Random: anyone else use this method to OC?

Xaeryan · Apr 25, 2008

[EDIT: thought I was in the 'General CPU' section when I posted, sorry!]

Anyone use the same (possibly backwards) method of overclocking as I have adopted? I like to set a relatively safe goal, say 3.2GHz on my 2.4GHz C2D. Then I work from 'fail to success' instead of from 'success to fail'.
Example:
1. Set my 'safe' goal speed, but with a low voltage.
2. It fails POST... GOOD

3. Bump voltage and adjust other settings as necessary until it POSTs.
4. It fails to boot Windows... FINE
5. Repeat step 3 until Windows boots
6. It fails Prime immediately
7. Repeat step 3 until Prime can start
8. It fails Prime after several hours
9. Repeat step 3 until Prime can run for many hours
10. Add a bit of margin room and continue testing all apps/games.
Of course, temperature monitoring in between all steps.

This is probably less-preferred than the 'creep-up' method; getting something stable then *slowly* increasing speeds and adjusting settings as you go. But it sure is quicker for me, and in my older age I don't have the testing patience that I once did. Results! NOW! lol

So, anyone else do this? Or am I crazy in my methods? :bday:

Nebulous · Apr 25, 2008

Well evryone has their own methods. I do the opposite which takes less work

As long as your method works for you, that's all that matters now, doesn't it?

Xaeryan · Apr 25, 2008

Nebulous said:
Well evryone has their own methods. I do the opposite which takes less work

As long as your method works for you, that's all that matters now, doesn't it?

Oh absolutely, just curious as to how others do it. :santa:

VinnyTAMU · Apr 25, 2008

Xaeryan, I use your method when I have a pretty good idea about what the CPU should clock to. Like when I bought my E8400 I knew it would most likely it would do 4GHz so I just set 4GHz right off the bat and adjusted the voltage till I got a 12+ hour stable orthos.

Albuquerque · Apr 25, 2008

I take a far longer route typically...

Start with stock speeds and find the lowest stable volts I can. And then work up from there. All at the same time, I key in my data to Excel to generate a chart of volts versus clockspeed so I get an idea of where the chip starts to cross over the line between sensical linear voltage adjustment versus non-linear voltage adjustment.

My E8400 would do stock speed at 0.9125v, and was pretty much a straight line at all overclocking vs voltage points until around 4.1ghz where the volts suddenly needed to move a little bit more for the associated clockspeed -- until again at around 4.3Ghz where the volts then needed to move a LOT more for the associated clockspeed.

The Q9450 will do stock speed at 0.925v, and needs 1.275 to do 3.6Ghz. Unfortunately, the stopping point on that processor wasn't voltage

ratbuddy · Apr 25, 2008

Xaeryan said:
[EDIT: thought I was in the 'General CPU' section when I posted, sorry!]

Anyone use the same (possibly backwards) method of overclocking as I have adopted? I like to set a relatively safe goal, say 3.2GHz on my 2.4GHz C2D. Then I work from 'fail to success' instead of from 'success to fail'.
Example:
1. Set my 'safe' goal speed, but with a low voltage.
2. It fails POST... GOOD
3. Bump voltage and adjust other settings as necessary until it POSTs.
4. It fails to boot Windows... FINE
5. Repeat step 3 until Windows boots
6. It fails Prime immediately
7. Repeat step 3 until Prime can start
8. It fails Prime after several hours
9. Repeat step 3 until Prime can run for many hours
10. Add a bit of margin room and continue testing all apps/games.
Of course, temperature monitoring in between all steps.

This is probably less-preferred than the 'creep-up' method; getting something stable then *slowly* increasing speeds and adjusting settings as you go. But it sure is quicker for me, and in my older age I don't have the testing patience that I once did. Results! NOW! lol

So, anyone else do this? Or am I crazy in my methods?

That's how I do it too

Never bothered mentioning it cause I'm sort of embarrassed I don't OC the 'right' way

smokie mcpott · Apr 25, 2008

lol, i do it the same way. i go for a goal that i know i can hit, work my voltages up to it, and after its stable, i try going higher

Omsion · Apr 25, 2008

Same here, lol.
Works well if your first guess is good, saves plenty of time

moto211 · Apr 25, 2008

i do it a little different. i set my absolute maximum voltage that i'm willing to run on all my settings (CPU, RAM, NB, SB, CPU VTT). then, i ramp up the clockspeed as high as i can get it at those voltages. i then fine tune and back down any voltages that can be reduced. it allows me to very quickly find max clocks.

adelphia83 · Apr 25, 2008

smokie mcpott said:
lol, i do it the same way. i go for a goal that i know i can hit, work my voltages up to it, and after its stable, i try going higher

Same here. I've been doing this for a long time, and found out the hard way that starting at stock is a waste of time. Every chip has at least *some* potential and by the time I buy a new CPU, it's usually known what the average speed is.

For example I just put a new Q6600 CPU in today. I immediately started at 3.3, and surprisingly found that it wasn't stable. So it was necessary to tweak things, and I'm now priming at 3.5 and it seems solid.

We're not in the K6-2 days anymore where we had 25-50mhz increments and had to start at the bottom and work our way up the ladder. Today's chips will often get you into Windows at > 200mhz what they're actually Prime-tested capable of. This makes setting the bar high easy, and sometimes I find it's less time consuming working your way down to stability.

Th3 F4ll3n P1mp · Apr 25, 2008

I just bump everything to a safe maximum, and slowly go down. Works wonders. In less than a week, I am running 3.6 ghz @ 1.4325 volts on my quad, and 1050 mhz @ 2.15 @ 4-4-4-12 on 4 gigs of tracers.

Perfectly stable.

Random: anyone else use this method to OC?

Xaeryan

Member

Nebulous

Dreadnought Class Senior

Xaeryan

Member

VinnyTAMU

Member

Albuquerque

Member

ratbuddy

Member

smokie mcpott

Member

Omsion

Member

moto211

Member

adelphia83

Member

Th3 F4ll3n P1mp

Disabled

Similar threads