Orthos

[Ika]

So my computer's been bluescreening lately, and I've been doing some stress-testing to figure out what the problem is. First thing I'm doing is running Orthos on my overclocked E6420 overnight. Funny thing is, I noticed that one of the CPU threads is running much faster than the other. Half an hour in, CPU 0 is halfway through the second self-test (so 12 tests done), but CPU1 is only on test 3. Also, CPU1 finished the first self-test after only three tests... should I be concerned?

Here's a screenshot to clarify what I mean.

 

[Idontcare]

Checkout this: http://www.anandtech.com/cpuch...howdoc.aspx?i=3184&p=4

It looks as though Core 1 is having trouble keeping up with the other three

 

[imported_Scoop]

I had some difference with my two cores as well but only a couple of seconds.

 

[Ika]

Originally posted by: Idontcare
Checkout this: http://www.anandtech.com/cpuch...howdoc.aspx?i=3184&p=4

It looks as though Core 1 is having trouble keeping up with the other three

I remember seeing that, but that page is too technical for me to figure out what it means...

Orthos has been running for about 2.5 hours now, and it seems like CORE0 is about 50% faster than CORE1... I've never seen this behavior before, not when I last Orthos'd my computer (which was a few months ago, around the beginning of the year).

 

[Idontcare]

Originally posted by: Aflac

Originally posted by: Idontcare
Checkout this: http://www.anandtech.com/cpuch...howdoc.aspx?i=3184&p=4

It looks as though Core 1 is having trouble keeping up with the other three

I remember seeing that, but that page is too technical for me to figure out what it means...

Orthos has been running for about 2.5 hours now, and it seems like CORE0 is about 50% faster than CORE1... I've never seen this behavior before, not when I last Orthos'd my computer (which was a few months ago, around the beginning of the year).

I think the basis of the Anandtech page there is that if your chip is overheating or undervolting in some capacity or within the circuits of relevance for the computations being done at the time then it throttles the afflicted core/circuit/processor (not necessarily the same thing as thermal throttling with TJmax) so as to ensure reliable operation without the processor crashing.

Basically what it suggests is back off on your clockspeed (lower your multiplier one or two clicks) and rerun the bench and see if the discrepancy goes away. If it does then that suggests the regular clockspeed you are running is causing degradation in performance (not degradation in stability, but rather in performance).

In this case the next thing to try is increasing Vcore, or improve cooling but usually increasing Vcore is easier for generating quick-test feedback to isolate the root-cause here.