Intel E8400 Wolfdale Living Review - Page 12

Ace-a-Rue · #**111** 12-02-2008, 18:02

here is the wPrime on my 9650 at 61% CPU GTLREF:

https://home.comcast.net/~captain-is...d/sixtyone.gif

Clunk · #**112** 12-02-2008, 18:31

What happened to the temps?

The 9650 have the same affliction as the wolfies?

Ace-a-Rue · #**113** 12-02-2008, 18:36

yeah, kind of pathetic especially when my 1st 9650 produced those same kind of temps at 4.4ghz using about 1.48 vcore...big time difference between the various chips coming out of intel's fab factory!

EDIT:...should have mentioned that the TRUE with a strong 120mm fan would have lowered those temps down to 68-70C range...the Freezone Elite is sucking wind after passing 4ghz at 1.28 actual vcore...once passing that range, it is well above its maximum TDP threshold!

Clunk · #**114** 12-02-2008, 18:41

I remember anandtech saying they had a hot runner like that.

In your opinion, do you think that it's actually running at those temps, or is it the infamous sensor issue again?

What happened to your first 9650?

supershanks · #**115** 12-02-2008, 19:02

From what i've read re gtlref it's probably not doing much at those temps.
Quad-Core Overclocking, continued

Quote:

Originally Posted by bottom of page 3)

Our tests revealed that CPU load testing failures generally occurred when CPU temperatures exceeded 65C or so. Over 65C it becomes ever more difficult to tune GTL voltages as there is always some uncertainty pertaining to the cause of a CPU core failure: Was it temperature or VTT/GTL related?

I'm water cooled so my temps are genarally < 60C on my Q6600 Go with gtlref making a decent impact.

Clunk · #**116** 12-02-2008, 19:23

Quote:

Originally Posted by Clunk

The GTL ref didn't have any positive effect at all on the 8400/DFI X38/Phase combination for me. Things went better on that setup with it disabled.

We were wondering about this the other day - we know that the GTL ref is less effective at higher temps, and it *appeared* to be the case for very low temps as well (-40 ish).

That's why I was asking if he thought if the temps were correct

supershanks · #**117** 12-02-2008, 19:29

Yep but gtlref has a bigger impact on quads than dual core.
the 61% CPU GTLREF is on ace's yorkfield.

but yep good point

if cpu is stable at that temps are wrong

Ace-a-Rue · #**118** 13-02-2008, 05:03

Quote:

Originally Posted by Clunk

I remember anandtech saying they had a hot runner like that.

In your opinion, do you think that it's actually running at those temps, or is it the infamous sensor issue again?

What happened to your first 9650?

it's hard to say about the temp accuracy...my first 9650 ran cooler at the same vcore and it topped out at 4500 mhz but it was not stable for me at that frequency...i call stable when i can run wPrime 1024M stability bench...it was stable at 4400 mhz..BUT..the temperature readout from coretemp was a little over 80C on two cores...that was on the TRUE running a 2.7K rpm Panaflo at full bore

i wish i kept the first 9650...i was being a nerd about the individual cores not indicating properly...i had core #0 locked at 51C at stock speed...it would move upwards only when enough heat was being produced to push it above 51C...so...at stock speed, stock vcore, core #0 never moved off 51C even though i ran wPrime 1024M stability bench...there was another core which read about 43C and that stayed right at 43C until enough heat was produced.

the 1st 9650 could clock to 4400 mhz with about 1.48 vcore...if i try 1.48 on this 9650, the cpu signals the board to shutdown when i run wPrime.