Prime Currently running temps @71C - need quick reply!

Page 3 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Toggle sidebar Toggle sidebar
Used settings above but with a higher CPU Voltage of 1.4v (CPU-Z 1.344V)

Prime Results:
Small ffts: 1min, temps ~60-65 across 4 cores
Blend: 16min, temps ~57-62 across all cores
Large ffts: ~1hr, temps 55-60 across all cores

small increase in CPu voltage did help a little.
 
Well, I think you'll need more voltage (about 0.050 more)
My CPUZ reads an average of 1.344 (1.336, 1.344 and only sometimes 1.360) when using prime small ffts. You could say my settings are similar, but the difference is that I'm at 3.3GHZ. I just recently made a test lowering down about 0.050v on my cpu and it posted and everyting, but crashed on prime. Then I started ussing more voltage at 0.0125 steps and every time it crashed but usually needed more time to do so. At last I ended up with the same V setting from before which reported the values just I told you on CPUZ.
So my advice will be to use about 0.050V more, no matter what your settings are now and try.
I believe your settings on CPUZ will have to be near 3.4v to be able to be stable. That's why you bought a Xigmatek, so those settings won't overheat your CPU.
I also concluded that my CPU will need 2x0.0125 steps up (0.0250 voltage increase) to go from 3.3 GHz to 3.4 GHz so you do the math.
I can say almost for sure you need more voltage and that those voltages are safe, since I believe your motherboard under powers your CPU (That's why the settings on BIOS are higher than on CPUZ which I believe is telling the truth).
 
Since I'm almost sure CPU voltage increase will solve the problem, when you get a stable system, try lowering down the other voltages (primarily NB) one by one and step by step.
 
I believe I have hit a Wall.

CPU Voltage has been set to 1.425v (CPU-Z reads as 1.36v) and Prime Crashed after just 15min using small ffts! Worse still temps increased to 74C on cores 0&1 and 70C on cores 2&3!!!

I can't believe i've hit temps that high again! I'm assuming everything is installed ok in terms of seating the HSF and applying the AS5.

Do you think any of the other settings may be causing the CPU temps to increase so high?

or should i wait until my AS5 is past its "Break in" period as apparently this temps should drop further.

 
MMM... it's getting difficult.
1) Have you read about applying thermal paste to a HSF with direct touch technology? they need much more paste than the other kind. If you applied the paste and installed the HSF correctly, the thing is getting strange.
2) Try to do a memtest (google it) but you can first try using just one memory and then the other (if one is faulty when using the other it won't crash).
3) Make sure the fan of your Xigmatek is at max speed when temps reach around 65º
4) Try less and then more (but safe) voltage on the NB
5) If everything fails, get down to 3.2 GHZ (at 9x) make the system stable with the lowest voltages possible and then start going up until you find your max speed.
6) One other thing, if your cpu is overheating at that speed the two most possible reasons are you're using too much voltage or your HSF (considering the one you have) is not properly installed. There's also a possibility that your particular chip isn't able to get to 3.6 GHz, but if it's GO stepping, I doubt it
 
Ahh,
I think you are probably right about the HSF. I read the instructions from Xigmatek (no detail about applying the paste). And i also read the instructions online from Arctic Silver about applying to a quad core CPU. Nothing about extra paste for direct touch though.

I followed instructions to the letter and last night i read a Frostytech review which said the channels between the pipes and the base plate is good to take up excess thernal paste. That did make wonder.

I'm pretty confident that i'm stable in memtest i've run it a few times at 3.6ghz and under current voltage and timings but i can leave it running while i'm at work.

Fingers crossed this will help keep me at lower temps. I think i need about 1.43v in bios (i assume 1.37v in CPU-Z) and i should be stable.

Thanks again.
 
Good, report to us when you can.
I didn't quite get if you used more AC5 as the frotytech review said.
If you ran a memtest for more time than prime needed to crash your system your ram settings should be fine (no need to leave it running while at work).
Remember to watch the speed of the fan at high temps
And tell me what PSU do you have
 
My PSU is a BFG ES-800W - http://www.bfgtech.com/bfgr800wespsu.aspx

I think the motherboard is the culprit for the excessive vDrop. Lupiron posted saying that he was getting 0.9v drop (1.4v set in bios and 1.312v in CPU-Z!) using the same mobo.

I ran memtest for 4 hours it made 8 passes and found zero errors on both sticks.

Also, took the HSF off and checked the Thermal Paste and the spread was terrible! I followed the instructions on the AS5 site and the line of paste they told to apply was in the same direction as the heat pipes. I only had contact with one pipe!!! one was clean and the other had maybe 25% covered by the paste!

I filled the gaps with paste and applied a line of paste to each pipe. Everything is running again and i am getting idle temps at least 5C lower than before (idle as in while i am doing this!).

Arctic Silver need to update their instructions to include HDT's.

Fingers crossed i am going to find success soon.

I feel happier about testing now.
 
Good to know you had a hsf installation prob. I was getting out of ideas. I too believe the MOBO doesn't give you steady voltage, therefore you need a lot more. So remember to increase the voltage and watch out for your temps and fan speed.
If there was a 5º drop at idle, you should have about 8º drop at load. You should be loving your cooler 😍 .
 
Just got an 1hr30min on small ffts - got a BSOD saying that "A Secondary Processor Timed Out" do you know what this means?

none of the cores failed in prime, not reported anyway. When I rebooted my Mobo temp reported 45C so i am assuming it had got hotter.

Temps reached 74 deg on cores 0&1 and on 66C on 2&3. i'm not too worried that they've gone over 70C again because i am hoping a reduction in other voltages should reduce this and also the break-in period of the paste should knock a few off too.

I'll try reducing NB and see if this could be the "secondary processor" issue.
 
Can't get lower on the NB at all. Probably a crazy suggestion but do you think the NB could be still too low?

If the mobo is not holding a voltage on the cpu could this be true of the NB too?

I'll up the NB slightly to 1.54 and see if i can beat the 1 hour 30min.
 
Well you could use a little bit more voltage on the NB but touch it once in a while. Mine used to get very hot (I could only touch it for a second before having to remove my fingers) till I decided to buy a small fan for it, now it runs hot, but doesn't burn.
Make sure the HSF is perfectly installed (double check all push pins) if positive, I believe you should be using too much voltage. if the change on NB voltage fixes the crashes, try lowering the CPU one.
 
I wouldn't compare your chip to Lupi's 1.2 chip, that thing is insane. Anyways, I'm sort of out of ideas for you other than doing like you said and increasing your NB a bit since there seems to be a lot of vdroop. I know you hate to hear it, but you said it earlier, you may have hit the wall, simply the max OC you can get from your chip may not be what you want, but if you can get an 8-12 hour prime at 3.2 or 3.4, I suggest doing it.

Keeping temps low and volts low, you're much better off than finding a flakey 3.6 at high voltages.
 
why are you people still using an out dated stress test like prime? which was only designed to test single core CPU's? get with the times and use the test that Intel uses which can test your CPU at 100% and only takes less than 10 minutes to get a result! check it out it works!
IntelBurnTest v1.8
by AgentGOD
----------------------------
The original. All others are copycats.

Home Page: http://www.ultimate-filez.com

here's my set up
 
I managed 6hr15m running at 3.6. Used 1.425v on CPU everything else fairly similar to before. Core temps reached 73C on 0&1 and 68C on 2&3.

I can't understand why my temps are reaching so highI've read so many other threads of people pumping much more voltage into their q6600 for 3.6ghz and air cooling to below 60C for some.

I've given up (for now) on the 3.6ghz - 6hrs was pretty good. I'm going to try and get a rock solid 3.4Ghz and maybe, once i'm through my break in period with my AS5 i'll try the 3.6 again.

itzamedave - That IntelBurn is Scary! My 3.6 set-up 4 out of 5 passes, crashed on the final. i don't even want to tell you what my temps were!

I like it though, certainly a way of getting quicker results. I think i'll still test with Prime after though becasue it is nice to know that your cpu will last a "good" load for a longer period of time.
 
Too add to above - i really like IntelBurn!

3.4ghz settings passed 5 out of 5 temps hit the max i want to go to but i'm ok with that seeing as this programme seems to stress to the absolute extreme.

i'll try an overnight test on Prime this eve and see how that goes too.

I'm using 378x9, 1.357v on cpu (CPU-Z reads as 1.312v), NB is at 1.46v which doesn't seem to be running too hot. Cpu idle temps are going below 30C which is where it will be most of the time.

Thanks again guys!

still will monitor my cooler i'm still not 100% impressed with the load temps i am getting.
 
you're welcome.

I'm concerned about your cooler too, remember to double and triple check the pushpins, make sure it's perfectly installed. That cooler should be able to let you stay on 3.6 GHz at no more than 70º.

I had a similar problem with my cooler (scythe Infinity) my idle temps were excellent, but at load the would escalate to much, so after some checking (which was difficult, because the cooler is too big and my hands didn't reach all the pushpins) I noticed one pushpin that wasn't perfectly in place, so after solving the problem my idle temps went down only about a degree or two, but my load temps were about 10º less.

Also check that yor fan is running at max speed when reaching above 60º, if you don't know how, use speed fan. I know I told you this before, but since you haven't said anything about it, maybe you haven't read it.

I will say almost for sure, there's a problem with the cooler.
 
ok. Its worth a look. I did install it without removing the motherboard and as you say it it difficult reaching the push pins.

I have some time so i will give it a go and see what happens.
 
OK report on that.

I have to tel you that 3.4 GHz is a very good speed for a Q6600, but I'm at that speed with an inferior cooler and won't ever get more than 71º
 
I suspect something isn't quite right too.

I've read other people's posts and they report load temps using Prime's small ffts at 3.6GHz below 60c sometimes and this is cooling on air.

Are you reporting (& others) reporting max core temps or Cpu temps? my CPU temps stay around mid 60's at load but core temps can reach 74 running at 3.6GHz.

I agree 3.4 is massive, i was just thinking that i remember being amazed at Asus' AiGear utility Oc'ing to nearly 2.6Ghz for me. I thought WOW!
3.4/3.6 is in a whole other world of their own in comparsion.
 
I'm reporting max temp readings of the hottest core (usually 1 & 2) on coretemp and the newest speed fan reports them correctly too.
Are you going to tell us some day about the CPU fan RPM's at full load?
I'll be out for a while but will be back at most on the evening (it's 11:30 now)
 
Well you were right! One of the push pins had come unstuck. it was one of the top ones so i don't know if it was because it wasn't installed correctly (most likely) or from the weight of the HSF - hopefully it was just due to me.

idle temps are more consistent across the 4 cores now. Cores 0&1 at 32C, Core 2 at 31 and core 3 at 27.

I'll test loading temps later. on the above i am expecting Cores 0&1 to be lower and 2&3 to be similar to before. but we'll see.

Fan Rpms don't change much at all +/-30rpm at any time including full load.


 
1) Push pins are extremely tricky, that's why I supposed one was off. I would give only 20% of guilt to the installer and 80% to the designer.

2) Ok, so the rpms don't change, but will it be that the fan is running at full speed all the time (around 1500) or at any other speed? Xigmatek's fan can go from 800 to 1500 rpm.

Tell us you new situation.
 
Yes Push pins are a pain.

Fan rpm is set to 100% Duty Cycle, which is 1400rpm and to be fair its quiet at full speed.

it doesn't ever drop below 1385 nor go over 1410rpm so its fairly steady.

Just did a quick prime run on small ffts and i am now getting max of 69 on Cores 0&1 and 65 on 2&3 although they only hit those temps briefly and then come back down to 67/63c. That's back on 3.6ghz by the way.

so its better and will allow me to continue testing but still not great - i was hoping it would drop to low 60s and give me more scope for increasing my core voltage.

 
You must log in or register to reply here.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom