Jump to content

My Opteron 165 gets LESS stable as I increase the CPU voltage??


Recommended Posts

Hi guys, first post!

 

Ok, so I just got this new rig because I heard that 165's are great overclockers and I was pretty keen to learn how to play around with all the acronyms like FSB and HTT :)

 

So i'm slowly increasing the FSB and not running into any problems.. at about 2.1Ghz Prime95 starts to report errors on a CPU stress test, so I increase the core voltage on the CPU by ~5% to about 1.25V. Still not stable, so I increase it to 1.3V but not luck. Anyway, I get up to about 1.5V and core 0 is still giving errors (core 1 is a champion, I've never managed to get it unstable).

 

Anyway, to cut a long story short, I finally got the CPU stable at 2.2, 2.3 and 2.4 by DECREASING the core voltage to 1.184. Any less than this and I can't boot, and more than this and core 0 is unstable unless I decrease the FSB. I don't think it's the RAM because core 1 is 12 hours Prime95 blend test stable. It's not heat because the CPU has never broken 43 degrees according to nTune.

 

What am I doing wrong? Why is my CPU so.. umm... "special" .... ? :)

Share this post


Link to post
Share on other sites

Well for starters there is technically no such thing as a FSB when talking about A64 overclocking.

 

I think you need to read the overclocking guide. Your CPUs stock voltage is 1.4v, so you're doing something wrong.

 

I don't think it would even run at 1.184v.

Share this post


Link to post
Share on other sites

At 1.4 I can't run above 2.0 Ghz...

 

As to it not running at 1.184, see attached... If I raise the voltage from this, even by the smallest increment I can, it is not stable at exactly the same speed. Note that Core 1 is totally solid at whatever voltage I give it.

Share this post


Link to post
Share on other sites

Please read the overclocking guide...its very informative.

 

Just a thought - you may be running 1.184 +special VID % so you may be running (for instance) 1.184V VID +123% = 1.45V hence its stable. So when u go to 1.3V you may be running 1.6V and stability issues may occur because of heat...

 

Think 450W PSU should be ok...are you using a divider for your memory to see what your CPU will go to? Bearing in mind that not all Opteron's run the same.

 

Also have you checked your HTT multiplier is low also...cus 260x5 would mean a 1300Mhz HTT bus - which would be unstable also.

 

I started with low memory divider (1/2) to cancel my memory out of the stability equation and managed to get my Opteron 170 to 2.6 at stock volts, no higher. Didn't want to push for more. Then I lowered my processor multiplier to take it out of the equation and began upping the memory to 260, managed to do it with the TCCD guide. Once I knew both would do 260 and at what settings I tried them together - to get 2.6Ghz, 260HT and 260Mhz Memory 1:1! Woolah!

 

The guide helps a lot though!

 

Matt

Share this post


Link to post
Share on other sites

Thanks for the replies.

 

Hmm.. I suppose it could be the PSU, but I'm hardly asking for much, just to be stable at 2.1 at stock voltage! I also think that core 1 would be having problems as well if it was the PSU, but it's totally stable up to 1.5V (at which point my computer won't boot!!).

 

Yes, I've run with my memory at a 1/2 divider, and as I mentioned I'm pretty sure it isn't the memory because a simultaneous blend test running on core 1 is totally stable, and that should isolate the RAM as being fine. At any rate it's well within it's specifications. Right now it's at 180. Additionally, all the memory timings i'm using are stable according to the OCZ forums. I've found that running at drastically relaxed timings has no effect at all on the stability of core 0.

 

I'm aware that there is a +% modifier to the voltage (in fact, i'm using 1.1V + 110% in order to get a 1.184 readout from CPU-Z).

 

Let me reiterate that below this voltage I cannot boot, and above this voltage (1.1 + 113%) core 0 is not prime95 (FPU stress test) stable.

 

I've tried using higher LDT voltages, but it has no effect. I've swapped out the RAM with another set of 2x generic PC3200 DDR400 and found no change in stability. The only parameter that seems to have a bearing on the stability (the difference between failing in <1min and stable for 5+ hours) is the CPU voltage.

 

Starting at 200MHz HTT, the stability of the system is roughly linearly related to the CPU voltage. To be stable at 2.1, V=1.25. At 2.2, V =1.21, At 2.35, V=1.8.

 

Lastly, I have read the overclocking guides, fairly carefully. That's why I'm so surprised!! My computer is behaving almost exactly contrary to the dogma: "If the CPU is unstable (and the RAM is okay), increment the voltage upwards"

 

thanks again.

Share this post


Link to post
Share on other sites

Not that this has much to do with anything, but my memory does the same.

Rated at 2.6v, but more and more unstable as I go up. Most stable is actually at 2.5v

Never could figure it out, even RMAed it twice. I know theres a voltage limiter from 2.8v to 3.2v, but this is ridiculous. Starting to think it's my MB. Hey maybe I should try Yellow slots, with this new bios...hmm...

Anyway, Hope this opty gets resolved; maybe I'm missing something obvious. :eek2:

Share this post


Link to post
Share on other sites

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
×
×
  • Create New...