762 Help

[Assassin48]

For the life of me i cant get 3d06 to finish the cpu test with out it closing down @ 4.6ghz

i was able to run 4.7ghz on the UD5, i must be missing some setting or something

any tips ?

[Neuromancer]

temps are good I assume?

You up the IOH core?

List your voltages.

I assume the classified has the option to set PWM frequency as well, have you tried that? (should not need to but figured I would ask)

[Gutterz]

Is your VTT above 200mv? To high of QPI PLL an IOH vcore causes instability.

[Assassin48]

These are my settings for 4.54ghz 3d06 benching
http://img18.imageshack.us/img18/4892/p1010756q.jpg
http://img404.imageshack.us/img404/427/p1010755s.jpg

[Kal-EL]

Sry won, silly me fell asleep after being up for 24 straight hours, I'll see what I can on the ole 762 now on air.

[Assassin48]

Sry won, silly me fell asleep after being up for 24 straight hours, I'll see what I can on the ole 762 now on air.

thanks bro,

Its weird tho because 4.6 and higher wont complete the cpu test in 06 no matter the voltage

[DrNip]

Try upping your cpu and vtt pwm frequency. Also you might try less vtt. Mine does better with less vtt. Was able to do 3D 4.6ghz runs on air at 100 and 5ghz runs on dice at 200. Couldn't do that before with more vtt. Might need IOH Vcore at 1.3 or so. I leave my QPI PLL Vcore alone because as soon as I touch it, it doesn't run for shit. I leave mine on auto or at the 1.2 setting.

[Assassin48]

Try upping your cpu and vtt pwm frequency. Also you might try less vtt. Mine does better with less vtt. Was able to do 3D 4.6ghz runs on air at 100 and 5ghz runs on dice at 200. Couldn't do that before with more vtt. Might need IOH Vcore at 1.3 or so. I leave my QPI PLL Vcore alone because as soon as I touch it, it doesn't run for shit. I leave mine on auto or at the 1.2 setting.
Thank YOU!

i managed to finish the cpu tests at 4.6ghz

going to keep going

will post pics in a sec

[Kal-EL]

The wife is sick so I had to run out for some sudafed but yah, 3D benches get funky with excess vtt. You can wprimes up and down the track with excess vtt but not 3d.

You still want me to give your settings a try Assassin?

[DrNip]

Thank YOU!

i managed to finish the cpu tests at 4.6ghz

going to keep going

will post pics in a sec


Which setting worked? The cpu/vtt pwm frequency? Too much vtt? The suspense is killing me. :D

[Gutterz]

what did you change to get the 4.6 to run?

[Kal-EL]

what did you change to get the 4.6 to run?

Just his mind.....................

[Assassin48]

Here are the settings used
http://img29.imageshack.us/img29/3356/p1010757fa.jpg

http://img20.imageshack.us/img20/8012/p1010758m.jpg
http://img20.imageshack.us/img20/3420/p1010759y.jpg

[Splave]

cpu pll can be lowered 1.5 sometimes helps me get through threaded benchies

[Neuromancer]

Yeah CPU PLL up or down is popular :)

[Dentlord]

I use cpu pll 1.425v

[Assassin48]

Ok

i tried a bunch of different voltages and i even remounted the cpu and water block twice !

and still cant go above 4.58ghz , i'm thinking of selling this cpu since it does do 4.2 @ 1.28v and use the money for another 920

[Kal-EL]

which bios were you using?

[Assassin48]

which bios were you using?

I updated to the newest one then i flashed to bios E76220N

[Neuromancer]

Phase Locked Loop; it's a circuit which gives you your multiplied clock pulse. it essentially takes your bclk and multiplies it by the multiplier you set in bios (or stock). it feeds a input waveform into a differential amplifier (correct me if i'm wrong; been a year since i did my module on PLL) then into low pass filter, leaving only the 'error' margin, this gets amplified, which is what the voltage relates to (and thus higher voltage gives an easier lock-in error margin) and then is fed back (the loop bit ) into the diff amp (looks at the difference in input A and input B [fed back waveform]) then back through the LPF & amplifier circuit to the output.

i7's seem to be contrary to the usual belief that more volts means more stable waveform, so experiment with higher and lower to see what is more stable.

Now this is really interesting. This must be coorelative to some other setting. No way should lower volts mean better performance at high bus speeds UNLESS it means more current... which could actually be dangerous... I need to research this some more..

[Assassin48]

I tried everything i possibly could to hit 4.6ghz

I'm getting an E5502 tomorrow and ill pop it in and see if i can get something out of it.

It would be nice if i had the other 920 still to see if its my chip or the board, anyone with a 4-way want to try out my chip and see if you can get 4.6 out of it ?

[Neuromancer]

Sucks that you could do it with a GB board and not with a classy... I would not be blaming the chip TBH.