Sign in with
Sign up | Sign in
Your question

Is 1.44v too much for 24/7 usage?

Last response: in CPUs
Share
November 16, 2007 12:47:43 PM

Not under load, just having the computer on 24/7.

I've overclocked my computer and the autovoltage is around 1.44 (if I don't use auto it will crash prime95 within 5 hours).

Is this too much voltage? Will it significantly shorten my cpu lifespan or risk harming my mobo/cpu?

I'm using an e6850. 3.0@3.6.

More about : 44v usage

a b à CPUs
November 16, 2007 1:18:24 PM

well if you use autovoltage while overclocking you deserve to get your chip burnt :p 

just set it manually to the lowest stable voltage, autovoltage usually sets it unnecesseraly high...
November 16, 2007 1:21:20 PM

What are your temps?

What motherboard are you using?

Did you try manually setting it to 1.44v?

If not I would do that and run Prime and if it's stable then gradually reduce the voltage one step at a time until it becomes unstable.

With increased voltage comes an increased chance of Electromigration http://en.wikipedia.org/wiki/Electromigration

You have a very good CPU and depending on which board your using and what BIOS it has you should be able to get it stable at a lower voltage than that.
Related resources
November 16, 2007 2:13:55 PM

p35 ds3r, award bios



temp is 64C under MAX load, prime95 smallfft. blend is about 58C.

i hate testing prime95 cuz it takes friggin hours to find the instability. the thing is, auto-voltage auto-regulates all voltage on my mobo, not just CPU. when I adjust fsb voltage it can affect stability, as well as all that other stupid voltage. fsb, mch, and ddr voltage confuse as to their necessity when oc'ing.

Manually, I've tried 1.4125 and prime95 crashes after about 5 hours. temps with 1.4125 are virtually identical to autovoltage temps at around 1.44. My board also has significant vdroop.
November 16, 2007 2:33:10 PM

1.44V is fine for 24/7 use, as long as you don't intend to keep the CPU until it reaches antique status. ;) 

November 16, 2007 3:45:11 PM

Quote:
1.44V is fine for 24/7 use, as long as you don't intend to keep the CPU until it reaches antique status


So what is considered antique use? Would the cpu last about 4 years at 1.44?
November 16, 2007 6:02:23 PM

It should last for years but 1.44 is too high for 65nm processor. On the positive side unless you turned off speed step it will throttle way down when idle so it's likely not an issue. The reference voltage above is not the acceptable voltage for CPU like say ram voltage range.
November 16, 2007 6:10:49 PM

1.44v is too high? People keep their systems clocked at 1.45-1.55v using conroes. From what I've read I should be able to get 3.6 @ 1.4 or 1.41. Not sure why I can't.

I don't know about speed step, I looked for it in the BIOS but I couldn't find it, unless it's called something else in gigabyte mobos. CPU-Z reports vcore for the cpu LOWER under load and HIGHER when idle..
November 16, 2007 6:55:55 PM

I think it's C1E if I remember right. It downsteps the multiplier to 6 on my motherboard when there's no load. My E4300 runs at 3 Mhz (9 x 333), and when no load, multiplier goes to 6 and I'm running at 1998 Mhz, slightly above stock 1.8. My MB is p965-DS3 rev 2.
November 16, 2007 7:55:22 PM

bydesign said:
It should last for years but 1.44 is too high for 65nm processor. On the positive side unless you turned off speed step it will throttle way down when idle so it's likely not an issue. The reference voltage above is not the acceptable voltage for CPU like say ram voltage range.


It's only a 10% overvolt, well within 'acceptable' limits. You see some people running 1.5V with high end air or water cooling, and I don't see many dead C2Ds? Hmm...
November 16, 2007 7:57:13 PM

As long as you have plenty of cooling, it's fine.



November 16, 2007 9:54:31 PM

It's quite high (my opinion) for a 600mhz overclock.
You should try to reduce the voltage to stock (or around 1.25V) and work your way up until it's stable (saves a lot of time).
November 16, 2007 10:01:58 PM

Should I touch FSB/MCH voltage?
November 16, 2007 10:51:19 PM

Try stock settings up to say 1.35V cpu.

If it's still unstable you could test with a very small fsb overvolt.

MCH shouldn't really need overvolting at 400MHz.
November 17, 2007 12:53:57 AM

So set voltage to 1.35V and see the highest I can get running prime95 smallfft?

I couldn't imagine getting remotely close to 3.6ghz considering I would get errors a few hours in at 1.38.
November 17, 2007 2:05:43 AM

EricVPI said:
So set voltage to 1.35V and see the highest I can get running prime95 smallfft?

I couldn't imagine getting remotely close to 3.6ghz considering I would get errors a few hours in at 1.38.


Really the limit for air cooling is 1.5 Vcore. any higher than this and you processer won't last. If your using a high quality water cooling system then up to 1.6 Vcore is possible.
November 17, 2007 2:14:06 AM

2003 i set my 3.0c to 1.45v

they said the same things!

2007 24/7 it still works


c2d should be 1.41-1.45v
November 17, 2007 2:55:11 AM

1.44 is pretty high u know.
I managed to get 2.7Ghz out my q6600 with 1.15v (bios)
It ranges 1.09-1.12 (load-idle) 100% stable

EDIT: Thats on a B3 stepping too.
November 17, 2007 3:10:46 AM

Q6600 G0 3.0GHz (333x9) using 1.2750 in BIOS, and shows 1.2400 in CPUID, and runs perfectly fine under full Prime95.
a b à CPUs
November 17, 2007 3:40:42 AM

EricVPI, no CPU's are 100% identical. Every silicon semiconductor nano-circuit has slightly different basic properties of electronics such as resistance, capacitance, inductance, impedance, and transconductance, and there are nearly 292 million transistors in a C2D. Although two consecutive serial number CPU's from the same fabrication, with the same stepping may appear identical, they're yielded from different location on the silicon wafer from which they're manufactured, and like diamonds, each has it's own unique flaws.

Even though their dynamic operational characteristics may be very similar, no two CPU's will overclock to exactly the same stable maximum speed, at the same voltage, at the same temperatures. Additionally, in a dual core processor, one core will always become unstable before the other. Successful overclocking is achieved through small increments, and requires a methodical approach, and a great deal of time and patience. Regardless of effort, there are never any gaurantees; we can only say what is typical. There's no way to predict an overclock until you try it. That's just the luck of the overclocking draw.

Comp :sol: 
November 17, 2007 3:33:58 PM

If at 1.41V only crashes prime95 after 5 hours, I doubt it needs additional 0.03V to stabilize it. 1.42 would probably be enough
November 17, 2007 4:21:44 PM

You're fine.

I know people with 1.5v+ on water 24/7.

If it's going to fail "early" it will still be after you upgrade.

Hell my Pentium 133 MHz is still running as a firewall on someones network.
November 17, 2007 5:30:12 PM

EricVPI said:
Not under load, just having the computer on 24/7.

I've overclocked my computer and the autovoltage is around 1.44 (if I don't use auto it will crash prime95 within 5 hours).

Is this too much voltage? Will it significantly shorten my cpu lifespan or risk harming my mobo/cpu?

I'm using an e6850. 3.0@3.6.

While 1.44 volts isn't to much it will significantly shorten your CPU's lifespan. Most CPU's can last about 12 years if left to stock settings. At that voltage you could burn out your CPU in under 4 years. OCing no matter how little your trading higher clocks for lifespan. Upping the voltage from stock really cuts into your lifespan and should never be increased by more than 10%.
November 17, 2007 5:32:13 PM

elbert said:
While 1.44 volts isn't to much it will significantly shorten your CPU's lifespan. Most CPU's can last about 12 years if left to stock settings. At that voltage you could burn out your CPU in under 4 years. OCing no matter how little your trading higher clocks for lifespan. Upping the voltage from stock really cuts into your lifespan and should never be increased by more than 10%.


You are correct.. but cmon.

Like OCers really keep the same CPU for 12 years... :lol: 
November 17, 2007 5:45:50 PM

cnumartyr said:

Hell my Pentium 133 MHz is still running as a firewall on someones network.


My Pentium 133@187 worked for 7 or 8 years... CPU is probably still OK
November 17, 2007 7:00:50 PM

cnumartyr said:
You are correct.. but cmon.

Like OCers really keep the same CPU for 12 years... :lol: 

I have a Pentium 60 with the div error in working order thats about 12 years old. OCers keeping a CPU for 12 years most likely not but 6 years likely. The 10% suggestion is important on keeping about half the CPU's lifespan. Going higher really cuts into the lifespan quickly so I strongly suggest not going any higher.

Hafnium may give the Penryn a bit more than 10% but only a major materials changes makes this safe.
November 17, 2007 8:07:14 PM

Has anybody's Core2duo ever died because of electromigration? It's really rare I'd say...

I would think under 1.5v should be fine
November 17, 2007 8:28:00 PM

So let's say a C2Q's lifespan is 10 years under stock 2400MHz @ 1.3V
Let's say it works 2400MHz @ 1.07500V and 3000MHz @ 1.24375V.
How long will it last ? My bet would be 4-10 years.

Those old Pentium chips had a ton of material in them (0.25um technology).
They were a lot less likely to die due to electromigration.

While 1.44V is high it is within Intel's spec but I wouldn't want to run my chip on 1.44 anyway. The choice depends greatly on how long you plan to have your CPU (at least 3 years for me).

If you get errors 5 hours into your Prime95 Run you should play around with the RAM Transaction Booster/RAM Performance Enhance.
When set to Extreme (Gigabyte's Transaction Booster with Standard, Turbo, Extreme options) I got errors after 5 hours of Prime95. In a review on anandtech I read that EXTREME was useless when overclocking (I'll try to find it).
While I was only overclocking the CPU the errors where always within the first 90 minutes or not at all.
November 17, 2007 8:36:35 PM

life span of computer is very short and gets short each year
in july of 2006 the c2d came out
in nov of 2006 the qx67600 came out
in july of 2007 the q6600 go came out
in nov the qx6950 comes out

in `1 year all the above chips will be obsolete the value of the $1300 qx6950 will be around $400 on ebay while a qx6700 may fetch $200

turn your voltage up on low cost cpu's like the q6600 and replace it in 3 years - so the whole discussion is mute.

the fact is the increase in performance is increasing explanationary therefore the required life is dropping in proportion -


THE KEY IS TEMPS - RUN YOUR VOLTAGE SO THE CPU RUNES AT 65C OR LOWER AT MAX OUT PUT! IT WORKS!
November 17, 2007 8:43:00 PM

dragonsprayer said:
life span of computer is very short and gets short each year
in july of 2006 the c2d came out
in nov of 2006 the qx67600 came out
in july of 2007 the q6600 go came out
in nov the qx6950 comes out

in `1 year all the above chips will be obsolete the value of the $1300 qx6950 will be around $400 on ebay while a qx6700 may fetch $200

turn your voltage up on low cost cpu's like the q6600 and replace it in 3 years - so the whole discussion is mute.

the fact is the increase in performance is increasing explanationary therefore the required life is dropping in proportion -


THE KEY IS TEMPS - RUN YOUR VOLTAGE SO THE CPU RUNES AT 65C OR LOWER AT MAX OUT PUT! IT WORKS!


Pretty much what I was getting at... By the time my Q6600 dies even if I ran it at 1.6v I could by a better replacement for 200 bucks.
a b à CPUs
November 17, 2007 9:21:27 PM

I strongly agree with dragonsprayer and cnumartyr. If you do not overvolt more than 10% (Vcore 1.5) and you're diligent about maintaining safe temperatures, then your CPU will die of neglect and old age years before electromigration begins to affect overclock stability. I recommend that you stop agonizing over it; just crank it up and run it. It'll be fine.

Comp :sol: 
November 18, 2007 9:24:31 PM

dragonsprayer said:
life span of computer is very short and gets short each year
in july of 2006 the c2d came out
in nov of 2006 the qx67600 came out
in july of 2007 the q6600 go came out
in nov the qx6950 comes out

in `1 year all the above chips will be obsolete the value of the $1300 qx6950 will be around $400 on ebay while a qx6700 may fetch $200

turn your voltage up on low cost cpu's like the q6600 and replace it in 3 years - so the whole discussion is mute.

the fact is the increase in performance is increasing explanationary therefore the required life is dropping in proportion -


THE KEY IS TEMPS - RUN YOUR VOLTAGE SO THE CPU RUNES AT 65C OR LOWER AT MAX OUT PUT! IT WORKS!

Your pretty much correct but not everyone can buy a new CPU, mobo, and RAM every 3 years. The CPU socket changes and your force to buy new atleast those 3 items. The thing here is the drop in lifespan isn't leaner but more of on a segma curve. A small drop in the top OC could add years to your CPU's life. Got to remember these systems are quite useful even after 3 or 4 years.

I like OCing my new PCs for top performance then when its ready to be replace, after 3 or 4 years, I set it to stock and use it for file servers or my sons game server. You can put an old Pentium in the back of your car and use it as an MP3 player.

Don't get me wrong 10% is safe but its about the max I would suggest for the average OC'er. Its not to say the OP's OC to 3.6GHz couldn't be done within that 10% given a +3 bin.
November 18, 2007 10:05:21 PM

If you are always looking for a gaming rig, a upgrade everything 3 years sounds reasonable
November 18, 2007 10:17:18 PM

itotallybelieveyou said:
If you are always looking for a gaming rig, a upgrade everything 3 years sounds reasonable



Every 3 years is pretty rediculous if you want to game with decent graphics settings.

Mid-range every 2 is more like it.

If you can't afford $1500 for a new rig every 2 years.. don't OC. Infact.. if you can't afford a new rig, don't OC period. You know there are risks involved with OCing, period.
a b à CPUs
November 18, 2007 10:26:19 PM

People are so finicky about voltage. Run it at 1.65V+ I say!

If my mobo still worked I'd start pushing 1.8V through my A64 3700 to see if I can get it to 3+GHz stable on air (it's not a good chip).

But seriously, if you aren't running 75C at load, theres really no problem. It will not seriously shorten the lifespan of the CPU, you may lose a year or 2 at the most with 24/7 operation, it is only a small voltage increase.

Also, consider what "normal" full load temps you are getting. TAT may give you 70C, yet running a multi-threaded game may get you to 58-60C. 70C is bad, but unless you bought your PC to run TAT, who cares?
November 18, 2007 11:00:35 PM

So long as you are operting within Intel's stated design standards, there is not any risk to your CPU.

They publish both safe temperatures and voltages.
And your 1.44v is well withing that range.

With that voltage and speed, however, you may need good cooling to keep temps in check.
November 19, 2007 2:47:31 AM

of coarse i am i am mr right (my wife told me that) so i guess i am always right?

wait - i am always right accept for when i am wrong since i say i am wrong then i must be 100% right.

that confirms it - i am 100% right


crank your cpu 1.42-1.45 for most c2d duals and g0 quads - the b3 are another story!

you get a highly repsonsive computer with proper overclocking - 1.6v is a good question why not go up to 1.6v?

answer: you do not need too and there is very little benfit
November 19, 2007 2:53:10 AM

water: i do not think water makes much of difference - the limit is the thermal compound, the mass of the cpu cooler above the chip

good heat pipes can cool high voltage to almost the same with air as water- in the real world outside of running orthos for fun quad cores do not see that much load. increasing your speed will greatly improve your responsiveness.

I took the position of only building and selling overclock computers in 2003 to prove that it can be done - every system we tune to the sweet spot which is can be found on these fourums.

i think 3.6ghz with quad is pretty sweet

3.7 is nice with an x6800
3.4 for most dual core chips


amd? i don't due wallmart systems so i do not no much - but i was not impressed with the fx-60 overclocking 1 or 2 ago and amd chips are just not nearly as smooth. the sweet spot is 2.6-2.8ghz same as the fx-60


ramble ramble.......
a b à CPUs
November 19, 2007 10:00:26 PM

Well the FX-60 was basically teetering at the edge of the K8 architecture limits, so that was expected. You only buy those chips if you want e-peen.
November 19, 2007 10:51:29 PM

Is 3.6GHz really necessary? What do you do that requires such speed? I'd say drop down to 3.2GHz and find the lowest stable voltage.

But in answer to your question, 1.44V probably won't harm it unless your temps get too high (yours are getting up there at 65C).
!