Venice 3000+ @ 1.1V

ziphnor · Post by **ziphnor** » Tue Apr 25, 2006 11:01 am

Hi,

Due to some incompatibility with my DVB-C TV tuners i cannot enable CnQ without suffering stuttering when watching TV as well as in recorded TV. Im going to upgrade my HTPC soon anyway, but until then i am going to be running at full speed.

I have used RMClock to set a multiplier of x9(full speed) while setting the vcore to 1.1V. This seems to be working okay, i ran Prime95 for 30min until it was needed for HTPC duty, and so far it has not displayed any problems.

Is it usually the case that that E6 Venice cores can run at [email protected], or am i just lucky/waiting for my PC to crash?

Do anyone have an idea how much of the energy saved by running CnQ is achieved by this voltage reduction?

Btw, how come my computer crashes if i try to set a 4x @1.1V ?It runs fine at x5 @ 1.1V.

QuietOC · Post by **QuietOC** » Tue Apr 25, 2006 11:57 am

ziphnor wrote:Is it usually the case that that E6 Venice cores can run at [email protected], or am i just lucky/waiting for my PC to crash?

1.8GHz @ 1.1V is normal for an E6.

Do anyone have an idea how much of the energy saved by running CnQ is achieved by this voltage reduction?

The additional clockspeed reduction (1.8GHz -> 1GHz) might reduce cpu power usage when idle by an additional 44%.

ziphnor · Post by **ziphnor** » Tue Apr 25, 2006 12:10 pm

The additional clockspeed reduction (1.8GHz -> 1GHz) might reduce cpu power usage when idle by an additional 44%.

I wish i knew why my TV tuners stutter when the CnQ is enabled. I suspect they are using some unreliable timing mechanism, but people have used the DVB-S version of the card with SpeedStep without issue.

teejay · Post by **teejay** » Tue Apr 25, 2006 12:30 pm

Have you tried CrystalCPUID? It's software-based but the result is the same: dynamic undervolting & underclocking.

FWIW, my 3000+ E3 runs at 230 MHz x 5 at 0.9V while idle and at stock voltage and multiplier at high load (230 MHz x 9 x 1.4V). Not the max OC probably, but it runs very stable at this point.

QuietOC · Post by **QuietOC** » Tue Apr 25, 2006 5:01 pm

teejay wrote:Have you tried CrystalCPUID? It's software-based but the result is the same: dynamic undervolting & underclocking.

C'n'Q is always software based whether you use AMD's "driver" or another program to control it. CrystalCPUID works well, and it does let you tweak all the parameters.

ziphnor · Post by **ziphnor** » Wed Apr 26, 2006 12:08 am

Using the pre-configured RMClock profiles also gave me the stuttering problem, but i think ill try to just add a x5 1.1V P-state in addition to the x9 1.1V Pstate and see if it starts stuttering.

Will CrystalCPUID allow me to set my voltage lower than 1.1V, because RMClock wont (maybe an Asus A8N-E mobo limitation?).

Is there any tutorials on all the advanced options in RMClock and CrystalCPUID, because i have no idea what most of the options do.

ziphnor · Post by **ziphnor** » Wed Apr 26, 2006 1:33 am

I just tried Crystal it did allow me to choose lower voltages but there was no effect, CPU-Z still reports 1.1V. Is it just the motherboard that has this limitation? Its actually pretty annoying because if it can do [email protected] then im sure it can do 1.0 at less than 1V.

QuietOC · Post by **QuietOC** » Wed Apr 26, 2006 6:58 am

ziphnor wrote:I just tried Crystal it did allow me to choose lower voltages but there was no effect, CPU-Z still reports 1.1V. Is it just the motherboard that has this limitation? Its actually pretty annoying because if it can do [email protected] then im sure it can do 1.0 at less than 1V.

I have this problem with all the "E6" C'n'Q enabled processors on my BIOSTAR Tforce 6100. The "D0" Sempron 3100+ undervolted beyond that using CrystalCPUID on a SIS-based motherboard. It has something to do with the processor and maybe the motherboard.

nutball · Post by **nutball** » Wed Apr 26, 2006 7:37 am

ziphnor wrote:I just tried Crystal it did allow me to choose lower voltages but there was no effect, CPU-Z still reports 1.1V. Is it just the motherboard that has this limitation? Its actually pretty annoying because if it can do [email protected] then im sure it can do 1.0 at less than 1V.

I've had similar issues on one of my motherboards -- CrystalCPUID won't undervolt below 1.1V, but the BIOS will quite happily do so (and indeed CPU-Z confirms that the voltages set in the BIOS do hold). So try setting a lower voltage in the BIOS is you can.

teejay · Post by **teejay** » Wed Apr 26, 2006 12:12 pm

QuietOC wrote:C'n'Q is always software based whether you use AMD's "driver" or another program to control it.

Mmmm ok... I always assumed it had something vaguely hardware-ish about it, since you can turn it on and off in the BIOS.. shows how much I know

ziphnor · Post by **ziphnor** » Thu Apr 27, 2006 7:34 am

I have measured my system power draw and it turns out that running at 1.1V gives some savings, but 1Ghz vs 1.8Ghz at 1.1V is only ~1W difference, so i guess thats acceptable for now especially since its such a small part ofthe toral system power draw (just below 70W) in idle. For details see this thread:
http://forums.silentpcreview.com/viewtopic.php?t=31045

Venice 3000+ @ 1.1V

Venice 3000+ @ 1.1V

Re: Venice 3000+ @ 1.1V

Re: Venice 3000+ @ 1.1V