Alright so I have this wonderful Q6600, G0 stepping, undervolted to 1.15, stock speed, with a Zerotherm FZ120, AS5, stock fan.
I used the line method that Thermaltake describes in their instructions, a thin line across the two (er four) cores under the heatspreader. Lowered the FZ120, gave it a little twice (1-2 degrees), and tightened it down (Pushpins are for sissys).
I burned it on Prime 95 for about an hour, the thermal material probably hasn't had time to really cure yet, but I was wondering what you guys might think of these temps.
Speedfan registered 3 cores (the 4th one was on holiday). I'd sort've forgotten to take a a screenshot during the burn, but I took this one a few seconds in, so it'll just have to do. The two hotter cores never reached 50C, and the cooler one sat at 42-43C the entire time (it must've been at the north end of the die).
Do these temps look good to you guys? On the one hand, and don't quote me on this, but I'm almost certain that the CPU fan sat at 1300rpm the entire time. I was sort've expecting it to speed up, at some point, and it may well have, but it didn't while I was around ( Beginning, Halfway point, end).
I'd had to have to re-seat that heat-sink, but I will if it'll help (considerably).
-SnuffleKitty
Message edited by snufflekitty on 09-27-2008 at 09:31:11 PM
Haven't used speedfan, but my Q6600 at stock speed with a Xigmatek heatsink is around 42-45C using realtemp and coretemp. I didn't try undervolting it though.
Well there's no doubt that the Xigamatek's HDT build gives the best results you can really except from air. I need to reinstall, I got some strange error in the Disk Defrag earlier, but I'll do another burn when I'm up and running again. It's no fecking suprise that I'm a little hotter if that fan sits at 1300rpm the entire time.
I rarely if ever see my temps go above 50c, when I stress the processor, I have seen around 55, but never higher. my second cores are normally around 40c in core temp, and between 35 and 40 in real temp.
That really isn't in line with the temps I've seen in bios, but if it'll make you feel better, I'll go ahead and run the Burn again with Core Temp and Real Temp.
Use Core Temp for monitoring temps. It's the way to go. Also, make sure it recognizes your 4th core. If you're only able to find 3 cores, your chip might be bad (your Prime95 image shows 4 cores working, so you should be fine. Just download Core Temp).
I've got an off brand cooler on mine (I know it's horrible). But at stock speeds I idle at around 37C (just reseated processor 2 days ago, so it hasn't really cured yet). Under load I'm sure I'll hit 50C or so.
When I reinstalled my system recently, I forgot to adjust the fan control settings in my BIOS. Therefore, it was set to some ridiculous default settings (1% fan speed at 30C, and 100% fan speed at 70C).
After adjusting this to something more reasonable (50% @ 30C, 100% @ 55C) my system runs cooler, since the fan wasn't running in gimp mode all the time.
If your fan truly runs at 1300 RPM all the time, you need to adjust the settings in your BIOS to let it throttle up and cooler your chip more. Play with that until you find reasonable temperatures and noise levels.
Message edited by jerreece on 09-27-2008 at 11:05:16 PM
You're probably right about the fan. Honestly, I can't tell if the damn thing is spinning, and it's not because the rest of the fans in my system are considerably louder (1x140mm in the Psu, 3x80mm in the case, and an 80mm on the Gpu). I'd thought that I'd set the fan to scale with rising temperatures, but if Speedfan is registering correctly, that fan is spacing out.
Speaking of which, it would appear that Speedfan is in fact registering correct temps on this cpu (Ha, ha ha, ha, ha). I'd like a little bit of clarification though. Core temp, and Real Temp; I'm seeing a bit of a discrepancy (~4C) between their two readings (We have a confirmation, there is a 4th core). Are the two using different sensors (like the backup sensor they put on the ass-end of the chip) or is that just...normal?
Regardless, here be my Prime95 burn (25 minutes in), using the small FTTs (Because theres 100% load, and then there is 100% load), just like you asked. ;D
No, the difference is in what they have as a default Tjmax. Core Temp uses 100c and Real Temp was using 95C, which I thought they changed to 100C. I guess that one is still in beta.
The number is actually subtracted from Tjunction max to get the core temperature.
Read the Temp Guide sticky I linked earlier for more information.
So, it looks like your temps are good enough to take advantage of the free OC to 3G.
Speed Fan 4.35 tjmax is set to 100C. I'm getting the same reading for it on CoreTemp.
I use the same HS, ZeroTherm NV120, but I use MX-2 thermal grease.
To really see how hot the Q6600 will get, like Zorg mentioned, use the Max FPU stress test. The 800000 iterations is what makes the cores get has hot as they will get.
Message edited by Grimmy on 09-28-2008 at 04:19:07 PM
Well, I may still re-seat it, just to see if it'll make any difference; later perhaps, but I'll leave it be for now. Thermal Paste tends to be a bit messy.
No, the difference is in what they have as a default Tjmax. Core Temp uses 100c and Real Temp was using 95C, which I thought they changed to 100C. I guess that one is still in beta.
The number is actually subtracted from Tjunction max to get the core temperature.
Read the Temp Guide sticky I linked earlier for more information.
So, it looks like your temps are good enough to take advantage of the free OC to 3G.
The reason Realtemp uses TjMax of 95 degrees is because thats the correct value; Intel has since confirmed it.
Realtemp: TjMax 95 (Proven correct)
SpeedFan: TjMax 100 (5 degrees high)
CoreTemp: TjMax 105 (version I last used anyway) (10 degrees high)
That's what it says for the newer.. They just didn't mention the 6000 series.
Edit:
Forgot to mention, if I use realtemps default (tjmax 95C) my cores register lower then my Tcase. So it more of a matter of calibration them, if I want the best accurate temps. Just going with defaults on apps that show readings of cores are not going to be totally correct.
So I'd say the tjmax for a Q6600 is around 98-100C, rather then 95C.
Message edited by Grimmy on 09-29-2008 at 08:05:07 PM
[unclewebb] "I guess I'm not 100% convinced that switching RealTemp to TjMax=100C is going to result in more accurate temperatures. There was more important information left out of the IDF presentation than included in it. The only thing I'm really happy about is that I didn't spend a $1000 + travel to get my hands on that presentation. My rant for the day.
Having said all that, about all I can do is hop on the TjMax=100C bandwagon and then go back and do some more testing to see if this is reasonable across the entire temperature range or not.
Here's version 2.75 which incorporates the new IDF TjMax values for 45nm as well as similar changes to TjMax for 65nm.
Message edited by Zorg on 09-30-2008 at 04:07:49 PM
Well, whatever the case may be, I'm still at a bit of a loss as to why my fan won't increase it's rpm under load.
CPU Smart FAN Control - Enabled
&
CPU Smart FAN Mode - PWM Mode
And yet Speedfan reads the rpm between 1295-1298 (Basically 1300rpm). Honestly, I even stuck my head down there during load, and I didn't notice any increase in airflow, or noise, which would seem to backup Speedfan's story.
Zerotherm rates that fan as supporting speeds between 1100-1800rpm. It's a 4pin fan so PWN mode should work. I'll try voltage but that's really meant for 3pin fans. Auto doesn't seem to do any good either.
Does CPU Smart FAN control only work if I have easytune installed?
-SnuffleKitty
Message edited by snufflekitty on 09-30-2008 at 07:50:17 PM
Well.. I have my Zerotherm on a fan controller. So my fan speed will go up to 2580 rpm. I normally run it in the 1800's.
As far as speedfan, it needs MB support, as well as having 4 pin. You might have to check it out on the website. Also, the MB fan voltage, it may be using 5-7V to power the fan, that might be the reason for the 1300 max speed. Since I use a fan controller, I'm actually using up to 12v.
Smart Fan in the bios on the other hand, is more of a switch to turn it off/on to max speed if I remember correctly, since it kicks it on at a certain temp.
Disable the fan control and see if it sounds different at max. If it is louder, then the mobo apparently doesn't have as radical a fan speed profile as you want.
You can use Speed Fan to set a CPU fan speed profile, depending on your mobo. I made a stock cooler fan for a Dell 3G Prescott ramp up faster and it worked well. It doesn't seem to step down as well, but that is of little concern to me.
Message edited by Zorg on 09-30-2008 at 10:20:27 PM
The EasyTune5 software that shipped with my board is registering my temps in line with Real Temp 2.70. Real Temp 2.75 is reading the same as Core Temp.
I'm just not going to worry about this anymore. Unclewebb has decided to err on the side of caution; better that we think that the core is 5 degrees hotter than it actually is than the alternative.
-SnuffleKitty
Message edited by snufflekitty on 10-01-2008 at 08:26:58 AM
I installed EasyTune, and lowered the threshold at which the fan should be running 100% from 60C to 45C in it's Smart Fan interfance, and now both Speedfan and EasyTune are reading my Fan's rpm between 1300-1800 depending on the load. It's a quiet fan, I'm not entirely sure that I would notice the difference between 1300rpm and 1800rpm; I'll lower the threshold a bit more and report back on this.
I haven't done anything else that would affect my motherboards Smart Fan settings. I know that 1800 rpm isn't much, but any change in this fan's rpm is a welcome change.
-SnuffleKitty
Message edited by snufflekitty on 10-01-2008 at 08:38:40 AM
RPM Silent Mode 700 ±10% RPM
Max mode 2600 ±10% RPM
Air Flow Max 84.7 CFM
Noise Level Under 39 dBA
So your fan speed is basically where it needs to be at 100%@1800rpm. Sorry if I confused ya. I thought you had the NV120, and over looked you had a FZ120 instead.
My mainboard is the Gigabyte X48-DS5. I'd considered the DS4, but with a price difference of ~$5, I couldn't really think of a way to justify going with the board with fewer bells and whistles.
I have two sticks of Corsair Dominator 2GB DDR2 PC8500 running in Windows Xp 32bit.
For whatever reason (Bios), I couldn't get the damn thing to post on anything but it's default settings (266x9). Whats more, the board sets the Ram timings to something like 5-7-7-24 (AGGGGHHHHH). Corsair lists this kit's timings as 5-5-5-15, but when I set them manually, the system wouldn't post. The DDR2 voltage is correctly set to 2.1v. This system is rock solid at stock (4hrs Prime95), but I purchased these components with the expectation that they would be viable for overclocking when the need arose.
Yesterday I flashed my Bios to the latest Gigabyte lists for this board (@Bios is sorta neat), so I'll be playing around with the settings tonight, to see if I cannot get a little more out of this Q6600. I've read a few other threads, it looks like the F5 Bios that shipped with my board are trash for overclocking, but the F8a Beta Bios are golden. I'll let you guys know how this turns out.
If that doesn't work, I'll try playing around with my ram. I currently have the kit installed in the two red slots; using the yellows slots would cause problems between the kits fan and my HSF. I really don't know how much good it (The Memory Kits Fan) does, but it is quiet, and I can't see how a little bit o' airflow over the ramsinks could hurt.
-SnuffleKitty
Message edited by snufflekitty on 10-03-2008 at 07:09:42 PM
You are about to answer a thread that has been inactive for more than 6 months. If you still wish to proceed, please ensure that your posting is original and does not duplicate or overlap any prior responses to this thread.