decided to get volt-happy with my qx9650

[JAG87]

ok not thaaaaat happy, but more than my usual :p:

pardon the resolution :rolleyes:

http://img155.imageshack.us/img155/3674/45sk2.jpg

[Leeghoofd]

Well I wish to go 400FSB with 11 multi, as soon as I get my hands on it, reseller told me it will be 15th of january before I get mine pppppffff... can you also achive 4.5 with a higher FSB and lower Vcore ?

[Hornet331]

...45°C at idel.

i hope you enjoy your 70+°C @ load. :)

[JAG87]

Well I wish to go 400FSB with 11 multi, as soon as I get my hands on it, reseller told me it will be 15th of january before I get mine pppppffff... can you also achive 4.5 with a higher FSB and lower Vcore ?

yea i can go a lot higher with FSB. my 24/7 is 400 FSB, and I can easily hit 450-460.

400 x 11 wont be easy my friend, lower your goals a tad. this is no where near 24/7 stable its just suicide shots. I can get 4400 with 1.47 vcore, and its pi stable, but it will BSOD as soon as I do anything multi threaded. the most I have been able to do is 4200 with 1.45v prime stable. but the increased volts + temps do not justify it over 4000 with 1.33v prime stable. you will find that above 4 ghz will need a LOT of volts to be 24/7 stable, unless you like BSOD every time..



...45°C at idel.

i hope you enjoy your 70+°C @ load. :)

haha no way sir, I wont even dare.. :)

[Leeghoofd]

well 24/7 I wanna run 3.8-4ghz with low volts and 4.4 or sorts off for some benches and fun, though indeed I'm already a warned man as I read more deaths on the QX9650 then any conroe yet...

[Hornet331]

well, no wonder ppl give them insane voltages. :p:

intel recommends max of 1.45V for 45nm cpus while they recommend a max of 1.55V for the 65nm ones. ;)

and when i see ppl give 1.6V+ to this cpu.... no wonder they die so fast. :D

[h@RRy]

that doesn't account for vdroop.

[Hornet331]

that doesn't account for vdroop.

actually vdroop makes it worse.
People give more voltage to compensate Vdroop and when the cpu switches from load to idel they get even higher voltage overshoot. :yepp:

[E.R]

Is 1.475V in BIOS and 1.456V read by CPU-z 1 considered "safe" vcore on water or in dangerzone for this puppy? 1,45V in BIOS does not do 4.2GHz OCCT-stable for me yet. I know i read a thread somewere about this but can´t find it now.

[nick2crete]

actually vdroop makes it worse.
People give more voltage to compensate Vdroop and when the cpu switches from load to idel they get even higher voltage overshoot. :yepp:

Thats the truth ..

..but ..idle ..:)

[LowRun]

Is it that much unstable that it wouldn't even let you resize that giant pic? :D

[luihed]

My 24/7 setting on my watercooler and not a glitch ( just over a week ), max spi is 4850mhz also on water.... Max cpuz valid (http://valid.x86-secret.com/show_oc.php?id=288218)

http://ocidb.com/albums/userpics/10062/normal_sg.jpg

[JAG87]

actually vdroop makes it worse.
People give more voltage to compensate Vdroop and when the cpu switches from load to idel they get even higher voltage overshoot. :yepp:

thats not true, vdroop is designed exactly to fix that. your overshoot will never be higher than the voltage set in bios. so for example, if you set 1.55v in bios to get 1.52 idle and 1.47 load, your overshoot when going from load to idle will be 1.55 max.

on the other hand, when you use methods to get around vdroop, like mods or bios options, then your overshoot can be dangerous because you dont really know how high its going. but I still believe that working with no vdroop is better, because the amount of time you are going from load to idle (milliseconds, if not microseconds) will always be less than your actual idle time, and when you work without vdroop you can keep an overall lower idle voltage. to make a comparison, id rather take 110v into me for a split second, then be constantly hooked into 55v :D

Is 1.475V in BIOS and 1.456V read by CPU-z 1 considered "safe" vcore on water or in dangerzone for this puppy? 1,45V in BIOS does not do 4.2GHz OCCT-stable for me yet. I know i read a thread somewere about this but can´t find it now.

I wouldn't recommend it. Anything above 1.4v is really excessive for these 45nm. People dont realize it, they are used to 65nm voltages, but 1.4v generates the same heat that 1.5v did in the 65nm. You gotta be careful.

[Ace-a-Rue]

well, no wonder ppl give them insane voltages. :p:

intel recommends max of 1.45V for 45nm cpus while they recommend a max of 1.55V for the 65nm ones. ;)

and when i see ppl give 1.6V+ to this cpu.... no wonder they die so fast. :D

which guys have you seen having 9650's die at 1.6v+?...i think NONE!;)

[irev210]

well, no wonder ppl give them insane voltages. :p:

intel recommends max of 1.45V for 45nm cpus while they recommend a max of 1.55V for the 65nm ones. ;)

and when i see ppl give 1.6V+ to this cpu.... no wonder they die so fast. :D

intel core CPU's die fast? Thats news to me, lol.


Intel has made some extremely durable processors.... hard to say they die fast.

[safan80]

well 24/7 I wanna run 3.8-4ghz with low volts and 4.4 or sorts off for some benches and fun, though indeed I'm already a warned man as I read more deaths on the QX9650 then any conroe yet...

deaths?? who, what, where, why, and how? links please!

[ExcessFC]

deaths?? who, what, where, why, and how? links please!



lol, only 1.45v, 4.6G, got a blue screen when installing OS.
cooling, no prob, voltage, i think it's normal.
things make me so said is that... I dont even have a chance to throw this baby under LN2 :(
or it will be died with some value :D (http://www.xtremesystems.org/forums/showthread.php?t=170358)

Seems to me there are a couple more I have found in my trolling...

[CyberDruid]

About 3 months ago I got a hold of an ES non production York (12MB Cacher but 2.1 stock ghz) that was tentatively called an E9350 by the seller.

Used to being able to pump 1.5 Vcore through my Kentsfields I did not realize such a voltage would kill my baby...but apparently it did.

What is so frigging weird about the "death" is that it would POST as if the BIOS were corrupt...but no matter what I did I could not get past that Boot Block routine on the P5K3...a real -pisser.

At first I thought I killed the board...( I was running some high chipset voltages) but soon learned that it did the same thing on any board...

Now that is a weird "death" because I would think if it was damaged it would not even POST.

So yes I would agree that voltages os 1.5 Vcore and 1.6 PLL are dangerous to the Penryns.

[Supertim0r]

vPLL IS dangerous

[safan80]

Seems to me there are a couple more I have found in my trolling...

that's only one.

I've put 1.6 volts on my chip and I've used a 1.7 v pll and my chip still works.

and to get 4.4Ghz 64bit stable I use 1.54 vcore.

[btdvox]

Hi Jag87
This has been the most useful thread for me, I have incoming a QX 9650 and 2 8800 ultras on the 780i (evga)

Im wondering if there are any FSB Blocks that i hear of giving an error.

Im thinking about doing 4.2 @ 1.45 ish (400X10.5) on Water though. So it should be fine. I know with water I can push this sucker but I dont care much about that anymore- Just want something thats nice and fast and ready for my gaming:)

Got any advice from overclocking on this motherboard?

[JAG87]

I am glad it was useful.

you want advice? take the heatsink off the board as soon as you get it. take that crap putty off, take the sponge things off (you will know when you see it), and clean it up. put ceramique on the mosfets, and put as5 on the chipsets, and surround the die with good quality electrical tape so the sinks dont short the resistors. put some tape on the board too where the resistors stick out under the heatsink.

my next advice is leave the fsb alone. I have done tons of tests and there is absolutely no gain at all from faster fsb. My memory read and write speed increase, and my latency decrease in Everest, but as far as actual performance there is zero improvement. games, same fps. encoding, same times. cpu bencmarks, same numbers. Bringing the fsb higher will only do 2 things, make your NB and SB hotter, and make your cpu hotter too, because the GTLREF voltages are automatically increased with higher FSB. use that unlocked multi that you paid for!

currently running 333x12, and I see no difference from 400x10. I'm running 1.35v at load after vdroop. BTW vdroop on this board is BAD, a whole 0.1v, so you will need to do the pencil mod if you want to go higher than 4 ghz, otherwise if you need lets say 1.45v at load, you will need 1.55 in bios and you will cook your cpu at idle with 1.5+

4.2 sounds like a good goal, it should be very achievable. just watch your temps, because above 1.35v these chips burn like lava. people say 45nm runs cool, yea at stock lol!!! once you overclock, these things get out of control with the wattage. I am running custom water, with apogee GTX block, and I get 33-35c idle and 55-58 load. thats with 1.42v idle and 1.35v load.

[sdumper]

About 3 months ago I got a hold of an ES non production York (12MB Cacher but 2.1 stock ghz) that was tentatively called an E9350 by the seller.

Used to being able to pump 1.5 Vcore through my Kentsfields I did not realize such a voltage would kill my baby...but apparently it did.

What is so frigging weird about the "death" is that it would POST as if the BIOS were corrupt...but no matter what I did I could not get past that Boot Block routine on the P5K3...a real -pisser.

At first I thought I killed the board...( I was running some high chipset voltages) but soon learned that it did the same thing on any board...

Now that is a weird "death" because I would think if it was damaged it would not even POST.

So yes I would agree that voltages os 1.5 Vcore and 1.6 PLL are dangerous to the Penryns.

Cyber thats odd...

I run mine at 1.61/1.62 and 4.8 ghz as my prime setting and have had no issues even when i pumped 1.72 into it for a basic 32m test at 5.1ghz...

That said ... you know me and know i never run my chips warmer than -30c or so...

[dunn2953]

i ran 1.9v PLL through my cloverfeild for about a week and it runs fine now.