Simple law of physics.....the power consumed by your proc will be released as heat

cpu power consumption = TDP X (overclocked speed/stock speed) X (overclocked vcore/stock vcore)^2

note that for vcore its squared...so the effect is much more powerful

ok.....so assuming you overclocked your proc from 2.0ghz to 2.4ghz(TDP=65W)

overclocked power consumption = 65 X (2400/2000) X (1.25/1.25)^2

= 78W

Which isnt that much of an increase, proly u'll notice your temps rise by 2 or 3 degrees depending on your cooler.

now.....say you have a nice windsor 3800+ F3 and oc to 3.0ghz with stock volts

overclocked power consumption = 65 X (3000/2000) X (1.25/1.25)^2
= 97.5W

but if you have a brisbane 4000+ and oc to 3.0ghz with 1.4v instead

overclocked power consumption = 65 X (3000/2100) X (1.4/1.25)^2
= 116.5W

As shown, raising vcore has a much larger impact on power consumption(hence heat released) than raising clockspeed. Hope my long winded explanation answers all doubts

My final contribution to AMD overclocking . bro pws plz updated my score yea. thnx



AMD Sempron LE-1100 | joe_star | Biostar TF560 A2+ | Team Value RAM DDR2 667 | 782Mhz | 5-5-5-15-18-2T | Tuniq Tower

Main AMD rig sold off ady, darkside too tempting

Probably never.....unless see big improvement from AMD. 3 years be their fan and evytime kena ffk only

1st time....when they totally cut off s939, that time I had just bought my X2 3800+ 4+ months ago
2nd time....after so long still suck so bad compared to c2d (EDIT: despite so many broken promises)

Am having fun with my pentium now. Can reach 3.6 but gonna need water to keep it cool

Hehe my ex proc spotted. Try to break my score of 3012mhz bro

This post has been edited by joe_star: Mar 27 2008, 12:53 PM

No i wont really. My main frustration was the s939 to AM2 socket change. After just 4 months a high-end system became obsolete

Currently is just pwnage frm intel. Considering my upgrade just cost me RM130, I tink its still reasonable

I dont tink you get my point bro. When i bought my old rig 2 years + back, s939 was looking good and looked like it still had future, but then AMD totally abandoned the platform when they switched to AM2, leaving me with few upgrade options.

Agreed....i still prefer AMD C n Q. Intels speedstep voltage cant really be customized much, unlike my C n Q setup with crystalcpuID lastime. That doesnt work with intel proc. Haih....wonder how my electric bill will look like this month

Now u tell......2 weeks after i join darkside

Good thing come too late . Hmm but maybe can get it for my backup rig. Abit NF M2-s got no voltage controls so this proc will be perfect

As you loosen the timing your memory access latency will increase, thus negating the effect of increasing the memory clockspeed. So yes, probably running it at 1066mhz with loose timing is similar to running it at 800 with tight timing. You can check out the memory oc-ing thread in the guides forum for more info

Any reports in from the new F3s??

try running small fft test bro, see wether its the prcessor or the ram. From my experience 1.5v is enuff for 2.97ghz.

Quite high. I used to increase the volts high on my mem lastime too, but then I realized that sometimes it actually makes the system unstable. Hmm......maybe try increase vcore abit too. But lastime i was fully stable at 3.0ghz with 1.52V

hmm.....is it that the other applications might cause orthos to have an abnormal error not related to oc-ing?

or

does heavy multi-tasking increase the chance of an error occuring?

If its number 2 then i say its a sign of instability.....and the oc needs to be re-tweaked, because there are often real-life situations that i myself do stuff like that, like encoding video+bt+IM+surfing all at the same time. Is my train of thought correct, or can anyone correct me if i'm wrong

He could be reaching the limits of his current cooling method as well.

1.7v is very high bro, even with tuniq tower......

Then again.....im pushing the same limits with my current proc too

Limit of your cooling also very likely . Lol.....lastime I had brisbane 3600+ which did 3.1ghz at 1.6v

now joined darkside, maximum reached 3.62ghz on 1.68v

You let it ru without running any program? I dont get that part I think probably its the mobo, usually when psu KO there wont be any reaction

I agree with you bro. I have read somewhere b4 that the actual value of the rail is safe between +/-5% of the actual value (11.4V-12.6V). However, the actual drop in voltage when going from idle to under load is more important, and is best if its 0.05v or less.

Dodgy integrated sound from personal and experience of 2 frens(all 3 of us experienced problem where sound comes out only on 1 speaker, 2 with MSI K8N Neo4 Platinum, 1 uknown). Quite slow RMA also, but on the other hand, quite understanding. They repaired my mobo eventhough it had a scratch

This post has been edited by joe_star: Apr 14 2008, 08:55 PM

Wow.....very very high bro. Although I think mobos can now support it, it is still better to keep the HT at around 1000 for guaranteed stability. Moreover the HT speed doesnt really impact performance either. In your cpu tweaking page in the bios, there should be an aoption to change the htt link. The options available range from 1X-5X

Eh......didnt know blend test was 1 of the requirements?? Sienz la....lastime i tested my sempron with small fft only coz i was targetting cpu stability

This post has been edited by joe_star: Apr 20 2008, 02:22 AM