Full Version : Opteron 165 Overclocking
asustech >>Overclocking Section >>Opteron 165 Overclocking


<< Prev | Next >>

Merc- 12-08-2006
Cris-

I'm running the following:

CPU:
334x9 (3006MHz)
CPU Voltage Set at 1.275v
CPU Overvolt enabled
Idle/Load voltage 1.47v

HT: 3x=1,002MHz

RAM:
Corsair XMS DDR500 PC4000 2.65v 2.5/3/3/6 1T (133) 215MHz

My dividers are 133 and 166. 166 won't run very well so I am better off with the tight timings at 133. I don't think I have a 150 divider.

It will boot and run stable at 3105MHz (345x9) with the same RAM timings but Orthos fails at about 5 minutes or less. Ran SuperPi to 1M in 28 seconds though. biggrin.gif

Crisao23- 12-08-2006
QUOTE (Merc @ December 08, 2006 07:26 pm)
Cris-

I'm running the following:

CPU:
334x9 (3006MHz)
CPU Voltage Set at 1.275v
CPU Overvolt enabled
Idle/Load voltage 1.47v

HT: 3x=1,002MHz

RAM:
Corsair XMS DDR500 PC4000 2.65v 2.5/3/3/6 1T (133) 215MHz

My dividers are 133 and 166. 166 won't run very well so I am better off with the tight timings at 133. I don't think I have a 150 divider.

Ok. My board is a little bit different. I have 3 vcore settings here, initial vcore a vcore control value and a overvolt value.

I'm using a fixed 1.37v value without using the overvolt option. One bug on that option and cpu goes down in flames.

Up to 36+ minutes now without errors.

RAM here at 3-4-4-8 1T 2 x 1GB sticks, 2.56v

It's UCCC so it likes low voltage and 3-4-4-8 settings at clocks above stock.

RAM at 245 Mhz with 166 divider. I didn't -*test*-('") 133 divider at 3Ghz clocks.

I think that the higher clocks create some real stability issues at the memory controller, so we really need to use low dividers.

Maybe that is the key to low voltage and stable high clock settings.

Oh man... I miss my single core 280Mhz memory biggrin.gif

40 minutes +

I'll stop at 1hour -*test*-('")ing and try a 3Ghz run.

Impressive low temps here. We really need to discuss the core temp data. At least for this Opteron it's completely innacurate. How can I run core 1 at 55°C and Core 2 at 42°C ???

Lol... maybe AMD put a heatpipe inside Core 2 rolleyes.gif

Nice temps and AS5 it's not even close to the 200 hour cycle smile.gif

Merc- 12-08-2006
...And HT is at 3X? My core temps stay fairly close, 2c-5c if I remember correctly. Was further apart the first day or so then settled in. I think that the cores take awhile to fully seat against the IHS.

Crisao23- 12-08-2006
Almost one hour and I got a nice blue screen.

any 333 or 334 setting doesn't even boot at 133 divider. I need 1.5v minimum to boot and of course it fails on orthos.

trying 2700 (300x9) with 1.4v now

Merc- 12-08-2006
What are your full specs Cris?

Crisao23- 12-08-2006
I had a bad dat yesterday and a terrible night, so I should be a little bit "sleepy". I forgot a "fas-*test*-('")" setting activated on the bios. That probably was the source of my bluescreen.

Orthos fails but never shows a blue screen. biggrin.gif

My specs:

Opteron 165 (trying with usual ram settings now and 1.35v only) 2700 (300x9) ram at 166 divider.

2GB RAM Corsair value select (UCCC based)

nvidia 8800GTX at the moment

700W PSU

DFI SLI-DR Expert motherboard

4 HDs 3 sata/1 IDE: 2 of it on raid 0 (2 x raptors 74GB)

Merc- 12-08-2006
Doh!! Oh well, I left my HT at 5x when I was first running this chip. 1,536MHz HT on CPUz had my sleepy eyes wondering what I was seeing. biggrin.gif

Good news though. I figured it was some errant setting. The BBE stepping is consistently a strong overclocker.

So, after a couple of hours of frustration and disappointment, mixed in with some anger and bitterness, we slap our collective foreheads and start over. biggrin.gif

Crisao23- 12-08-2006
I got used to the la-*test*-('") official bios. This hacked one is better but they added some extra values, for TCCD based memory and so on.

I have here lotsa pics as reference of my previous bios adventures so i'll not be a problem.

For the 10th time I need to say a big THANK YOU to all of you with this thread here and of course Arlie who gave me a chance to -*test*-('") this overclocker's gem biggrin.gif

*looking at Orthos and repeating: "thou shalt not fail"

Crisao23- 12-08-2006
Ok. 2700, 1.31v. Feels stable.

More than 1h now running Orthos.

user posted image

This is the first stable run due to the wrong setting on bios.

For me, even at "only" 2700Mhz 165 is definitely the WOO on WOOT.

Before 2800 I'll update my nvidia drivers and why not a single 3DMark 2006 run ? tongue.gif

Stopped orthos at 1h32min. No errors.

user posted image

Time for a driver update and 3dmark -*test*-('").

WannaGoFast- 12-08-2006
Looked like you were off to a rough start Crisao23....
Glad things are working better for you now....
Took me a bit of playing with things to get to all play nice,
but once I got it right it works above any of my expectations.
cool.gif

Crisao23- 12-08-2006
Just did the nvidia video card driver update and a 3dmark 2006 run.

I think that futuremark has issues with their servers, it cannot connect and submit/show scores.

A good score for my AMD system: 9597

Let's try 2800 and up now.

Merc- 12-08-2006
Looking great Cris. You're at about where Wanna and I were. Tobight I'm going for the temperature record. Lowest temp running Orthos for 30 minutes biggrin.gif Man it is cold here.

Crisao23- 12-08-2006
Not so great.

Memory settings are everything on this overclock.

One "0.x" change and it's all gone.

Doesn't matter if you have a high speed memory.

All my g.skill settings are gone, I'll need to learn all from scratch.

For the first time I had blue screens on orthos. 3Ghz will be a looong way to go.

Edit:

2800 fails at every single memory setting and voltage, currently trying it at 1.5v. And that's no memory error.


Crisao23- 12-08-2006
How can one core reach 60+ celsius and the other only 48 celsius ?

My core 1 temperature is completely messed up, it goes up and down, while my core 2 temperature is steady and only shows usual changes during voltage and higher clocks.


Something isn't right with this processor, maybe a damaged sensor ?

Maybe this one deserves water or a IHS surgery.

When I received the processor I noticed that some of that "glue" who holds the IHS was all around the left side of the processor.

user posted image
user posted image
user posted image

I really don't know. Maybe I should stick to 2700 Mhz sad.gif

Somehow the core temp issue fixed itself...

I'm now at 2808Mhz, stable for more than one hour on Orthos, vcore at 1.44v. I set the divider really low (memory is running at 200Mhz) so I can max out the processor first and after that find a good ram config for it.

user posted image

3D Stable:

user posted image

Merc- 12-08-2006
Cris-
You mean the CoreTemps have settled down? Same thing happened to me. Took about a day or so. Like I said, I think it takes a little time for the IHS and the cores to mate up. Little heating and cooling to get them settled.

Ok check this one out boys. I'd be happy with these temps on air at idle. biggrin.gif
user posted image

Free Forum Hosting by Forumer.comTM!