Core i7 runs TOO HOT with Prime 95 & side panel attached!

I was thinking about getting some improved performance out of my current build which uses a Core i7 965, a single GTX 295 video card and a 910w PSU from PC Power & Cooling. I currently only have the stock cooler on my CPU which I intend to change before I overclock, but first I wanted to get a good reference point as to how well my processor performs with the Intel heatsink.

What I've noticed is that while running Prime 95 with the side panel on, my core temperatures hit 90C after only 2-5 minutes, at which point I quickly shut off the test. However, if I remove my PC's side panel, my cores all perform in the 70s until I voluntarily turn the test off after 20 minutes. I don't seem to have any extreme temperature issues while playing games with the side panel still attached. (Readings in all scenarios are the same whether I use RealTemp or CoreTemp.)

The case, if anyone is interested, is a limited edition version of the Cooler Master Stacker 830. Instead of the side panel with the mesh screen for air flow, it now has an acryllic window (so you can watch your parts baking like a pie in the oven :lol: ). I realized before buying it this might be an issue, but I actually have 9 120mm case fans inside and an 80mm fan keeping my hard drive cool. (Don't worry - I can't hear any of them. The CPU & PSU fans are definitely louder.) There's an intake on the bottom of the case (which is raised about an inch off the top of the desk), and two exhausts (one in the back and one on top, although the back is definitely warmer). The video card also exhausts both GPUs out the back instead of the newer single-PCB version which exhausts one of the two GPUs back into the case.

So I'm curious if other users have similar experiences - or is it the general opinion of everyone that this case has horrible air circulation?
25 answers Last reply
More about core runs prime side panel attached
  1. If the components are getting too hot...turn the fan speed up. I encounter a similar issue when I place my hand on the red hot burner of my oven. My hand gets too hot, so I remove it from the burner. Then it becomes cool again.
  2. your temps should not be that high with a cpu at stock clock and voltage...even at full load for hours.....eithed the heatsink is not installed right....or the voltages are higher than stock...or something with your airflow is way out of whack....
  3. You've already answered your own question. If your temps drop by that much with the side panel removed, you should seriously reconsider the cooling inside your case. Either the case itself is restricting the airflow or your components and cables are wreaking havoc with the airflow.

    For comparison, I've got a Core i7 950 in a P182 with two 120mm fans, one fan attached to an HR-01 Plus and one fan on the rear exhaust. After running Prime95 large FFTs for 15 minutes, my max core is 65C. Both fans are running at 730-760 rpms and I'm at stock clocks with undervolting. Ambient temperature is 21C.

    On the other hand, TjMax for the i7 is 100C; so if your temperatures reach 90C and stay there, "technically speaking" you should be safe. I know that if my cpu were running that hot I would not be very happy.
  4. thats much too hot it should be round 80c after a few hours prime
  5. the cmstacker 830 is a beast if u have 4x120mm fans on the side, it provides way better airflow than antec1200 but if you put that plexiglass window on, it changes it's title to one of the worst. The stacker 830's focus was to have huge airflow from the sides
  6. So, what's the point of the OP's post? You obviously have shown the problem is terrible airflow in the case. And the fix is pretty obvious too. Add more fans or buy a better case.
  7. Is the airflow through the front of this case really that bad? I'm looking at pictures of it and the front is completely mesh.
  8. Wow! OK, you clearly have an air flow problem. Your cpu cooler is getting the heat out of the cpu "OK" (mostly), but you are not getting the hot air out of the case. As the ambient temp rises in the case, so does the cpu temp once you "cool" with hot air.

    You say you have 9 fans. When I look at a Limited Edition case like here:

    I see two only fans stock. Could you tell me where you have installed each fan and what direction it blows (intake or exhaust).
  9. Thanks to everyone for their replies. I've actually never overclocked before and since Prime 95 is a torture test not typically experienced in a real-world scenario, I wasn't sure if it was standard practice for most overclockers to remove the side panel before running it. (Sounds like the answer is a definite no.) So that points to a problem with either the case not getting enough fresh air, not exhausting enough hot air, or both.

    This is actually a modified version of the Stacker 830 (like the one Twoboxer linked). Cooler Master has created some limited editions of the 830 with nice paint jobs. But as a trade-off, the front mesh is covered, and the side mesh is replaced with an acryllic window.

    I set it up with a fan in the back and at the top for exhaust. Four additional 120mm fans are mounted where you remove the side panel (but with no way to draw in fresh air). Two fans are mounted at the bottom of the case intended to pull in air through the mesh underside of the case. (Again, the case bottom is raised about an inch above the desk.) Finally there is a fan in the front to cool the hard drive bay holding three HDs, and inside the hard drive bay I have an 80mm fan mounted to the top HD so it runs the coolest.

    None of these fans are audible, but I was hoping they would prevent the inside of the case from getting too hot. (Apparently not.) In fact, only the rear exhaust is warm. The top doesn't exhaust much hot air, if at all.

    Sounds like I might have to accept this is not the ideal case for overclocking.
  10. Well, we can try something quick. Turn off the power to the four fans mounted on the side panel and to the 80mm on the HD

    Now we should have four fans, two intake at the bottom of the case, two exhaust at the top. The air flow may now be undisturbed. Test again, see if there's any improvement. (I'm a little worried about how much air can get past the 4 side fans and the vid card, too.)

    That 965 is going to run hot, especially running 8 threads at max. I honestly don't know what the "expected" temperature is.
  11. Doesn't seem to be much difference with those side fans off. Idle temperatures were the same. Ran Prime 95 again with the side panel attached and within a minute my CPU temperature had exceeded 70C; 45-50 seconds later, Core 0 hit 80C. I stopped the test after that.
  12. Well, the last longshot I can think of that has a chance of improving what you have got is to power up the 4 side fans again, but set the lower two to exhaust.

    I honestly doubt that will set up two "chambers" and yield an improvement, but it might. If it does, it would be because the 295's heat would get exhausted rather than sent up to "cool" the cpu.
  13. I bought one of the last GTX 295's with a dual PCB, thinking it would run cooler than the single PCB version since it exhausts both GPUs out the back of the case. (The single PCB exhausts one GPU out the back, and the other inside the case where the hot air just gets recycled again.)

    I might try an aftermarket tower cooler with two fans on either side in a push/pull configuration to draw hot air out of the exhaust. I'm hoping that will be a lot better than a stock cooler that's just blowing directly onto the CPU's heatsink.
  14. Oh, it will be better.

    Google for reviews of the cooler you're looking at - those that actually report numbers compared to other coolers lol. Push-pull does not in itself guarantee better cooling than a different single-fan cooler.
  15. Actually you did learn something from turning off the side fans: since the results before and after are the same, you know that those four fans aren't really doing much for the system.

    What you could do if you get a tower heatsink is fabricate a duct to direct the hot air from the cpu to the rear exhaust, similar to what Dell has done with their systems. This will definitely reduce the amount of hot air recirculating inside the case.
  16. Yeah, I was surprised how little those four fans affected temperatures (at least with the side attached). And I actually have one of those ducts from an old Dell. Maybe I can put it to some good use.
  17. One more thought: you didn't mention what motherboard you have but, if it's an ASUS like mine it could be feeding more voltage to your cpu under load.

    To show the difference I just changed my vCore back to auto in my bios and ran Prime95 large FFTs again. After 15 minutes my max core temp was 81C with an ambient temperature of 23C. My two fans also spun up to 850-880 rpm. Compare this to my undervolted temps from my previous post and you can see what a difference the vCore can make.
  18. ecmslee makes two good points.

    BTW, if you are interested in playing around some more - and with your case *I* would be interested lol - I'd go back and try putting the lower two side fans as exhaust.

    Then I'd take a piece of cardboard and use it to divide the case into two chambers. Basically by building a"wall" from the mobo to the side of the case, from the rear of the case to the drive cages, running right in between the two lower and the two upper fans. (If this works, I'd later use a sheet of modeller's plastic cut to fit with an exacto knife.)

    Two lower intake and two side exhaust fans for the lower (295) chamber, and two side intake plus two upper exhaust for the cpu area. Not perfect because you'd bring some warm air back into the cpu "chamber", but it might improve overall air flow.
  19. Interesting idea Twoboxer. But I wonder how will the side fans exhaust the air? rennervision stated that instead of mesh on the side panel, CM replaced it with acrylic. Wouldn't the fans just be blowing against the acrylic?
  20. Yeccch! I guess I read that and couldn't believe 4 fans would be mounted with their backs to solid plastic trying to blow on the mobo.

    Time to make some holes in that plastic I guess.
  21. Since the top exhaust doesn't get very warm, I've thought about changing it to an intake as well. Doesn't seem to be doing much good as an exhaust anyways.
  22. Just thought I would return with my findings in case anyone with a similar problem stumbles onto this topic. The good news is I've had some success getting my CPU to run cooler.

    First, anyone with this case that assumes the top fan should be an exhaust since hot air rises, may want to consider reversing the direction and changing it to an intake instead. I have noticed the CPU gets hotter faster when this fan is an exhaust. (Changing it to an intake doesn't make a big difference, but it certainly won't make things worse - and as I've already explained, this case needs all the fresh air it can get.)

    But the biggest help was changing the cooler. I got rid of the stock cooler and replaced it with a Nepartak S983 and separate mounting bracket for socket 1366. It's an HDT tower heatsink, not as tall as other models currently on the market. I had to go with this one since despite my case being huge, it's actually not wide enough to accommodate the other monster coolers out there. (Not too wild about those coolers which cover your RAM slots anyways.)

    Then I added a second 92mm fan to the other side of the cooler, so it now has a push/pull configuration for airflow. The air then gets sucked outside with an exhaust fan about two inches away in the rear of the case.

    So with this new setup, I'm able to run Prime 95 at the same temps with the side panel on, that I originally could only achieve with stock cooling by removing the side panel off. Load temps in the 70s may not be a whole lot to brag about - but it's a start, considering how hot this case can get. Unfortunately, Intel's stock cooler is just not designed to keep down case temps, and therefore is not very effective.

    My only question now is - can I get this to run even cooler? I applied the heatsink with Arctic Silver 5 using a number of different techniques. (A thin layer spread out over the CPU and the line method both produced horrible temps.) I got the best results following the method outlined here:

    Running Prime 95 for 20 minutes, RealTemp showed my highest temps to be as follows:

    Core 0: 77 C
    Core 1: 72 C
    Core 2: 74 C
    Core 3: 70 C

    Do you think the huge difference between core 0 and 3 is an indicator I did not seat my cooler very well?
  23. "Do you think the huge difference between core 0 and 3 is an indicator I did not seat my cooler very well? "

    Could be, but frankly your temps for running 8 threads of Prime95 are pretty damn good. The possibilities of you doing the equivalent with real world apps is slim and none.

    Converting the top fan to intake will probably continue to work well for you as long as neither the back nor the top of the case are enclosed. That way you are bringing in the coolest possible air to apply to processor cooling.

    The only issue with this is the likelihood your gpu is running hotter.

    As for applying the thermal paste, I would have followed the advice in your link, but would have tried using 4 *thinner* lines of paste than shown. Each line would be about 2/3rds as long as the processor, centered side-to-side, and spaced up-and-down so that they divided the processor into 5 equal segments. I think it makes the spreading out process easier and more certain.
  24. Hi rennervision, it's been a while since your original posts. Just thought I could add some more data for you. I ran Prime95 large ffts again and my individual max core temps are as follows:

    After about 15+ minutes-
    Core 0: 63C
    Core 1: 61C
    Core 2: 62C
    Core 3: 60C

    I use AS5 and I follow their recommended method of applying one very thin line across the cpu cores and giving a slight twist once the heatsink is mounted. My heatsink has a standard square base.

    I also agree with Twoboxer that your core temps are possibly caused by uneven contact between the IHS and heatsink. Your heatsink is mounted with spring-loaded screws right? That being the case, the pressure on the heatsink should be fairly even. Disregarding mounting pressure, the only two other factors that may be causing your difference in core temps are uneven contact and TIM application.

    I was reading a review on a heatsink that should be fairly similar to the one you have:

    One observation they made that stood out to me was that-
    "The grooved plate which hold the heatpipes is not really designed to aid heat transfer. It may be made from cast aluminum. It's not clear whether the heatpipes are just press-fitted into the grooved plate or soldered. Performance might have been improved slightly by making this plate from copper and soldering it to the heatpipes."

    I think we can agree that for this type of heatsink the most important consideration is good contact between the heatpipes and the CPU. If I were to experiment with your heatsink I would try applying a very thin line of TIM on each heatpipe. I would also make sure that I have as thin a layer as possible since too much TIM can be detrimental to the transfer of heat.
  25. Thanks ecmslee for those temps. I was curious how hot other Core i7 systems were running.

    I think 70s aren't horrible for stock speeds, but it pretty much negates the possibility of trying any serious overclocks. I'm not completely dissatisfied with the cooler I'm using, since it's obviously much more effective at exhausting heat out of the case than the Intel stock. I've thought about adding an air duct that travels directly from an intake fan to the CPU cooler, but unless I can get my Prime 95 temps in the 60s, it'll will be a lot of work for nothing. I think I'm going to settle for 3.2 GHz for now.
Ask a new question

Read More

CPUs Intel i7 Overclocking Product