Safe Voltages for Venice?

[bjc112]

Originally posted by: pnb263
289x9=2601 at 1.32V... 100% stable [meaning there is nothing I can do to cause it to crash short of pulling parts out or throwing something on it, like liquid... I ran Prime95 for 4 days straight with no errors or warnings, and I had video encoding going on in the background at the same time... ] OR i can run 317x9-2853 at 1.65V with 80% or so stable... with my VCore at 1.6, I can do 300x9=2700 100% stable... I just dont like to push the CPU that high with voltage because I need 110% stability and for the CPU to last 5 years or more...

Oh well, at least I know that I can count on the CPU with voltages like I run them, at 1.32

So you saying it takes almost 20% more voltage to get 2700?

Its rock stable @2600/ 1.32v Which is DAMN low but not 2700 @ 1.6v????


Seems odd to me..its definitely stable at 1.32?

 

[Diogenes2]

Originally posted by: bjc112
1.67 is too high for my taste, but it should still last for a couple of years..

unless you have EXCELLENT case cooling, I find 50C loaded @ 1.67v with only a Zalman 7700
kinda BS.. But you never know..

With an XP-90 @ only 1.49-1.51v I can't even get below 52C..

Back to my 3500+ at 1.7v ...

I have a 92MM Zalman pushing about 50CFM as well..

I have a ZALMAN CNPS7000B, I keep the case open..
I have a temp probe up against the IHS and I'm reading ~47 loaded..
Core temp might be a couple degrees hotter, but I'm comfortable with any stable temp this side of 60..

 

[theMan]

you definately shouldnt run that voltage on someone elses computer! maybe max out at ~1.55v. all those volts moving through the millions of tiny transistors wears them down, and eventually, they cant do their job. current leaks out into places it shouldnt, and then there is the cpu funeral. it doesnt matter how cool it is. those people running 1.8v on a venice with phase change dont expect it to run for over year at the most.

bjc112: thats crazy!!! you have the sweetest overclock ive ever seen! wow! i wonder what that can do on a 10% voltage increase! you may be able to hit 3000mhz!!

 

[imported_g33k]

Originally posted by: bjc112

Originally posted by: pnb263
289x9=2601 at 1.32V... 100% stable [meaning there is nothing I can do to cause it to crash short of pulling parts out or throwing something on it, like liquid... I ran Prime95 for 4 days straight with no errors or warnings, and I had video encoding going on in the background at the same time... ] OR i can run 317x9-2853 at 1.65V with 80% or so stable... with my VCore at 1.6, I can do 300x9=2700 100% stable... I just dont like to push the CPU that high with voltage because I need 110% stability and for the CPU to last 5 years or more...

Oh well, at least I know that I can count on the CPU with voltages like I run them, at 1.32

So you saying it takes almost 20% more voltage to get 2700?

Its rock stable @2600/ 1.32v Which is DAMN low but not 2700 @ 1.6v????


Seems odd to me..its definitely stable at 1.32?

Oops NVM

 

[imported_g33k]

Originally posted by: theman
you definately shouldnt run that voltage on someone elses computer! maybe max out at ~1.55v. all those volts moving through the millions of tiny transistors wears them down, and eventually, they cant do their job. current leaks out into places it shouldnt, and then there is the cpu funeral. it doesnt matter how cool it is. those people running 1.8v on a venice with phase change dont expect it to run for over year at the most.

bjc112: thats crazy!!! you have the sweetest overclock ive ever seen! wow! i wonder what that can do on a 10% voltage increase! you may be able to hit 3000mhz!!

You are right of course. If it was mine, no problem, but since it is not...

I dialed down vcore to 1.52volts which is less than a 10% increase. So it should be good there. At this voltage I was only abe to squeeze 2500mhz out of this chip.

 

[theMan]

that kinda stinks, but at least it is a lot faster than what you paid for! :beer:

 

[zendari]

I use 1.6 V onm my new venice.

 

[KoolDrew]

I still don't understand why it would be ok to go 1.65V on water, but not on air. A good air cooler such as the Zalman 7700cu and water cooling can be the difference of maybe 5C. But if temps don't matter than why would higher voltage be ok with different cooling?

Sorry. I should of reworded what I said. Temps do matter to an extent, but once you start getting into higher voltages low temps will not save your CPU. Also a Zalman 7700cu is not going to be near as good as water cooling. The difference would be more like 15C.

 

[bjc112]

Originally posted by: theman
you definately shouldnt run that voltage on someone elses computer! maybe max out at ~1.55v. all those volts moving through the millions of tiny transistors wears them down, and eventually, they cant do their job. current leaks out into places it shouldnt, and then there is the cpu funeral. it doesnt matter how cool it is. those people running 1.8v on a venice with phase change dont expect it to run for over year at the most.

bjc112: thats crazy!!! you have the sweetest overclock ive ever seen! wow! i wonder what that can do on a 10% voltage increase! you may be able to hit 3000mhz!!

This chip definitely has 2.9ghz in it but my board can't go higher than 314-315HTT without quitting on me..



With 1.6 or even 1.65 3ghz is probably within range!

 

[CHOPPER GOD]

Venice 3200 LLBLE 0516EPEW (Newegg) @ 2500 stable @ 1.45 with xp-90c @ 32 idle 41 prime
I want to know max volts too! I prefer my own cool n Quiet by using clock gen. I only run 2.5 during gaming from the desktop.
I want more Overclock but I am actually scared of going to 1.6v OMG!...
I cant get stable after 2550 ish with this voltage..Should I try 1.55 maybe...Will it really shorten My lifespan?
Im actually very happy at 2.5....its only greed and jealousy that makes me want to hit 2.7 on air...I can't stop tinkering heh..If im drastically gonna reduce Lifespan for an extra 100-200 Mhz then forget it as I should be happy with my allready 500 Mhz OC....
Input?

 

[bjc112]

1.55 isn't one bit scary.. There are people here that have been running 1.55v through 90nm for a couple years.. Or close to it..

1.65 -> 1.7v gets REALLY hairy.. I kinda want to run one at 1.7 just to see how long it last...

 

[CHOPPER GOD]

Originally posted by: bjc112
1.55 isn't one bit scary.. There are people here that have been running 1.55v through 90nm for a couple years.. Or close to it..

1.65 -> 1.7v gets REALLY hairy.. I kinda want to run one at 1.7 just to see how long it last...

whew! ok off i go to 2600

 

[pnb263]

Originally posted by: bjc112

Originally posted by: pnb263
289x9=2601 at 1.32V... 100% stable [meaning there is nothing I can do to cause it to crash short of pulling parts out or throwing something on it, like liquid... I ran Prime95 for 4 days straight with no errors or warnings, and I had video encoding going on in the background at the same time... ] OR i can run 317x9-2853 at 1.65V with 80% or so stable... with my VCore at 1.6, I can do 300x9=2700 100% stable... I just dont like to push the CPU that high with voltage because I need 110% stability and for the CPU to last 5 years or more...

Oh well, at least I know that I can count on the CPU with voltages like I run them, at 1.32

So you saying it takes almost 20% more voltage to get 2700?

Its rock stable @2600/ 1.32v Which is DAMN low but not 2700 @ 1.6v????


Seems odd to me..its definitely stable at 1.32?

Yea, Seems kinda odd to me too, but I have had zero problems at 1.32 and 2601MHz. I dont know if the problem lies in my southbridge or the ram, but to eek out 2700 I need to push the CPU voltage to 1.6. (100% stable).

I just wish I had a different motherboard to test the ability of the CPU from a different point of view. I know the motherboard I have is rock solid like I run it, but to assume that the CPU has a limit would be foolish when I can run it the way I do.

If I were to go with a better overclocking motherboard and a liquid/phase change setup, then I think I really could find the limit of the CPU that I have. But, since those are really not the way I wish to go, Ill just be thankfull that I have a CPU that will do what I have it set at and not crash.

--update--
thus far, i have had a perfect uptime of about 3 weeks. I have done about 300 hours of encoding (video) and have done three times as much in capture. If I can get water cooling later this summer, I might have to push the CPU a bit more just out of curiosity.

I am now running the CPU at 2700 and 1.55Vcore - seems that its about all I can do for the voltages that I feel safe at...

long live AMD!

 

[Lonyo]

Originally posted by: pnb263

Originally posted by: bjc112

Originally posted by: pnb263
289x9=2601 at 1.32V... 100% stable [meaning there is nothing I can do to cause it to crash short of pulling parts out or throwing something on it, like liquid... I ran Prime95 for 4 days straight with no errors or warnings, and I had video encoding going on in the background at the same time... ] OR i can run 317x9-2853 at 1.65V with 80% or so stable... with my VCore at 1.6, I can do 300x9=2700 100% stable... I just dont like to push the CPU that high with voltage because I need 110% stability and for the CPU to last 5 years or more...

Oh well, at least I know that I can count on the CPU with voltages like I run them, at 1.32

So you saying it takes almost 20% more voltage to get 2700?

Its rock stable @2600/ 1.32v Which is DAMN low but not 2700 @ 1.6v????


Seems odd to me..its definitely stable at 1.32?

<snip>
long live AMD!

Slightly different, but I have a mobile XP which hits 2400/2450 @ 1.7~1.75v or so, but to get 2500+MHz stable I have to go to about 1.93v or more.
As you near the end of its clock speeds, you seem to get huge voltage jumps.

 

[Diogenes2]

Originally posted by: Lonyo

Slightly different, but I have a mobile XP which hits 2400/2450 @ 1.7~1.75v or so, but to get 2500+MHz stable I have to go to about 1.93v or more.
As you near the end of its clock speeds, you seem to get huge voltage jumps.

This is interesting.. My 3500+ I mentioned above is P95 stable to ~2550 at 1.4v..

I have to go to 1.7 to get stable at 2600.. It will go on up to 2686 with 1.7..

 

[pelikan]

Originally posted by: g33k

Originally posted by: bjc112
1.67 is too high for my taste, but it should still last for a couple of years..

unless you have EXCELLENT case cooling, I find 50C loaded @ 1.67v with only a Zalman 7700
kinda BS.. But you never know..

With an XP-90 @ only 1.49-1.51v I can't even get below 52C..

I have a 92MM Zalman pushing about 50CFM as well..

I have a Lian-Li PC60 and it is an excellent case. No BS. I peak at 50C at 1.67 volts according to Smart Guardian on my DFI Ultra-D.

I lowered voltage to 1.61 and I had to drop htt to 255. But my freind will still be very happy and I don't have to worry about his proc crapping out in a year or so.

Wow, those temps are very good then.

[/quote]


You guys have different motherboards so your temps are not comparable. Even if you had the same motherboard with the same parts and the same ambient temperature, you may still get quite different temp readings. For example, I've seen a 10C temp difference with a new bios before (due to re-calibration of the temp sensors in the new bios).