The Official X2 Overclocking Discussion Thread

[knopflerbruce]

awesome, bent pins and newegg has no more stock of these. Can't get a refund either even though I just got the chip the other day...

You could always try to make them straight again. I know, it's not what you expect when you buy a new chip, but if you can save some time (and a few bucks, perhaps), then why not?

[Dave_Sz]

cause then hey might say hat I'm the one that bent he pins cause the cpu was weak or something. It's a whole outer row of pins..

[ClintE]

Wow that sucks. Can't imagine the egg doesn't have at least a doa policy on all their products, even refurbs or restocks. If nothing else I'd think they would give store credit.

[Dave_Sz]

bent pins back as much as I could. Just tried to slowly get it to what my current x2 3800 can do and it was a no go. Temp difference between cores is always over 10*C. and the chip is SUPER hot. Goes to 57*c at 1.43v with my xp90 with 120mm fan. Much hotter than my 3800. Definitely sending it back...

[scott14]

Yes, that is what I said, and I think that the point that most people and you miss ...is that if you dont over volt or over heat the chip it will go on for as long as you need it to. Most people will want to upgrade a very long time before they have used a few % of a CPU's life time. Processors dont wear out with a few years use.

I have an old 300B Celeron that has been running for the last 7~8 years with a 100% overclock (no core voltage change) it runs 24/7 and because of the very strong crypto I run on my network (its used as a print server/router/firewall) it runs above 85~90% CPU all of the time, I expect that it will be doing so for another 10 years without putting a dent in its longevity and that is with the standard HS/f, as long as you use good cooling and keep things clean and dust free there will be no problems.

I expect to get an easy 10 year life out of my 4200 and I very much doubt that running it at or just below a 1.4 core voltage will do anything to stop that.

A) Re: Voltage to cpu
Which vcore reading do you most believe?

My DFI mobo bios says one thing, say for example 1.43, while at the same time both cpu-z and Everest say something different, like 1.408 v.

Which one would you go by?

B) Re: cpu temperatures
When there are 3 different temp readouts given for (1) CPU, and (2) CPU core 1, and (3) CPU core 2, and they're all different, which one would you go by when you're trying to keep cpu temperature below some level, like say maybe 55 deg C during stress testing max load?

[mad_skills]

anyone knows what's a probable max for a G1 steping Brisbane on a high end air cooler like Tuniq Tower or TR Ultra 120 eXtr. ?

[ThatGuy16]

Maybe someone could help me get 3.4 stable? i doubt it, but maybe someone could suggest away.

My current settings are:
HT: 256
Vcore: 1.425
Multi: x13
HT Link Multi: x4
Ram: 667 @ 835; 2.0v (5-5-5-12 2T)
PCI/PCIe: is auto locked
Spread Spectrums: off

at this setting its rock solid (full orthos load is 46*c), i can't get 3.4 to get orthos stable, i get a blue screen. I tried up to 1.525v to run orthos, load temp is about 50*c @ 1.5v. And i tried a number of cominations, could the NB not be getting enough voltage? or do you think i just hit the wall?

[Dave_Sz]

so, surprisingly they are giving me a refund. Yay...

[Hughyhunter]

I have a 4600+ X2 manchester 939 on asus a8n-sli and get instability when overclocking even slightly. The mobo has latest bios 1014 and my driver has been updated as well.

I am running 4x512 ddr400 set to 333mhz in bios. Everything else for mem is auto.

I set HT multi to 4x

Set voltage to 1.4 for CPU

Set CPU to 233 w/12x for 2.8

It fails PRIME 95 in first 5 mins.

Please help! I'm new to overclocking but am learning fast!

[muzz]

I have a 4600+ X2 manchester 939 on asus a8n-sli and get instability when overclocking even slightly. The mobo has latest bios 1014 and my driver has been updated as well.

I am running 4x512 ddr400 set to 333mhz in bios. Everything else for mem is auto.

I set HT multi to 4x

Set voltage to 1.4 for CPU

Set CPU to 233 w/12x for 2.8

It fails PRIME 95 in first 5 mins.

Please help! I'm new to overclocking but am learning fast!

Have you tried dropping HT multi to 3x ?
Tried more Vcore?
Vdimm?

[Hughyhunter]

Alright... i'll try that 3x. I thought though that as long as you ran it below 1000mhz it wouldnt be unstable?

I raised my vcore of cpu voltage (same thing right?) to the max my mobo would allow that being 1.45 and the temps were to high. Under load they were 57c. Is this too high? I'm really not worried about it if you all think it might be alright to run that hot.

I havent touched the memory besides lowering the FSB to 333 and then even one step lower I think it was 266. When I was unable to boot I lowered it to 333 and it booted and went higher on cpu then wouldnt boot and lowered once again to 266 and then it posted again.

My big concern is if I cant get it to run prime 95 for at least 15 mins then it's just not stable. I ran super pi at 2.95 yesterday (before I reset my bios) on the 32m setting it took 30mins but I really dont know what that means. Does that mean it's stable or what? I just dont think it's stable because it failed prime and crashed in 3dmark.

[Hughyhunter]

http://

here is what my cpu-z settings were at 2.9 ish

[muzz]

http://

here is what my cpu-z settings were at 2.9 ish

Is it prime stable at that frequency/VCore?

[Hughyhunter]

No... like I said... i've never been able to get it prime 95 stable during an overclock. I am able to run prime all day long though at stock clocks. So nothing is wrong with prime... just my overclock.

Do those settings in cpu-z look right? What should I set my voltages too and all other settings?

[muzz]

Well 2.95 may be too high for that cpu/mobo/ramset
I saw 1.5v on the CPU-Z shot- that may be too much for THAT chip, I don't know how accurate that temp of 57 is, as i have no idea if Asus has any temp bugs, as i don't own anything Asus.

Did you try and run just 2 sticks of ram?

At what speed WILL it pass Prime?

Your saying "Fails at even just a little OC'

What kind of PSU?
Are the rails stable?
Cpu HSF?
Something is obviously failing, and it's difficult for me to tell what it is, it could be just a crappy clocker.
I would figure out what speed it IS stable at, with decent temps @ stock vcore, and then go up from there.

[Hughyhunter]

http://

Here is my overclock that i just finished. It ran for about 7 mins.

As you can see my voltage is lower than before... I am going to raise it once more to see if I can even run it stable at this speed yet.

I dont want to try 2 sticks because that is not what I am going to use. My goal is to have the fastest overal performance. I would rather not go without the extra gig.

The only speed that i've gotten it to pass prime is stock...

The fail with little overclock i'm refering to was yesterday it failed at 205mhz with voltage in auto and ram set to 400mhz.

My psu is xion 600 watt power real... how do i check the rails?

my hsf is a thermalright xp-90 w/92mm fan. I think it is running at 2000 rpm. I can tell from two diff programs. I dont know if that's fast enough as the temps were in the 50's during loads. I dont know how accurate that is two but got it from two programs. I lapped my hsf so it should be at it's max performance. I used right amount of arctic silver five.

[muzz]

I understand that you don't want to run 2 sticks full time, I was just asking if you TRIED it, because of the load involved running 4 sticks.

You mentioned that you are new to overclocking, well it takes a bit of twiddlin' to figure out what is the culprit, sometimes alot of twiddlin'.

Most folks run 12-24 hrs to determine stability.

I haven't even touched on software programs (that have been called culprits) running in the background, or usb devices...

You said you were just learning, but that you are learning fast, but it looks to me like you are just going too fast.

I used to use Hexus Prime(this is a fast bench), and ramp up clockspeed until it failed, then bumped up vcore/vdimm/chipset voltage etc... to make it happen, then repeat until it was maxxed out ( at that point, no matter how much voltage ya feed the system,it will not pass).
Then I would back it off a bit, and run longer instances of prime95 or SP to make sure it would be stable at full load for longer time periods.

Most folks run fans over the over volted parts to keep them cool (better chipset coolers, fans over ram + regulation circuit, etc..).

I always payed attention to the temps, in case that was the culprit.

Have fun, and GL

m

[muzz]

Have you looked at this guide?

http://www.xtremesystems.org/forums/...ad.php?t=46539

[Hughyhunter]

Alright... i am probably going way to fast. I just want some sort of results.

Right now i'm thinking that it really isnt any problem besides maybe that my proc and mobo just really arent that overclockable... I will continue to try though.

I was able to get 2.78 for 15 mins stable on prime 95. Last time I failed prime I raised voltage one notch and lowered frequency one notch. I also raised cas latency to 3 from 2.5 (I dont know if that makes any difference but I was noticing on this forum that everyone else has theres set at 3).

So I decided to be brave and overclock my 8800gt too. I raised the fan speed to a steady 60% and clocks to 695/970 and on my fan control panel and PSU turned all fans up to super high till the computer just started to lift off... then ran 3dmark06 and it DIDNT CRASH

I'm very happy with improved score over stock clock which was 8409 and previous project which was in low 7k.

I will continue to try to push this cpu to it's limits as well as gpu... I will read that thread and refer back for any questions... thanks for all the help.http://

[muzz]

I would PERSONALLY leave the 8800 at stock until ya find the max stable mobo/cpu/mem speeds.
THEN I would adjust the 8800gt.

If you try to do too many things at once, it's hard to tell what is causing the problem.

Try to focus on 1 thing at a time, that way you will have a much better chance at success.

What Brand+MODEL of ram is that(most ram can run that speed easily at Cas 2.5)?

Gl

m

[muzz]

BTW- Your PROBABLY looking at 2.7- 2.9 for that combo IMO.

Maybe you'll get lucky and get a bit higher, I dunno.... it does happen.

[ThatGuy16]

Installed my ram today making it 4gb instead of 2. Got some beep codes and it reset cmos

So im at 3ghz now instead of 3.3

I still havn't ran orthos yet to see if 3 is stable on stock volts. I know if was before...

Oh well this sucks, i miss my 3.3 5600+

[Hughyhunter]

Right now i'm running a 4600+ X2, Asus A8N-SLI with 4x512 sticks of Centon DDR400 ram. I cant get it to pass super pi at even stock speeds. I took two sticks out to see if that was the problem and still not passing.

Any suggestions?

[madfaze]

stress testing my rig to find out the max fsb then suddenly my rig shutdown?!

vcore default (1.35)
NB 1.125
multi 10
fsb 290
vdimm 1.95

...is this normal?did i hit the wall (fsb)?!

thanks for the help!

[palma]

hey guys, maybe someone could gimme a hand here?

this one is stable:



and if i set 250x10 it will only be stable if i really loose my mem, like ddr400 6-6-6-15 1.9v.also if i try 255 on htt it doesnt boot anyway, not even with mem 6-6-6-15 1.9v

4000+ brisb
m2n-sli dlx
2gb xsm2 800

any ideas?