Sign in with
Sign up | Sign in
Your question

2600K + Archon: Temps going from 42-52 and I'm worried

Last response: in Overclocking
Share
August 5, 2011 10:08:23 PM

This is a common question found on this site but that's because we are worried for our PC and proper cooling.

I have a 2600K I7 CPU that is tweaked to be overclocked to go to 4.42,
P8Z68 Deluxe motherboard,
Nzxt Phantom,
all four Ram slots are filled with 4GB or memory making it 16GB total,
and the CPU Cooling is a Thermalright Archon with the one stock fan that came with it.
the room the computer is placed in is warm in the area of 70.

Usually I see the temps be around 35 usually with a high of 42 when idle or doing small things on my computer (looking at pictures, writing a letter, handling some programs, ect). Well today while online watching some youtube videos and having a program called Picasa located, obtain, and scan pictures from my two hard drives. During that time according to AI Suite 2 software the CPU went around 49-60 averaging around the mid 50's in temperatures. After Picasa was done the temps went slow going down and is right now 49.

Second I have the case open so I could feel the heatsink of the Archon to see if it heats up. The heatsink on my graphics card heats up greatly to the point you could burn your fingers if left for to long however I don't feel the same heat from the heatsink for the Archon. I don't feel any heat from the Archon heatsink.

Is the temperatures I see normal for my 2600K CPU that's overclocked to 4.42, Cooling and programs running or is it high and what do people recommended? Also shouldn't I feel a good amount of heat coming from the Archon heatsink like I feel on my graphics card (I'm afraid I messed up the installing of the cooling and thermal paste which was an Arctic Sliver I think)?

I understand these question are common and annoying but I want to make sure I'm not hurting my computer and all the parts was installed correctly.
August 5, 2011 10:19:32 PM

Haven't you checked if your OC is stable by using Prime95 or anything like this? And haven't you checked temp. of your cpu during that operation? What voltage are you using? My i7-2600k@5ghz was maxing at 90 something using archon. At 4.8GHz and much lower voltage I max out at 80. All these temps are during prime95, of course. And check that temp with Realtemp. Some say to stay below 75 degrees, I'd say stay below or at 80 during prime95 and you'll be fine. Good luck.
August 6, 2011 2:26:14 AM

Are you sure you applied your thermal paste properly? Also be sure to check if the cooler and all case fans are working as they should. All it takes is a malfunction in just one and your temps can be royally screwed.
Related resources
August 6, 2011 4:11:42 PM

I followed the instructions to applying the arctic silver thermal paste applying a small rice like on the CPU and another on the heatsink and applying pressure and spreading the paste. The fans are spinning and looks like they are working.

I will use Prime95 torture test Small FFT option (or should I pick another option). I downloaded Realtemp and which should I use:
RealTempGT (This one is what I'm planning to use and doesn't look different from the other RealTemp program)
RealTemp
i7TurboGT

Also how long should the test go for?

The CPU voltage according to AI Suite 2 is set appears to be 1.220 but when idle it appears on my AI Suite 2 sensor to be much lower at going usually to 1.000 to 1.100. Also in AI Suite in EPU area on the side it has "Current CPU Power" going from 26 to 34 Watts and back (I'm not sure what this means).

After I run the Prime95 with RealTemp GT and see what happens and post what I find.
August 6, 2011 4:49:36 PM

Looks like I'm OK as my CPU was stable during the Prime95 Test with no errors or problems. During the test it maxed at 70 degrees. I also put my hand on the heatsink and felt a slight warmth to it not at all to the same heat from my graphics card heatsink but I guess that's how it should be with good cooling maybe and the fact my graphics card is not high quality and thus produces more heat for the heatsink to have. The graphics card I'm using is this one: http://www.newegg.com/Product/Product.aspx?Item=N82E168... ASUS EAH5450 SILENT/DI/512MD2(LP) Radeon HD 5450 512MB 64-bit DDR2 PCI Express 2.1 x16 HDCP Ready Low Profile Ready Video Card

The card is a weak graphics card but ok for what I'm currently doing (not a gamer on my computer so I didn't need a high end graphics card and this card is doing ok for my needs but the heat it produces is alot and I read online that nowadays people need at least 128-bit for standard stuff on the computer so I somewhat regret not getting a slightly more advance but still cheap graphics card with a fan and 128bit at least as if I pay for a graphics card fan now it just would add the cost overall for the card which I could of applied instead to a better card plus slightly future proof my system should I ever get a Blu-Ray player/burner and want to watch/burn blu-rays on my computer)

I guessing as far as my CPU and CPU heating goes things look fine (unless I'm wrong and if so please inform me). I might get the second fan to afford a push pull config for the Archon but I'm in no rush if it won't make a big difference and my temps are fine (plus the NZXT Phantom has a fan pulling air out in the back already which helps my system.

Any other recommendations, tests, teaks, and possibly what I might want to buy for the case, RAM (using G.Skill Sniper 4GB in all four slots), Cards (like my graphics card or should I in the future just replace the card itself that has a fan next time), and cpu overall?
August 6, 2011 7:34:36 PM

How long are you running Prime95? When it's stable for 2-3hours you might want to test for 12 or 24 hours. You just need to start realtemp.exe. Use prime95 default options, the ones you get when you open the program. To check cpus voltage use Cpu Z. If you really have 1.2V at 4.4GHz then you are lucky to have such cpu. Really, I wish I had it. And it's pointless to buy another fan to put it on archon, temp stays the same or drops 1degree, I checked myself. If you're not a gamer, don't need much gpu power I'd stay with what you have now or even use Intel's gpu that you have inside your cpu.
August 6, 2011 8:13:48 PM

Asus board? Update to the latest BIOS (non-beta!). Made my CPU temps drop by 10C.
August 10, 2011 12:08:46 AM

Here are my findings after running Prime95

First before updating to the latest Bio I ran the Prime95 torture test in "Blend" (testing both CPU and RAM at the same time).

The Voltage was mostly according to CPU Z: 1.280 and the CPU speed was 44xx (forgetting what the two numbers in xx was)
On RealTemp their was four temperatures due to the 2600K's four cores (with 8 hyperthreads thus 8 workers running in Prime95 with the CPU load at 100%) the average temperatures was in the lower 70's. In 3 hours and 30 minutes the highest that the four temperatures shown was 75. However at the 3 hour and 35 minutes mark about worker 3 and worker 5 had a "Fatal Error: Rounding was 0.5, expected less than 0.4" and thus was stopped. Shortly I stopped the rest manually.
After I updated the ASUS bios to the latest upgrade and some treaks I ran the test again this time for 25 hours
The results:

The voltage was higher at 1.351V
Speed: 4429.92 MHz
Highest after 25 hours for the four temperatures was 77 however when the test first hours began the temps average in the low 70's then in the last few hours of the test mid 60's in temperatures
All 8 workers was going fine with the CPU load at 100% till mid-way to the blend torture test two of the eight workers (number 6 and 7) had a "Fatal Error: Rounding was 0.45, expected less than 0.4" with one going mid-way and the other following an hour later and thus the CPU load decreased to 85% when the two workers stopped but kept the same CPU speed 4429.92 all though the test. After the 25 hours was done I stopped the rest of the workers with the temperatures not rising past 70.

Is this good or should I be worried about that two out of 8 of the workers stopped during the 25 hour Torture Blend test?
August 19, 2011 4:47:08 PM

Bump up voltage by only one measure. You'll have something like 1,355. That will do, cause you are facing a vdropp which sometimes is big enough to make a core make errors. Nothing to worry about, really.
!