• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Q6600 and overclocking.... HELP!

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

souper

Registered
Joined
Oct 1, 2007
After reading these boards for the past couple days I started screwing around with overclocking my Q6600 last night. Before I get into anything, my system breakdown is as follows:

-Q6600 G0 w/ AS5 paste & Thermaltake Big Typhoon CPU cooler
-2-2GB (4GB total) of Geil DDR2-800 @ 2T 5-5-5-15 Value Series w/ Thermaltake Aluminum heatsinks
-Asus P5K
-EVGA 8800GTS OC 320MB w/ Thermaltake GPU cooler
-Enermax Noisetaker 600W PSU
-Win XP SP2 32bit

All of this sits in a Coolermaster COSMOS case, so all things considering, I feel cooling is adequate...

The first thing I learned with my limited overclocking experience is that my memory will not operate at a higher frequency than stock. I believe the recommended voltage is between 1.8V-2.2V for this memory, so even feeding them 2.2V I still cannot POST at anything higher than 400MHz. The stock timings are 5-5-5-15. Changing this to 5-5-5-12 works fine, and I was able to run Orthos for 20 minutes without any errors ( I'll run a longer test once I figure out what i'm doing ). But changing them to 4-4-4-12 will not allow me to POST. Again, no matter how much voltage I feed them. I guess this is due to the fact that they are value sticks? Or maybe i'm doing something wrong, or could it even be due to the amount of ram that I have? There's also the issue with a 32bit OS not supporting more than 3-4gb (in my case, 3.25gb). For the record, I don't mind leaving it at 400Mhz @ 5-5-5-12. I just think it's odd that I can't get it to OC beyond changing 15 to 12 (I apologize for not knowing the proper vocabulary when talking timings - i'm still new!). My goal is to reach a stable OC with my FSB and RAM operating 1:1.

Before I started screwing around with my memory, I adjusted most of my BIOS settings found in the 'Sticky' C2Q OC thread on this forum. Then I started to play with my CPU freq, memory freq/timings and all associated voltages. I couldn't for the life of my find a stable OC and I don't know why.

I eventually set all my settings back to default and changed only a few items that resulted in a semi-stable OC. The settings I changed are as follows:

C1E - disabled
Speedstep - disabled

(I disabled the above two so that CPU-z would reflect accurate frequencies)

CPU Multi - 8x
CPU Freq - 400Mhz
RAM Freq - 400Mhz
RAM Timings fixed to 2T 5-5-5-12
PCIe - 100Mhz

Everything else I kept to auto. Including all voltages. I don't know if this is good or bad. I understand that you always want your hardware to operate at the lowest voltage possible, and setting it to auto may provide more voltage than needed. Is this true? Is it bad to leave the voltages on auto?

With the configuration above I can run a small orthos test with no errors. This is fine, but for some reason orthos is detecting my CPU to be operating at 266Mhz with a multi of 13.5x (3.6Ghz). Not only that, but my device manager also detects 3.6, as well as Intels TAT program. The only program that seems to be detecting the correct frequencies is CPU-Z.

oc.gif


Why is this happening? My CPU voltage is very high too for a OC that should be only 3.2Ghz. I've been hearing stories that the G0 version of this processor can operate at stock voltage up to this speed (or close). Could this be conflicting with my memory some how? Causing them to not being able to OC at all? Any help would be greatly appreciated. Thanks!
 
Well Orthos is crazy when it comes to CPU freq. I would trust CPU-Z. I am pretty sure that Windows is also about as crazy. My one tip would to be manually set your chipset voltages, vcore, and memory voltage. If your value memory does not have heatspreaders on it I would be careful feeding them 2.2+ volts. If it does have heat spreaders anything beyond 2.2v should have a fan blowing on it. It looks good at 3.2 GHz so long it is stable.

Edit: Just noticed you have ram sinks on your memory. 2.2v should be fine. If you go higher then hang a fan over them.

Edit2: I would suggest testing your memory at your current settings with Memtest86. Get the bootable ISO and boot it from a CD. It can be had here: http://www.memtest86.com/download.html
 
You might want to use Prime for stress testing as it will test 4 cores, orthos only supports 2 (you would need to start up another instance and set affinity etc). Also use CoreTemp for your temp. monitoring program since TAT only reads 2 cores.
 
Good advice from the others above. Orthos usually displays what your theoretical cpu speed would be given the FSB you are using and the maximum multiplier that your chip can use. So, if I am running at an fsb of 500 with a multi of 7 on my E6600, Orthos will display 500*9 which is 4.5GHz (obviously incorrect!).
 
Thanks for the help guys. I'll give it another go tonight based on your comments. I think i'm gonna focus on finding the proper voltages for my configuration,
 
Just as a comparison, I'm only able to get 3186 (9x354) out of my G0 6600 @ 1.275v. I try 3.2 at up to 1.4v and no go. I gots a 1mhz window. One it's rock solid, the next it is not, I always get at least 1 core dropping out.
 
It seems I can't get my CPU to operate stable at anything beyond 3Ghz @ 1.35V which seems high (volts) to me. Maybe cause i'm using a low multi and high FSB? I wonder if using 9x333 would allow less of a voltage than 7*428.

I was able to OC my memory to 450MHz @ 6-5-5-14 2T @ 1.8V and ran Memtest for 1 hour without any errors.

I also ran Prime95 for 6 hours, no errors, with my CPU @ 3Ghz @ 1.35V with my memory at stock freq. & timings (400Mhz & 5-5-5-15) @ 2.2V.

I also ran Prime95 for about 15 minutes with this configuration:

7x428 = 3Ghz @ 1.35V, memory @ 428Mhz @ 6-5-5-14 @ 2.2V. Does this seem like a good OC? I've been browsing this forum and noticed that not too many others go lower than 8 multi. Is there a reason for this? EDIT:Actually, running Prime95 at these settings caused XP to lock up after 1 minute. I forgot... I wonder why memtest would seem fine @ 1.8V but Prime wasn't @ 2.2V...

At this time i'm still feel like i'm at ground zero lol. I gotta go home tonight again and play around some more.

I wish I could take the damn thing into work so I can play with it all day.
 
Last edited:
Give it some volts man!

I have an almost identical CPU (x3210), and it likes voltage. I'm at 1.488v (as seen in CPU-Z), and at 3.6GHz. I'm stable and temps are around 70 under full P95 load.
 
haha, ok - i'll try. It's just that i'm already reaching 62C degrees at 1.35-1.375V and that was my innitial 'limit'. I guess it won't fry if I try 1.4V and check the temps.
 
Should I only be concerned with CPU Vcore or should I be pumping other voltages too?
 
Should I only be concerned with CPU Vcore or should I be pumping other voltages too?

It depends. You will probably be fine until you get to 420 or so FSB. The way I check is to drop my multi to 6 or 7, so that my CPU freq isn't factoring in so much. Then start raising the FSB, and test for stability. If it isn't stable then start playing w/ the other voltages and reference settings. I can see where you might have a problem w/ your RAM in this scenario, though, being as how it is value RAM. But, you did get it up to 450, so you might be fine.

If you are completely stable at 400FSB then if I were you I'd put the FSB at 400, the RAM 1:1 w/ timings you know work at 800MHz & voltage you know works at this speed. Put your multi at 9, and keep adding vCore until it boots at 3.6GHz. Don't put more than 1.55v. If it's too hot for you then back it down. I think 70s or even 80s is fine for testing purposes. You can always get a better cooler (U-120-E) later.
 
Do as jason said and see if your FSB (northbridge) is crapping out. If so then it needs more voltage. Personally I would set my multi low (5x) and my RAM as low as possible then find your max FSB. You know your max RAM so you are good there. Now that you know these two will not limit you at a given speed work the CPU up.
 
Not sure of the CFM for the fans.... but the case comes with four 120mm fans... I would think they are 45-50cfm+ each....

I wasn't able to spend much time tinkering last night as my router crapped out and I had to replace and set another one up.

I was unable to boot into windows with 9x400 @ 1.4+V so I gave up for a little... then went back and changed it to 7x430 @ 1.325V 1:1 and was able to run prime for 15 minutes with no problem. At that point I shutdown and decided I was done and would run an all day Prime95 test while I was at work today (while sleeping I wanted to run memtest one more time to make sure my RAM was ok). But for some reason this morning I couldn't run Prime95 for more than 2 minutes without getting errors. *shrugs* It was 6-8 degrees warmer in the room this morning if that makes a big deal.

This is all with my FSB volt @ 1.45. Ram @ 5-5-5-15 @ 2.2V.

I set it to 9x333 @ 1.35V 1:1 w/ memory at 2.2V before leaving this morning with Prime95 running. Hopefully i'll go home to good news.... if It passes maybe I'll try lowering the voltage a little.
 
Back