680i overclocking thread

So I started playing with my settings last night, trying to get a fast real world overclock. I didn't care what my final FSB numbers were, or what not, I just wanted to see what I could get the system to run smoothly at, and what combo of settings gave me the best real world overall performance. I was a bit shocked when I get these settings to blow away my old stuff in Everest (not that I am saying Everest is real world performance, but it's a good baseline to start from in evaluating changes to your settings):

FSB: 333 (1334 QDR)
RAM: 333 (667)
Timings: 3-3-3-10-1T

My final OC is only 3ghz, but these settings are smashing what I was getting at 3.3 and 3.4ghz in Everest, with much looser timings and no 1T. *shrugs* Just something to think about. It seems to me the C2D's are so fast once you get up above 2.8ghz, if your memory isn't timed tight enough to keep up, you start to bottleneck. It seems to me keeping the FSB and the final CPU speed in moderation, but tightening up your memory quite a bit, yields serious real world performance. Considering that's why I am OC'ing this machine, rather than to just get a swank FSB number, I am happy with this.

I am curious to see what I can get the system running at once I have my D9 based Dominator's in the box. Right now I am using some value level OCZ RAM and it's doing 3-3-3-10-1T @ 667 on stock volts. I am hoping I can get near those kind of timings with the C5D at 800mhz.

Quote:

---

Originally Posted by tylerw13

hmmm ok this is really wierd....i just tried replaceing my other e6600 with the one that i was having troubles with...and it stopped having problems with the 1500 fsb strap.....very weird...i can go up to 1700 fsb with it now...however my cpu cant go that hight booting into windows....very very crazy....my first one is a 625a*** batch and the other that has the 1500 fsb strap problem is a 628b*** that makes me wonder if that could be what is having problems with this strap problem at 1500 fsb...anyone heard anything about this problem...and hey terminator have what batch is your cpu is a b by chance??? maybe those are that processors that have that problem?????


EDIT: sorry it really isnt a cpu strap its jsut a problem with the motherboard only using and alowing one core when it goes over 1500 fsb...a strap is a motherboard that cant go over a certain fsb

---

my processor is core 2 duo 6700 l624a but i don't think the problem is ours motherboard that don't run over 1500mhz, all motherboard evga that i saw go over 1500.
I think there are the bus problem with ours processor nforce 680i was tested by nvidia over x6800 platform with this cpu bus go up 2100mhz.

the best processor for overclock bus in this motherboard are 6300 6400 x6800

in this review is explicated that their tested one evga 680i on 6700 processor and bus don't go up 360mhz but when on the same board was mounting core 2 duo 6300 bus go up to 490mhz

http://techreport.com/reviews/2006q4.../index.x?pg=21

WHAT DO YOU THINK?
i had tested with beta bios p17 p20 and p21 but there is the same situation don't go over 1500 if go don't recognize second core
do you think this problem'll resolve with new bios uprgade ?

quad core does pretty well on this board, becayse multi set to 15 = easy 4.3Ghz

Quote:

---

Originally Posted by Terminator82

my processor is core 2 duo 6700 l624a but i don't think the problem is ours motherboard that don't run over 1500mhz, all motherboard evga that i saw go over 1500.
I think there are the bus problem with ours processor nforce 680i was tested by nvidia over x6800 platform with this cpu bus go up 2100mhz.

the best processor for overclock bus in this motherboard are 6300 6400 x6800

in this review is explicated that their tested one evga 680i on 6700 processor and bus don't go up 360mhz but when on the same board was mounting core 2 duo 6300 bus go up to 490mhz

http://techreport.com/reviews/2006q4.../index.x?pg=21

WHAT DO YOU THINK?
i had tested with beta bios p17 p20 and p21 but there is the same situation don't go over 1500 if go don't recognize second core
do you think this problem'll resolve with new bios uprgade ?

---

Take in to account the difference in multipliers between the 6300 and the 6700. I bet if you stepped down the multiplier on that 6700 to the same as the 6300 (which is what, 7x?) it would also reach 450+ FSB. As you step that multi up, your FSB "ceiling" for your current voltage is going to drop. I can barely do 375 FSB on my E6600 if I stick to 9x multi, but if I drop the multi to 8x, I can do 425 or so stable. I think memory is the limiting factor in my case, but it still demonstrates what I mean.

(Also, he is running an engineering sample E6700 in that test...not exactly a clean test of retail vs. retail chips. Hard to draw conclusions from that.)

I recently got a Striker and have an encountered a weird problem: I cannot use all four DIMM slots without stability issues.

I discovered this while trying to install XP, and it kept crashing at random points with all four DIMMs installed. But when I use just 1 or 2 DIMMs, everything is beautiful. I am using a pair of Corsair TWIN2X2048-6400C4D sets. I tested dozens of combinations, and all four DIMMs work individually and in pairs fine in any of the four slots (both blue, both white, one of each, etc). So it is not the DIMMs themselves, and it is not the slots themselves, it is only an issue when all four are in use at once.

I was running them at 2.1V and also tried 2.2V, and tried a variety of timings (including extremely loose, extremely low clock rates). :confused:

Any idea what might be the cause of this, and any solutions to try? I would like to be able to run with 4 GB, but I would like to avoid having to return a board that is so scarce at the moment. :)

@ Vol...i know this does not help ya now, but always better off with 2 x 2 gb, as opposed to 4 x 1gb

maybe they need more juice...try the volts a lil higher see if that gets ya stable on the rams...good luck

icon57 is right, it's always better to use two modules than four. I can't think of many boards in the past year or so that haven't had some kind of issue with 4 DIMM's, even it was an automatic down clock or something of that nature.

Tried 2.3V, 2.4V and 2.5V and couldn't boot to Windows. Went back to 2 DIMMs and no problem.

Thanks for the tips. If anyone else has an idea, please pass it along.

Quote:

---

Originally Posted by Terminator82

my processor is core 2 duo 6700 l624a but i don't think the problem is ours motherboard that don't run over 1500mhz, all motherboard evga that i saw go over 1500.
I think there are the bus problem with ours processor nforce 680i was tested by nvidia over x6800 platform with this cpu bus go up 2100mhz.

the best processor for overclock bus in this motherboard are 6300 6400 x6800

in this review is explicated that their tested one evga 680i on 6700 processor and bus don't go up 360mhz but when on the same board was mounting core 2 duo 6300 bus go up to 490mhz

http://techreport.com/reviews/2006q4.../index.x?pg=21

WHAT DO YOU THINK?
i had tested with beta bios p17 p20 and p21 but there is the same situation don't go over 1500 if go don't recognize second core
do you think this problem'll resolve with new bios uprgade ?

---

do you know what though i put in my other e6600 that i have and it works fine....it overclocks clear up to 450 fsb and it recongnizes my 2nd core every time...its really weird...its like it doesnt like my other processor or something...then when i put in my other...its goes back to only seeing 1 core...its very weird and nothing explains it...ive tried lowering mutis and everything....it sux bad...i freakin have gotten really frustrated with this thing:slapass:

to me a bios update isnt going to fix it...at least that is how i feel!!!

but is normal that at default bus speed 266 and memory gskill hz 800mhz 4 4 4 10 2.050volt i' must increase chipset spp voltage from 1.2 volt to 1.25 because at 1.2 volt freeze windows?

Do you have a bios screen shot of what your talking about?

Ive got a whole bank of DIMMs not working and am hoping i can avoid an RMA.

Terminology may be incorrect so here is what I mean:

Bank1: Dimm slot a1 Blue and dimm slot b1 white (This bank is out if i put a stick in either slot and boot I get a continuos beep)
Bank2: Dimm slot a2 Blue and dimm slot b2 white (This bank work fine I can populate a2 and b2 and overclock like heck but of course no dual channel)

Quote:

---

Originally Posted by sdumper

Do you have a bios screen shot of what your talking about?

Ive got a whole bank of DIMMs not working and am hoping i can avoid an RMA.

Terminology may be incorrect so here is what I mean:

Bank1: Dimm slot a1 Blue and dimm slot b1 white (This bank is out if i put a stick in either slot and boot I get a continuos beep)
Bank2: Dimm slot a2 Blue and dimm slot b2 white (This bank work fine I can populate a2 and b2 and overclock like heck but of course no dual channel)

---

i must setting (for stability at default ) voltage of SPP at 1.25 volt standard voltage at default fsb speed is 1.2.
is that normal ?
CPU Core auto
CPU FSB auto
Memory 2.1V
SPP 1.25
MCP auto
SPP-MCP auto

E6600, evga680i.

Max FSB: 437,5 x6, super pi stable.
Max CPU: 425 x9, super pi stable.

Vcore 1.58

Because I'm new to Intel setups I've been concerned about my CPU temps. With that said would you guys take a minute and post your Idle/Load temps if you would along with CPU votlage and cooling method?

I ask becasue my CPU is running at about 2.924Ghz on 1.4v (BIOS setting) and my Idle temps hover right about 36C-42C and my Load temps jump upwards of 55C-58C. This might now seem horrible but I'm using the following cooling,

Thermochill PA120.3 w/ 3x Scythe 120mm fans (1200rpm)
Swiftech MCP655 Pump
1/2" ID all around
Swiftech Apogee block
Ambient Temp: 20C-23C

My CPU is the only block in the loop and I'm currenlty running a T-Line. I've worked very hard to get all air out of system so I don't think that is causing the problem. Then last night I went back to my old radiator which is BIP 120.3 and after purging all the air I still get the same exact temps and I'm positive all air is also out of this system. Anyway just curious if these temps are resonable or are they too high for current cooling method?

Quote:

---

Originally Posted by Roger_D25

I ask becasue my CPU is running at about 2.924Ghz on 1.4v (BIOS setting) and my Idle temps hover right about 36C-42C and my Load temps jump upwards of 55C-58C. This might now seem horrible but I'm using the following cooling,

---

Have you tried reseating the waterblock on the CPU? Make sure you clean the block and CPU well, then reapply some Arctic Silver 5 or something (just a pea size drop in the middle of the CPU. Are you sure you're getting good water flow through the system?

Thanks for the reply Trellame, much apprecaited! Since I wrote the post above I've tried a couple different things. I had been getting my temperature from nTune (didn't use any of the overclocking aspects of it), just used for voltage/temps. I found and downloaded both "Core Temp" and "TAT" which from what I've read should be more accurate than nTUne (but I could be wrong)? Anyway these two programs read my temps as,
Idle: 29C-31C
Load: 42C-46C

Am I right in thinking that these two programs are more accurate than nTune is? I truly hope so because it seems to fall in line with the tempersture other E6600 users are getting on water with around 1.4v? Thanks again guys!

I got 4GHz on my E6600, and 12,953s in Super-Pi. But vcore is at 1.775 in bios, so I should turn of now. 58C in Core-Temp while superpi is running, 50C idle.

4 GHz

ok so its safe to say the bios is not reporting the correct voltage lol.

Check your voltage using Speedfan 4.31.

Don't know if this is common knowledge or not but the things that helped me hit what I assume to be the max FSB (~470Mhz) on my BFG 680i are:

Enable "Execute Disable Bit"
Set nForce SPP voltage to 1.55v
Set HT nForce SPP <-> MCP voltage to 1.55v

Quote:

---

Originally Posted by sdumper

ok so its safe to say the bios is not reporting the correct voltage lol.

Check your voltage using Speedfan 4.31.

---

I set vcore to 1.775v, bios reports 1.73v and speedfan 1.74v.

>----
t was found that the GTL Reference Voltage, derived from CPU VTT/FSB Voltage, on the 680i SLI Motherboard, needed a little boost for better FSB overclocks on Kentsfield Processors. About a 0.03v-0.05v boost should yield better FSB overclocks on a Kentsfield Processor. Current boards do not have this implementation, but newer boards should be tuned with a little more GTL Ref Voltage with Kentsfields. Perhaps we can expect a Revision 2 of 680i SLI Motherboards...

Where can i change the fsb voltage?
I need more oc on my 6600 , it runs on standart core @ 3,0..
now i have waterchilling setup,what is a possible value to these hardware??can i get 4,0 without Nitrogen ?? best changes in bios for o clock??
please help!!

greets from germany

jAnBAtos2142

Has anyone figured where the actual voltage read points are on this board yet? I realize it just came out but I'm very curious to know what the actual voltages are as I've pretty much never had a board which showed accurate CPU/Chipset/RAM voltages. Plus I just feel better knowing exaclty what voltages really are!

Quote:

---

Originally Posted by Roger_D25

Thanks for the reply Trellame, much apprecaited! Since I wrote the post above I've tried a couple different things. I had been getting my temperature from nTune (didn't use any of the overclocking aspects of it), just used for voltage/temps. I found and downloaded both "Core Temp" and "TAT" which from what I've read should be more accurate than nTUne (but I could be wrong)? Anyway these two programs read my temps as,
Idle: 29C-31C
Load: 42C-46C

Am I right in thinking that these two programs are more accurate than nTune is? I truly hope so because it seems to fall in line with the tempersture other E6600 users are getting on water with around 1.4v? Thanks again guys!

---

That could be the problem, I don't know that any of them can really be trusted but those second numbers seem more likely. If your CPU was really getting as hot under load as you said in your first post, you should be able to feel the tube flowing away from your CPU block being very warm. I have an E6600 also, but I'm posting from work, when I get home I'll check the temps, I haven't looked for a while.

Quote:

---

Originally Posted by Roger_D25

Has anyone figured where the actual voltage read points are on this board yet? I realize it just came out but I'm very curious to know what the actual voltages are as I've pretty much never had a board which showed accurate CPU/Chipset/RAM voltages. Plus I just feel better knowing exaclty what voltages really are!

---

VR-Zone has some. On my board the memory meters .048 volts higher than set or indicated in bios. Looks like unless Vcore is set to "Auto" the true Vcore is is way the hell off (way low on my MB) from what you set.

Just pulled the Vcore mosfet heatsink off on mine too. There are 16 Vcore mosfets and the only ones the stock heatsink was making contact with were the 2 at each end. The 12 mosfets in the middle were not even touching the heatsink pad at all!!!

Viper

That would be greatly apprecaited Trellame, if you do get a chance to post would you also include what monitoring program you use along with CPU voltage?

ViperJohn,
Excellent, I'll have to go check VR out, their usually pretty good about getting that info posted quickly! I'll also post my actual voltages once I check along with what I had the voltages set to in BIOS to show the differences. Too bad about the VREG heatsink, although I'm not suprised unfortuenlty! Are you going to try and mod the heatsink to better make contact with the mosfets, if at all possble? Your talking about the larger of the two heatsinks next to CPU correct?