Couple of questions for the experts out there

Associate
Joined
8 Sep 2008
Posts
1,242
Location
Birmingham
Hey everyone. Im fairly new to overclocking. My first and current system was purchased last October from overclockers and came pre-overclocked (system spec: [email protected] GHz, artic 7 freezer cpu cooler, 4GB corsair dominator 1007Mhz, 500GB 7,200rpm Seagate barracuda (used to run vista x64 ultimate os), 7,200rpm 1TB Samsung spinpoint (750GB storage, 250GB partition with windows 7 RC), corsair 650W power supply, abit quadGT motherboard, Gainward GTX280 1GB, antec 1200 case)

I recently had to return this to overclockers.uk as there was a powersupply fault which resulted in the loss of the overclock. When i got my machine back they had redone the overclock but I noticed the temperatures were idling at 45 degrees whereas the pre-return overclock was idling at 35 degrees. A 10 degree difference was pretty huge in my opinion.

Looking at the voltage I could see why. 1.47V for a 3.02Ghz overclock seemed a bit high so after some research I read that if I changed my FSB:RAM ratio to 1.1.0 and changed the FSB to 333 MHz I could roughly get the same speed (3.00Ghz) with standard voltage.

So now my system is running at 1.32V, 3.0GHz CPU, 1000MHz RAM speed and it seems faster and more stable then the overclockers “professional” attempt. Tested it for 30 mins with prime95 and with everest ultimate ver 4.6 stability test with no issues (plays fallout 3 and dawn of war 2 flawlessly and with high fps)

I have 2 questions for anyone who would be kind enough to answer them

1) I just wanted to know if by leaving intel speedstep enabled (the voltage changes frequently 1.19v-1.32v at x6 and x9 multipliers respectively (as measured using everest ultimate sidebar). Will this have a negative effect on my current setup? (I just wanted to save some electricity and unnecessary heat)

2) If I installed another 4GB of RAM onto the abit quad GT mobo would that affect the current overclock I have set on the machine or have any other negative impacts on the system?

Thanks in advance
 
Hi and welcome to the forums.

1 If the system is stable leaving speedstep on wont do any harm.

2 If you add another 4gb of ram, youll probably have to up the north bridge voltage a bit, as running all memory slots puts a bit more strain on the NB.
 
So now my system is running at 1.32V, 3.0GHz CPU, 1000MHz RAM speed and it seems faster and more stable then the overclockers “professional” attempt
Hey icedbullet,

I echo setters welcome to OcUK forums! :)

A 3.0GHz Quad-Core with 1GHz ram is a good place to be, can't imagine that struggling with anything in the near future. Once you feel the system is rock-solid it may be worth trying to squeeze a few hundred extra MHz from your chip if you feel you need a little extra bump!

I suppose its a good thing that OcUK are selling pre-clocked systems but as with anything *mass* produced it will most likely never be as good as a custom overclock that you worked out yourself, nice fast system and the sense of achievment of performing the overclock yourself.

Enhanced Intel Speedstep (EIST) is a pretty good feature but it seems that the many LGA775 overclocking guides deem it a bad feature and therefore it's deemed best that its disabled. I think personally that a lot of folks never tested this themselves and just disable it because everyone else (almost) disables it :confused:

I've never had a problem with it myself and generally have it running. There is some good sense in switching it off during your initial overclocking and testing period then once you know everything is good and stable you can turn EIST back on with hopefully no undesirable side-effects.

I did observe that its better to have EIST *disabled* while benching though, the tiny periods where the processor slips back to a slower speed have a tiny impact on bench numbers although not really visable in the real world! :cool:
 
Thanks for the warm welcome Setter and Bigwayne :)

@Bigwayne. Yeah Ive read around the speedstep issues on other overclocking forums and generally people do seem to steer clear of having it enabled when overclocking. What you said makes sense and I have been using speedstep with my overclock with no real stability issues as far as I can tell. The only thing I have noticed is that my everest sidebar monitor sometimes shows the cpu drop @1999 MHz (when it should be 2000MHz) when the spedstep kicks in.

Does this seem like an indication of instability? (prime 95 & everest stress tests both ran for 30 mins, both ran fine). Maybe I should try a more prolonged stress test?
 
if it drops to a 6x multi then it should be 1998 shouldn't it? +/- 1mhz isn't something to worry about, it's probably just the speed the software is measuring.

I think you should have a safe overclock there, I run my Q6600@ 3.6 with 1.37V & its rock solid at that, 3.0 is "almost" a given with a half decent Q6600. I say almost because there are always some aren't there :)

Welcome to the forums, it is the place to get information & help. This is my 1st real venture into overclocking & this place has been invaluable for me.
 
Thanks for the welcome BeeP :)

Ive decided to keep speedstep off for now. Im still attempting to get a higher clock speed but i refuse to raise the voltage from stock settings.

Moronically I set my clock to give 3.4GHz, pc wouldnt even boot :P so i cleared cmos and went for something more realistic. I got the clock up to 3.2 GHz but it failed prime 95 (on only one of the cores). Dropped it to 3.1 GHz but same prime 95 error occured. Looks like 3.0 GHz is the highest stable speed im gonna get from stock voltages, im quite happy with that unless anyone has any ideas how to get a faster speed? (CPU voltage from everest says 1.32-1.33 without speedstep enabled)
 
Back
Top Bottom