on a stock Q6600 (Quad-Core 2.4ghz) running at factory settings (no overclock).

Anybody else running llr on Q6600 who might shed some light as to wether this is normal or not?

Any help is much appreciated![/quote]
seems a bit much for me intel recommends less than 70C for G0 stepping and 62.2 for the older B3 stepping
mine runs at 58 in a warm room when overclocked to 3ghz
when the room is cool it only goes up to about 48

Thanks both for the answers.

Seems there must be a problem with my fan / thermal paste or ventilation. I'l take the box apart tomorrow to apply the thermal paste again. hopefully that was the main issue.

[quote=mcduck;144491]Thanks both for the answers.

Seems there must be a problem with my fan / thermal paste or ventilation. I'l take the box apart tomorrow to apply the thermal paste again. hopefully that was the main issue.[/quote]

There's good information on thermal relationships in this [URL="http://www.tomshardware.com/forum/221745-29-core-quad-temperature-guide"]post[/URL].

Get rid of the stock HSF. I use Thermaltake Big Typhoons.
Mine runs at 30C or 86F with DNET OGR cranking away for weeks and months.
No case of course. Room temp is kept between 76F and 80F.

[QUOTE=mcduck;144462]Hello

I'm thinking there might be a problem with the cooling on my box.
When running 4 instances of llr.exe i'm running a temperature of 78-80C.
This is on a stock Q6600 (Quad-Core 2.4ghz) running at factory settings (no overclock).

Anybody else running llr on Q6600 who might shed some light as to wether this is normal or not?[/QUOTE]

I have a Q6600 cooled by a Zalman 9700 air cooler that maintains 58C under all four cores at P95 peak load. And I consider that "high" because that system is kept in an enclosed "hutch" under my wife's desks where ambient temps tend to rise due to less cool air coming in.

I had a QX6700 watercooled that maintained about 57C.

It's not fair to compare a 165W Kentsfield to a 95W Yorkfield (Penryn)... btut I will. I *just* upgraded this week from that QX6700 to a Q9650, and switched from water cooling to another Zalman. It is currently undergoing the overnight torture test (SFSG)... and is maintaining 40C under load peak, sometimes as low as 36C.

70C+ is *way* hot, even for a Kentsfield (65 nm) chip. *WAY* hot. Damaging hot. Seriously. I've fried several motherboards at that level (though, thankfully, not a CPU).

Jester

[QUOTE=IronBits;144507]Get rid of the stock HSF. I use Thermaltake Big Typhoons.
Mine runs at 30C or 86F with DNET OGR cranking away for weeks and months.
No case of course. Room temp is kept between 76F and 80F.[/QUOTE]

The Zalman CNPS 9700 is the current "nuts" (in my opinion) for air coolers. Find it at frozencpu.com.

I find that closed cases make for errors and higher temps. I tend to run "open" cases as well, with an external fan blowing good cool air into the case. I had some temp problems with the RAM in my upgrade. Even thought the CPU was hovering at 48C, the RAM was overheating, and a direct airflow over the RAM eliminated the roundoff errors I was having. The direct "open fan" didn't do much for the CPU core temp (from 44C to 40C), but it dropped the ambient temp in the case (and thus the temp around the RAM) by 10 to 12 C.... enough to make a difference.

It isn't all about the CPU core temps.

Jester

I had the same problem with the same processor on a new box. I did not have those problems with two other Q6600 processors. 80 degrees Celsius is the maximum temperature before the processor shuts down automatically. Turned out one of the four pins that attach the cooler to the motherboard was not attached correctly. Those connections are woefully inadequate and prone to unnoticed mis connection. Fortunately my processor is still in good shape and performs well at a temperature just below 60 degrees Celsius using the same Intel cooler.

When running 4 separate instances of P95, my G0 overclocked to 3Ghz runs in the low to mid 60s. This is with a Zalman 9700 for cooling. Having a stock 2.4GHz hit 78-80 degrees is indeed worrying.

Funny how CPUs can only take up to 80 degrees before shut down. But for a GPU (f@h for example), 80 degrees would be considered borderline ok - start panicking around 90-95 degrees.

I also had that problem one year ago and I found it was a problem of not cleaning the dust of the stock HSF.

Just for information, my Q6600(no overclock) runs at ~70 °C with the standard Intel fan and 4 llr instances running.
[CODE]hugo@hugo-desktop:~$ sensors
it8718-isa-0290
Adapter: ISA adapter
in0: +1.23 V (min = +0.00 V, max = +4.08 V)
in1: +1.81 V (min = +0.00 V, max = +4.08 V)
in2: +3.25 V (min = +0.00 V, max = +4.08 V)
in3: +2.94 V (min = +0.00 V, max = +4.08 V)
in4: +1.63 V (min = +0.00 V, max = +4.08 V)
in5: +0.00 V (min = +0.00 V, max = +4.08 V)
in6: +0.08 V (min = +0.00 V, max = +4.08 V)
in7: +3.06 V (min = +0.00 V, max = +4.08 V)
in8: +3.10 V
fan1: 1934 RPM (min = 0 RPM)
fan2: 0 RPM (min = 0 RPM)
fan3: 2070 RPM (min = 0 RPM)
fan4: 2083 RPM (min = 0 RPM)
temp1: +33.0°C (low = +127.0°C, high = +127.0°C) sensor = transistor
temp2: +58.0°C (low = +127.0°C, high = +127.0°C) sensor = thermal diode
temp3: -2.0°C (low = +127.0°C, high = +127.0°C) sensor = transistor
cpu0_vid: +0.000 V

coretemp-isa-0000
Adapter: ISA adapter
Core 0: +72.0°C (crit = +100.0°C)

coretemp-isa-0001
Adapter: ISA adapter
Core 1: +67.0°C (crit = +100.0°C)

coretemp-isa-0002
Adapter: ISA adapter
Core 2: +67.0°C (crit = +100.0°C)

coretemp-isa-0003
Adapter: ISA adapter
Core 3: +72.0°C (crit = +100.0°C)
[/CODE]

ogr?
 

[QUOTE=IronBits;144507]Get rid of the stock HSF. I use Thermaltake Big Typhoons.
Mine runs at 30C or 86F with DNET OGR cranking away for weeks and months.
No case of course. Room temp is kept between 76F and 80F.[/QUOTE]

I don't think OGR cranking would heat up a cpu much in any case since it doesn't use any floating point to start with.

[quote=ADBjester;144514]The Zalman CNPS 9700 is the current "nuts" (in my opinion) for air coolers. Find it at frozencpu.com.

I find that closed cases make for errors and higher temps. I tend to run "open" cases as well, with an external fan blowing good cool air into the case.[/quote]

I read a review that put the [URL="http://www.frozencpu.com/products/8432/cpu-noc-12/Noctua_NH-U12P_SE1366_CPU_Heat_Sink_Socket_LGA_1366.html#blank"]Noctua[/URL] ahead of all the rest for heat dissipation and noise level. And, they make a version for the Nehelem!

I'd have to politely disagree about running an open case. Properly engineered airflows (might not get this on just any case) are important. When running 100% 24/7 one should consider front and rear fans, check cable placement and other items that obstruct air flow, etc. I've even manufactured baffles at times to direct or block air flow.

I would tend to blame case or motherboard design before saying an open case is the solution. But if you've found what works for you, and you seem to be aware of the issues, by all means... Intel sometimes lists cases that are "certified" for a particular platform, which I think includes airflow. My Intel manufactured server chassis has very detailed airflow design.

[QUOTE=Freightyard;152404]I read a review that put the [URL="http://www.frozencpu.com/products/8432/cpu-noc-12/Noctua_NH-U12P_SE1366_CPU_Heat_Sink_Socket_LGA_1366.html#blank"]Noctua[/URL] ahead of all the rest for heat dissipation and noise level. And, they make a version for the Nehelem![/QUOTE]

I'm very happy with my Noctua, and it is very quiet.

Jeff.

I've got both the Zalman CNPS9700 and the ZeroTherm FS120 - running 2 PCs. Started off with the Zalman with an earlier build. Works great, happy bunny.

But when I came to build my next PC, I just couldn't get my head around the price. It's more than double of the ZeroTherm, so I went for the latter. On my overclocked Q6600, the ZeroTherm matches the Zalman in performance - maybe even slightly cooler. A bargin considering it's so much cheaper.

On the flip side, the fan on the ZeroTherm isn't the best. I'll replace it one day with a Panaflo (my favourite make for fans). The Zalman's fan I think is better, but can't upgrade since it's non-replacable.

I also found the Zalman's clip mechanism to be more fiddly than the ZeroTherms. The cooler can rotate a little and the clasp wasn't easy to get in place due to it being springy. ZeroTherm's was far easier - just tighten the screws in the four corners.

Weighing the factors, unless Zalman halves their prices, on my next build I'd go for the ZeroTherm again.

I use a Zerothem FZ-120 on my OC'd Q6600 (3.2 GHz) and my four cores run in the low 40's Celsius at full load. I also use one on my dual core E8500. My only complaint is that it is slightly too large. In one application it interferes with closing the case because it is too tall (so now I just leave it open) and in the other it hangs over one of my memory slots (so I use the other pair of memory slots, as I just have two sticks). This would be really bad if I wanted to close my computer case or run four sticks of memory, which shouldn't be asking too much of computer hardware these days. I consider that a bad design, but it cools so well and is so inexpensive that I like it nonetheless. It works for my applications. Right now newegg sells it for $44 with a $30 rebate, so it costs just $14. That's pretty hard to beat.

[quote=mcduck;144462]Hello

I'm thinking there might be a problem with the cooling on my box.
When running 4 instances of llr.exe i'm running a temperature of 78-80C.
This is on a stock Q6600 (Quad-Core 2.4ghz) running at factory settings (no overclock).

Anybody else running llr on Q6600 who might shed some light as to wether this is normal or not?

Any help is much appreciated![/quote]

Have you removed this transparent plastic sticker on the cooler or/and cpu?

[quote=mcduck;144462]Hello

I'm thinking there might be a problem with the cooling on my box.
When running 4 instances of llr.exe i'm running a temperature of 78-80C.
This is on a stock Q6600 (Quad-Core 2.4ghz) running at factory settings (no overclock).
[/quote]

It also depends on the tool which is reading out the temps. Because the temps can't be readout direct, but are given as a distance to the T-Junction temperature (the one at which the processor begins to throttle).

Poorly the T-Junction of each processor isn't announced by Intel, and varies from 80°C to 115°C (as far as i know), so the Tools must guess it.

So, you must either know what T-Junction value your tool uses, or you use a tool which gives out the difference to T-Junction. This way you know, how far away from throttling (and so from save usage) your processor is working.

I'm using Realtemp, please look at [URL]http://www.techpowerup.com/downloads/1139/Real_Temp_2.70.html[/URL]

A Value of 30°C to 20°C should be fine in most cases.

hot mosfets
 

I was measuring temperatures on my motherboard (with a hand-held IR meter) and noticed that the power mosfets around my OC'd Q6600 cpu were hot....~190 degrees F. Maybe this is normal for them, but it concerned me. I'm a few days away from building this system into a case, and for now it is free-standing. I had a cpu HS/fan, one fan on the NB and one on the SB, but nothing for the mosfets. I added a large fourth fan for them and got their temperature down to ~120 degrees F, which feels safer. I'm surprised more people don't talk about cooling these parts. I do see some motherboards have mosfet heatsinks and even active cooling for them, but not all. For fun, I wanted to see if my new cooling would allow for a bigger OC and it did. I raised the system to 1700 MHz FSB (from 1600), 3.4 GHz (from 3.2), and it ran stable for an hour. Previously it would crash in about 4 minutes at these settings. I'm not personally comfortable with these numbers, so I put it back to where I've been running stable for a month now. But I just wanted to mention this might be something people overlook with regard to temperature, OC, and system stability. Case enclosure will probably help get some air moving over them, but I'm still going to add a dedicated fan blowing on the mosfets when I build my case next week.

I have been running Prime95 on my Dell Studio 15 laptop that has a Core 2 Duo, and with both cores running the temperatures were 75-80 C (very high for a desktop CPU but I thought it wasn't horribly uncommon for a laptop). I switched the stock thermal paste with Arctic Ceramique and the temperatures dropped to 59-64 C!:surprised Which is a tremendous difference IMO

Just thought I'd share my experience as well as chime in that correctly applied, decent thermal paste can make a noticeable difference in some situations.

Edit: Forgot to mention that the temperature monitoring software that I use is Core Temp 0.99.4