SB Overclocking results?

Got her to 4.5ghz, 45x, 1.38v.

Been running Prime95 Small FFTs for just over 45mins so far. Will leave it running a bit more and then I'll try the multiplier at 46x.

sandyv.jpg
 
rUmX said:
Got her to 4.5ghz, 45x, 1.38v.

Been running Prime95 Small FFTs for just over 45mins so far. Will leave it running a bit more and then I'll try the multiplier at 46x.
Click to expand...

What settings did you use?
 
I've noticed something where I cant get above 4.5 no matter what voltage i use, well I tried up to 1.4 and I freeze and the windows boot screen.

So i went from trying 4.8 down to 4.7, booted into 4.6 with 1.35v but froze running prime95. So I dropped down to 4.5 @ 1.35v stable, decided to drop my voltage and now it seems like I'm stable @ 4.5 with 1.3v.

Just seems kind of weird that I can drop that much more voltage and be stable but i go up to 4.6 and I unstable. Need to play with some more settings I suppose.
 
smokin9 said:
I've noticed something where I cant get above 4.5 no matter what voltage i use, well I tried up to 1.4 and I freeze and the windows boot screen.

So i went from trying 4.8 down to 4.7, booted into 4.6 with 1.35v but froze running prime95. So I dropped down to 4.5 @ 1.35v stable, decided to drop my voltage and now it seems like I'm stable @ 4.5 with 1.3v.

Just seems kind of weird that I can drop that much more voltage and be stable but i go up to 4.6 and I unstable. Need to play with some more settings I suppose.
Click to expand...

have you enabled PLL overvoltage? not saying this will necessarily break through the wall you're seeing, but it's worth a try.
 
smokin9 said:
I've noticed something where I cant get above 4.5 no matter what voltage i use, well I tried up to 1.4 and I freeze and the windows boot screen.

So i went from trying 4.8 down to 4.7, booted into 4.6 with 1.35v but froze running prime95. So I dropped down to 4.5 @ 1.35v stable, decided to drop my voltage and now it seems like I'm stable @ 4.5 with 1.3v.

Just seems kind of weird that I can drop that much more voltage and be stable but i go up to 4.6 and I unstable. Need to play with some more settings I suppose.
Click to expand...

I had similar behavior. I am stable at 4.5 at 1.3 volts (1.27 at load) but it takes 1.4 to get to 4.8. I had more success (lower Vcore) by upping the VCCIO and VCCSA voltages.
 
ran cinebench at 5.1ghz for a score of 9.85

testing some F@H frames at 5.ghz atm
 
Well that sure is toasty, but it was fun hitting 5ghz. I only ran this for a few seconds and backed down to my default settings which are 4.5ghz ~1.38v with Prime95 load temps of ~65C max.

I ran it once more after this and it went for a few minutes, but the CPU started creeping toward TJMax so I cut it off.

Glad to know my chip will do it if pushed :) Hey Intel says 1.52v max so why not try it ey? Maybe with a bit more tweaking I can achieve this with more reasoanble temps, but 4.5ghz is plenty fast for now.

Cooling = Xigmatek 1283 with AS3



Uploaded with ImageShack.us
 
hmz said:
What settings did you use?
Click to expand...

The Asus P67-M Pro bios (my version is 0405) is real shitty... at least compared to the other bios screenshots I have seen around.

Here are my settings (whatever is not listed was set to Auto):

Settings: Manual
Turbo: All Cores
Multi: 45x
Load Line Calib. - Disabled
Ram voltage: 1.5v
Spread Spectrum: Disabled
C1E: Disabled

For Turbo to work I had to leave Speed Step enabled. I also have the TPU switch disabled.
 
Last edited:
smokin9 said:
I've noticed something where I cant get above 4.5 no matter what voltage i use, well I tried up to 1.4 and I freeze and the windows boot screen.

So i went from trying 4.8 down to 4.7, booted into 4.6 with 1.35v but froze running prime95. So I dropped down to 4.5 @ 1.35v stable, decided to drop my voltage and now it seems like I'm stable @ 4.5 with 1.3v.

Just seems kind of weird that I can drop that much more voltage and be stable but i go up to 4.6 and I unstable. Need to play with some more settings I suppose.
Click to expand...

I just came to the same conclusion last night (4.5 GHz @ 1.32V). Everything above 4.5 results in a BSOD/memory dump eventually. Waiting for a new BIOS in the future to play again.
 
BottomsUp said:
Well that sure is toasty, but it was fun hitting 5ghz. I only ran this for a few seconds and backed down to my default settings which are 4.5ghz ~1.38v with Prime95 load temps of ~65C max.
5ghz.jpg
[/URL]

Uploaded with ImageShack.us
Click to expand...

Thats rather high voltage for 4.5. At 1.38, I run at 4.8Ghz. Most people running at 4.5 are around 1.30-1.32V. Try reducing your voltage, it'll bring temps down.
 
Just promise not to be mad that im folding for OCN :p Were all forums hoppers looking to share/gain info anyways. Only pic i have right now of my stable clocks so far... since im at work. My CPU will boot into windows without bsod @ 50x multi and 1.36v tho... so ill wait for my loop to come in the mail and see what 5.0 will take to be rock solid stable


folding.PNG
 
@rmp: Just an interesting observation. My 2600K @ 4.7Ghz gets 45k PPD on a 2686.
Nevermind, your lower PPD is probably coz of the overhead of two GPU3 clients.
 
xbonez said:
@rmp: Just an interesting observation. My 2600K @ 4.7Ghz gets 45k PPD on a 2686.
Nevermind, your lower PPD is probably coz of the overhead of two GPU3 clients.
Click to expand...

Yea i dunno im starting to think something is up as well... most of the stuff i own is broken or defective... luck of the draw i guess...

The GPU3 clients dont seem to pull much CPU load.. like 2-4% tops.. maybe i need even more voltage I dont know. Ill run the 2600k alone for a while and see if it picks up a little... right now its at like 32k.

EDIT: w/ GPUs off im seeing like 43k atm
 
Last edited:
Thats more like it. I use a 8800GT on that rig, so I'm better off not folding on it. Since you have two 460s, you get a better PPD folding on them unless you can migrate them to another system.
 
From my experience (had a 920@4ghz briefly), if you're going to fold bigadv on 4C/8T it better be dedicated with no GPU folding.
 
I got my 2600k up to 4.7 but I have hit the wall. I upped the vcore to 1.4v and turned off all the throttling options and it is prime stable with temps maxing out at 67-69. I am not really sure what other options to change, but I really want to get this chip to 5.
 
i honestly dont know why some of you want to go over 4.6ghz.

i think 4.6ghz is a sweet spot with these cpu's, and its more than enough for today's needs.
 
Draax said:
I got my 2600k up to 4.7 but I have hit the wall. I upped the vcore to 1.4v and turned off all the throttling options and it is prime stable with temps maxing out at 67-69. I am not really sure what other options to change, but I really want to get this chip to 5.
Click to expand...

You have PLL overvoltage turned on, right? Try upping the VCCIO and VCCSA - those helped me reach 4.8 when I was having trouble.
 
@ Forceman

Would you mind sharing those voltages? Your settings?

Everything you're doing to get a stable 4.8Ghz

Thanks in advance.
 
DualOwn said:
i honestly dont know why some of you want to go over 4.6ghz.

i think 4.6ghz is a sweet spot with these cpu's, and its more than enough for today's needs.
Click to expand...

My Q6600 was more than enough for today's needs. This is [H]ardForum not [L]impForum.
 
Arkainium said:
Are you folks stable when stressing a single core? Stressing a single core with EIST and C1E enabled doesn't get the same voltage response as stressing all four cores. For example, my system is completely stable when stressing all four cores (full voltage response) but crashes when stressing a single core (partial voltage response). I don't know how to get around this without disabling EIST and C1E or bumping up the voltage excessively to compensate.
Click to expand...

You could try lowering the LLC and raising the voltage offset. This should keep your full load voltage the same while raising the voltage in other conditions. My problem was different in that I was crashing at idle, but it should work for you too. It's not elegant but I don't know any way to directly control the rate at which EIST changes the voltage (besides making it constant).

My settings for anyone interested are 46 multi, voltage on auto (0 offset), and LLC at level 4 (ASRock board; default is level 5). This results in a load Vcore of ~1.27. The processor seems p95 stable down to ~1.23 but I like the safety margin (and it's plenty low anyways). I have a 2600K, P67 Extreme4, and a Zalman 9700 heatsink.
 
fortnern said:
You could try lowering the LLC and raising the voltage offset. This should keep your full load voltage the same while raising the voltage in other conditions. My problem was different in that I was crashing at idle, but it should work for you too. It's not elegant but I don't know any way to directly control the rate at which EIST changes the voltage (besides making it constant).
Click to expand...

That's pretty much exactly what I did except I disabled LLC completely. The additional voltage that I need to keep it stable when stressing a single core gives me plenty of room for vdroop when all cores are loaded, so LLC isn't really necessary in this case. It's unfortunate because it means that I need a +0.03 voltage offset to be completely stable at 43x. I wonder if I should just disable EIST and C1E, and set the voltage manually. I'd be able to get at least 45x that way, but I really don't want to lose the dynamic voltage throttling. I guess I can't eat my cake and have it too. Oh well.

fortnern said:
My settings for anyone interested are 46 multi, voltage on auto (0 offset), and LLC at level 4 (ASRock board; default is level 5). This results in a load Vcore of ~1.27. The processor seems p95 stable down to ~1.23 but I like the safety margin (and it's plenty low anyways). I have a 2600K, P67 Extreme4, and a Zalman 9700 heatsink.
Click to expand...

On my Asus P8P67 Pro, setting offset = auto is not the same as setting offset = 0 (which is not an option). For me, setting offset = auto is about the same as setting offset = +0.030 at 43x.
 
5.0ghz stable @ 1.46v. 75C max using newest IBT that utilizes the SB optimizations. 71C using old IBT or Prime 95.

9.80 on Cinebench.






 
Last edited:
Nothing special really. Enabled PLL overvoltage, manual voltage set to 1.46, LL to Ultra High -- all C States and Speedstep enabled. Right now I'm trying to get it stable using the offset voltage function.
 
This is what I'm at so far.

2600k49001400v2.jpg


When I was messing with AVX, I was able to get up to 5.2GHz IBT 10 passes stable. Didn't want to go much further than that.

2600k52001520v4c.jpg
 
SilverSliver said:
Nothing special really. Enabled PLL overvoltage, manual voltage set to 1.46, LL to Ultra High -- all C States and Speedstep enabled. Right now I'm trying to get it stable using the offset voltage function.
Click to expand...

Can you sleep with those settings?
 
Forceman said:
Which version of IBT is that? I think I have 2.50.
Click to expand...

You have to download the new Linpack from Intel and import it into the IBT front end. If you set it to auto it gets the chip hotter and puts out like 110-130gflops overclocked.
 
NukeleaR said:
Is this normal? 2500K, stock air cooled...seemed a bit high to me.
http://i53.tinypic.com/2liiih3.jpg
Click to expand...

the auto tune feature on these asus boards can be a bit wonky. when i pushed the button it set my BCLK at 4.2 MHz - you're reading that correctly...instead of ~100 MHz it set it at 4.2 MHz. With a multi of 48 i was screaming along at 201.6 MHz. i cleared my CMOS and started over by doing it the 'old fashioned' way through the BIOS.

i'm considering removing the AI Suite altogether. it keeps giving me odd warning messages like "Motherboard Temperature is 126 Celsius!" when it clearly is not.
 
Forceman said:
You have PLL overvoltage turned on, right? Try upping the VCCIO and VCCSA - those helped me reach 4.8 when I was having trouble.
Click to expand...

I tried the F5c beta bios for my Gigabyte p67a-ud3p, which allows me to turn on PPL overvoltage, but it is unstable and wont boot. My dual bios just flashes back to F2 saying checksum error in bios.

I manually increased the CPU PLL voltage, but im not sure what to put it to. My BIOS doesn't have an option for VCCIO or VCCSA unless these are acronyms I am unfamiliar with.

I just remounted my HS but I guess Im stuck at 4.6 until a BIOS update comes out.

Anyone else here using the p67A-UD series? Perhaps you could tell me your bios settings?
 
Well, for some strange reason and I am not sure why. I've lost my ability to run 4.8Ghz stable. I hope I didn't do something to the motherboard or CPU but hard to say. It's doubtful but possible I guess. I even reverted back to the old bios. I've also even attempted to reinstall the OS but that didn't stop the machine from freezing at 4.8 during boot up, old OS installation or boot up during an OS install. So I am perplexed a bit.

I did however get it to 4.7Ghz rock solid stable, lower temps, highest core was around 82c only briefly. Also, at a higher memory speed, 1902. The vcore is set to 1.4 and never goes above 1.416 during stress testing.

437ri.jpg
 
You must log in or register to reply here.
Back
Top