The Official OC (OverClocking) Thread!

68c bad temp for the core?
so there's gonna be permanent damage if he doesn't?
and back to my original question, will Intel be able to tell what happened if the CPU were to die?
thanks,
-im1992

 

It doesn't on my gigabyte board which is admittedly a P965 chipset. Any processor activity will cause some minor voltage gains but not by that much, definitely not enough to for a high overclock. To date I have never seen an auto voltage setting that overvolts for overclocking.

 

Default settings both my P35C-DS3R and X38-DS4 ran my CPU at its default 1.325V. Auto the P35C-DS3R reported 1.39V in everest, the X38-DS4 1.41V.

 

Here is a link to someone overclock a Gigabyte board and you will note that he had to increase his vcore. I could surf some more but I'm certain enough to bet that if you push it you will need to go to manual vcore settings.

http://www.3dgameman.com/forums/showpost.php?s=e5cd54aff5a311c3e51e52ba6bed94f1&p=663456&postcount=7

 

Im sticking with manual settings unless you are doing a software overclock and not a bios overclock.

 

Here is a review from a tech site and you will note that they set their vcore manually for an X48 Gigabyte board.

http://www.guru3d.com/article/gigabyte-gax48dq6-review/6

{quoteAfter raising the core voltage of the processor too a way too high 1,65v we finally reached a stable POST. Obviously we made sure the FAN was set to maximum which made it sound like a JSF fighter on a runway takeoff.

Check temps, overclock speeds and Prime95. We measured and overclocked at a 20 Degrees C ambient room temperature.

Inevitable the temps rose to beyond 75 degrees ... not acceptable. We'll need better cooling. Point here is that the mainboard can manage the FSB, which is all we wanted to proof, yet requires a higher CPU voltage compared to other high-end specced mainboards we tested.[/quote]

 

That wouldn't surprise me - high vcore is bad news especially if people don't realise. However, that said if they manage to up to 1.41V, you can pull off a hefty overclock with that sort of voltage.

These are BIOS overclocks, not software. I don't think I'd manage 1.8 to 3.15 in software!

 

I never recommend software overclocks.

 

Yet another tech site that used manual vcore settings.

http://www.neoseeker.com/Articles/Hardware/Reviews/ga-x48-dq6/15.html

 

Yeah I never used AUTO settings. I always thought AUTO upped the voltage more than required and that gives bad temps and higher energy consumption.

Imran, 63C for CORE temps under LOAD is not too shabby considering his cooler is probably worse than the Intel stock cooler! lol Seriously, people can OC those Q9450s easier than the Q6600 as long as you don't push the FSB too much.

 

I don't really know what you're trying to prove here soph - I'm well aware you use manual volts if you want to do overclocking properly, but auto is still a viable method for getting decent results.

 

Actually I was trying to prove that you were right. I overclocked more system than I can hope to remember and I have never seen a successful overvolt on auto, but I am open to the possibility that I could be wrong.

I suspect that the advice that you gave was not accurate. You can use auto but only to a point. I've said enough in this area however and I won't pursue it any further.

 

That's fair enough, and I'm certain software overclocks are useless for scoring big numbers. However, for those who think it's useless for any overclocks, there's a 3.15Ghz E4300 telling you otherwise...

I did OC the CPU manually using my P5N-E SLI, but any real comparisons with that board are flawed because it was so terrible...

 

I've good results from every board I've ever purchased and on occassion excellent results. I was looking at a Gigabyte, MSI, and DFI for this build but fortunately I meandered into my final choice.

 

i didnt say 63
i said 68
thanks,
im1992

 

Indeed you did, and that's a bit on the hot side. Fortunately since it's realtemp it's not of huge concern, but it would be wise to check the voltage running through the CPU. At temperatures like those I wouldn't take it above about 1.28.

 

If 68C is a measure of the CPU temperature then it is quite high but if it is a core temperature then it is not too bad. Most people don't know that there is a difference between CPU temps and Core temps. The CPU temp is a measure of temperature as it is on the heat spreader but core temps sensors contact the actual core itself. If a CPU temperature is at 68C then the core temp can be as high as 70-85C or more. The T-junction max for the Q9450 is 105C. That appears to be a lot of headroom, but core temps can rise very swiftly and could pose a problem. Those types of high temps can also speed up Electromigration considerably.

 

sam,

supposedly realtemp is the most accurate core temp software for 45nm cpu's. it reports the core temps for my 45nm E3110 10c lower than OCCT, coretemp and everest.

so it is a big concern, as OCCT, coretemp and everest would all be reporting core temps in the area of 78c.

intel never officially released the T-J max for 45nm cpu's that I am aware of but I have heard it is 105c and that's according to Soph too. realtemp uses a T-J max of 105c where OCCT, coretemp and everest use 95c.

my coretemps under load reach about 65c with OCCT at an ambient room temp of 73f. OCCT uses a T-J max value of 95c so in reality and according to realcore, my core temps under load are about 55c. I would be very concerned with core temps of 68c with realtemp.

 

My bad, I don't know where I got 63C from lol.

 

Mort

The tjunction/tcasemax is coded into the CPU.

All that you need is a software that can read it such as core temp. The reason I was concerned about a temp of 68 is even if that is a CPU temp, it is darned close to the maximum for that processor.


http://processorfinder.intel.com/details.aspx?sSpec=SLAWR#