Ocing takes this long?

FATALiiTYz

FATALiiTYz

New Member
Was looking at a forum on Ocing a Q8400 just before out of curiosity. Few months ago I got my monthly issue of atomic, one awesome computer gamimg magazine, much better than those "PC authority" crap which are filled with tonnes of ads and simple stuff that I already know! Atomic has brought me to figuring out the awesomeness of computers just like this site did when i found it. Anyways my mag was saying about how easy they got to 4ghz by raising the fsb to 500mhz, Heres wat this guy said on the forum.

"When you first start overclocking you want to set the FSB/Ram ratio to the lowest this way the ram doesn't cause issues while overclocking. Then start upping the FSB and vCore in little bits and test for stability using LinX or Prime 95".

So every time I get 50mhz higher I should boot into widows and wait another how ever many minutes for Prime to finish! Wouldnt this take me hours just to get to 3.6ghz or a stable OC that doesnt cause too much heat!

One more question, answers are highly appreciated. Why would you need to lower the rams clockspeed? How could the ram "interfere"? I cant find a reason to why you would do this?

Thx for all answers!
 
Fatal1tyXx said:
Was looking at a forum on Ocing a Q8400 just before out of curiosity. Few months ago I got my monthly issue of atomic, one awesome computer gamimg magazine, much better than those "PC authority" crap which are filled with tonnes of ads and simple stuff that I already know! Atomic has brought me to figuring out the awesomeness of computers just like this site did when i found it. Anyways my mag was saying about how easy they got to 4ghz by raising the fsb to 500mhz, Heres wat this guy said on the forum.

"When you first start overclocking you want to set the FSB/Ram ratio to the lowest this way the ram doesn't cause issues while overclocking. Then start upping the FSB and vCore in little bits and test for stability using LinX or Prime 95".

So every time I get 50mhz higher I should boot into widows and wait another how ever many minutes for Prime to finish! Wouldnt this take me hours just to get to 3.6ghz or a stable OC that doesnt cause too much heat!

One more question, answers are highly appreciated. Why would you need to lower the rams clockspeed? How could the ram "interfere"? I cant find a reason to why you would do this?

Thx for all answers!
Click to expand...

When I overclocked my CPU it took me two days. I would OC it 100MHz then run prime95. Then if it was stable I would do it all over again. So depending on how long you want to run prime95 then it can take a while. I ran prime95 for 3 hours after every OC. When I got to far I got a BSOD so I lowered it a little then ran prime95 again. Finally when I decide I was done I ran prime95 for 5hours and didn't get any errors or BOSD's. Now I have already got 3 BSOD's in the past 4 days which I suspect is from overclocking but it could be anything. Also you should know when you overclock your CPU it also overclocks your RAM for example I OC'ed my CPU 400MHz it OC'ed my RAM 144MHz. So if your RAM starts causing errors or BSOD's from OCing the you will have to down clock it.
 
Last edited:
Thats why you always buy a motherboard that allows you to unlink the CPU & RAM so it only overclocks the CPU.
 
The first lesson in overclocking that many dont realize is that someone else cant tell you the exact procedure you need to use to overclock your rig. You can read other people's forum posts for guidance and for learning, but no 2 chips are going to act identical!

Also, 1 hour prime is a good start, but errors can happen way beyond that. I'm working on my i7 clocks right now and I got an error after 12 hours of blend testing. 1 more notch on the QPIv and off to prime again!
 
LOL thats one way of overclocking i guess:confused:

Or you can look online for people that have the same mobo/cpu as you and see what they overclocked too, like for me.

i looked online, a few guys overclocked there 5200 at 3.2ghz, so what i did was i went straight from 2.7 to to 2.9. Stable, so that was the big leap, i was able to reach a maximum of 3.1ghz, but after i did my big leap i did about 50mhz or so per time, because i knew eventually i would have to raise the voltage, because for my cpu the range to change the voltage is around 2.9-3.0, so just be sure of that.
 
if you don't want to invest the time, you can try automatic overclocking software. your results won't be as good usually, but if you don't like the time investment of doing it the manual way that's really your only other choice.
 
I understand that IntelBurnTest can do what Prime95 takes ~8hrs to do in 10 minutes.

I use it for my OCing at the moment. At least, my OCing is farily sub maximal. i7 at 3.6 at the moment. If I was pushing 4GHz+ I'd probably end up running Prime95 for a few hrs as a final stability test. But the IntelBurnTest is good to throw up any red flags right away without wasting time.
 
you don't want to raise voltage unless you have to for stability. The first rule in overclocking is to take small steps. The second rule is watch temps, know your maximum safe temps! The third thing you want to do is find your maximum OC on stock volts.

Here is how I went about OCing my current CPU. I first turned off all power saving and unnecissary settings in bios. Then I down clocked my RAM to eliminate instability of RAM. You don't want to mistake RAM errors for CPU errors. Then I started bumping FSB and testing with linpack test. Running like 20 passes is a good start. I found I started to lose stability with stock volts at just under 3.2Ghz. To get stable at 3.2 I had to increase Vcore slightly. Since I am now in unstable territory with stock settings I test much more rigorously. I started doing 50 passes of linpack (watch temps for at least 3 passes) and 24 hours of prime testing everything. And still take small steps. To hit 3.4 I needed to raise Vcore a little more and north bridge (MCH core) a bit. I found 3.4 a very desirable OC for 24/7 because it needed very little tweaking. In order to get 3.6 I needed to boost vcore much more than before. I also had to play with the CPU reference and termination quite a bit. I found that I actually had to lower the termination to get stable. Also my temps were borderline with linpack at 3.6 so it became my benchmarking OC rather than my 24/7 OC.

In higher OC's things get tricky because you only want to alter one bios setting at a time and check stability. But you may need to alter multiple things to actually get stability.

Also, 1 hour of prime is not a very good test for stability. And neither is 20 passes of linpack. I could often pass 20 passes of linpack and 4 hours of prime but could not do 50 passes of linpack and 24 hours of prime on the same settings.

Good luck and check out www.overclock.net for further help in OCing.
 
Cameron said:
I understand that IntelBurnTest can do what Prime95 takes ~8hrs to do in 10 minutes.

I use it for my OCing at the moment. At least, my OCing is farily sub maximal. i7 at 3.6 at the moment. If I was pushing 4GHz+ I'd probably end up running Prime95 for a few hrs as a final stability test. But the IntelBurnTest is good to throw up any red flags right away without wasting time.
Click to expand...

Yes, using multiple stress test is a must. IBT or any linpack test will stress your cpu to the max while prime will stress everything including your northbridge which is very important at higher OCs.

Edit: to test RAM stability I don't trust anything except for memtest86+ for hours....
 
When I overclocked my E8400, I did my research, read around on forums for a day or two and found out what was safe, what wasn't, etc.

Once I actually overclocked the chip, it took all of about 90 seconds including the reboot back into Windows. Some people suggest you bump up the FSB in tiny increments, but I don't think this is necessary unless you have already gotten near the threshold of what the chip is capable of, and are trying to find the limit.
 
Enjoi said:
When I overclocked my E8400, I did my research, read around on forums for a day or two and found out what was safe, what wasn't, etc.

Once I actually overclocked the chip, it took all of about 90 seconds including the reboot back into Windows. Some people suggest you bump up the FSB in tiny increments, but I don't think this is necessary unless you have already gotten near the threshold of what the chip is capable of, and are trying to find the limit.
Click to expand...

it takes 90 seconds to change bios and reboot. It takes more than a day of stress testing to know if you are stable.

The bottom line is; do you want to overclock correctly or just do a hack job? It's up to you, it's your hardware.
 
Backwoods166 said:
it takes 90 seconds to change bios and reboot. It takes more than a day of stress testing to know if you are stable.

The bottom line is; do you want to overclock correctly or just do a hack job? It's up to you, it's your hardware.
Click to expand...

Some people say you need to run stress tests for 24 hours, others say if your OC can withstand the most stressful program you will be using the computer for, usually gaming, then the OC is stable. I agree with the latter. My OC may fail at 22 hours of Orthos testing, but I can play games until my palms sweat and not have any issues...

My OC doesn't fail at 22 hours, I was just throwing that in as an example.
 
Enjoi said:
Some people say you need to run stress tests for 24 hours, others say if your OC can withstand the most stressful program you will be using the computer for, usually gaming, then the OC is stable. I agree with the latter. My OC may fail at 22 hours of Orthos testing, but I can play games until my palms sweat and not have any issues...

My OC doesn't fail at 22 hours, I was just throwing that in as an example.
Click to expand...

That is just bad advice plain and simple. If you are ok with a hack job of an OC that is fine but you should not encourage other people to do so. The steps and caution to take are well established and agreed upon in the OC community, albeit everyone is a little different by necessity. And these are established to protect people from frying their hardware and to find as stable as possible settings.

Reading around is very important but it will not tell you what settings you need specifically. Every CPU is different and requires its own OC. This becomes more important the higher the OC.

I encourage everyone to take caution when OCing. First because it is risky, and second because you probably just voided your warranty.
 
You must log in or register to reply here.
Back
Top