How much voltage can I "safely" pump into my cpu?

C

cosgrove

n00b
Joined
Feb 19, 2006
Messages
23
Max voltage kinda depends on the cooling and cpu. for convenience, here is what i consider the max safe overvolt from stock voltage based on your cooling type:
Stock: +0.1v
Good Air: +.15v to +.2v
Water: +.2v
Click to expand...

I've got a DFI lanparty ultra, x2 3800+, XP-90c with a panaflo 92mm. With nothing overclocked, I idle around 25C and load somewhere in the 30s. I've followed the OCing guide and managed to get 2.55GHz on my cpu by bumping the voltage up by 0.125 (load with 2x p95 is like 37C iirc). Does my cooling setup qualify as "good air" thus allowing me to "safely" go up by 0.2V on the cpu core?

I guess a better question is how much a 0.2V increase will affect the life of my cpu. I've heard that it will drasticly decrease the lifespan, in which case I would probably want to save OCing for when I really start to notice performance loss.
 
Which voltage readout should I trust? Whatever I set it to in the bios settings? The reading in SpeedFan? The reading in the bios? Whenever I set it to something, the reading in speedfan is always a little lower.
 
I run an XP90 with a smaller fan, and I've pumped as much as 1.575 through it just fine (A64 3000 winchester core). Though I wouldn't recommend that. I usually put about 1.5 in order to reach my 2.46 overclock.

As stated above, keep it under 1.5 unless you have water. You shouldn't need more than a couple bumps to reach the threshold of the core.
 
i have had 1.6v on stock cooling on my 3500+ idle ~34 full load ~45 but i backed it down to stock :p
 
ikurox said:
i have had 1.6v on stock cooling on my 3500+ idle ~34 full load ~45 but i backed it down to stock :p
Click to expand...

Well of course you CAN do that, but obviously if you kept it that way and it wasnt sufficiently cooling it (which would depend on its frequency) youd run into problems quickly. People say 1.5v with the A64s referring to the highest your going to want to have with sufficient air cooling for long peroids and not reduce the cpu lifespan enough for it to matter. If youve got a very good air cooling solution with very low temps, you really could dabble above 1.5 as its the temp you need to pay attention to (electromigration happens with air, water, or anything else you use).
 
It just depends on how much of a risk taker you are. Will you shorten the life of your cpu? Sure, but then you need to ask yourself "how long do I want my computer to last and when will I upgrade again"? If you are into short-cycle upgrading (every 2 yrs or less) then pump it up and enjoy the power. Just be aware you may need to replace a cpu if your temps go nuts. My sig system is at 1.64v (1.4v stock) at 33.5C idle and 42C load. Since the system does not run more than 15-20 hrs a week and the cpu is cheap ($120) I'm willing to gamble. If you want to push it get a very capable cooling solution and go for it. Its only money and you can't take it with you :p
 
Thanks for the great replies concerning voltage guys! How about which voltage reading to trust though? BIOS, speedfan, or just what i set it to?
 
having watercooling or phase change doesn't make it any safer to run more voltage to your processor, they are just more capable of cooling the excessive heat that comes from higher voltages, thats where those guidelines come from.

I would say go with the readings from the BIOS. But you need a voltmeter to be sure.
 
I'd second the bios, however, I have had issues with bios vs. CPU-z.
On my old Asus K8N4-E Deluxe I'd set bios to 1.55v but CPU-z would read 1.52v. Now with my new Abit NV8 I set bios to 1.60v and CPU-z reads 1.63v. Its a bizarre situation. A voltmeter would solve the mystery but I'm not that concerned about it.
 
I ran 1.76v at 2.94 Ghz for about 20 minutes before the system locked up. Now running 2.79Ghz at 1.52v. I can get it up to 2.73Ghz at 1.48v which I think is probably the safest solution, I want that damn 2.80 threshold though!

I am hoping that if I chill the water in my loop it will increase system stability and let me reach 2.8 without pumping too many volts into the proc.
 
any one ever actually kept an overclocked chip long enough to ruin it?

I wonder if its all talk and "what should happen" I overclock everything in my comp and as a result of that I tend to not keep it that long.

I never had something blow up that was overclocked or stop working unless I did something stupid as hell to it. but if I had a stable overclock it worked fine till I upgraded anyways.

I just wonder real world how much it effects stuff.
 
The only thing I ever did was burn a 512 mb stick of cheap ram because I volted it from 2.6v to 2.9v. The other stick was and still is just fine. I oc my cpu's at work and at home. The 3700+ Clawhammer runs at 1.625v-1.65v (215x12 - crappy K8V mobo) all day, every day and idles at 40C. I f I game on it the temps go to 50C. No big deal. Its been doing this for a year with no problems. I'll more than likely get a new system before this one ever thinks of crapping out. Hardware comes and goes so fast its worth it to me to get the most out of it I can. 2-3 years of heavy use won't hurt it in my opinion.
 
Hell, I gave a Celeron 300 @ 450 to my dad and it's still running strong to this day. The worst thing with a overvolt is frying the cpu so if you managed to get stable and keeping the load temps under 55C, you are good to go.
 
Dont trust anything but a multimeter for your voltages. My last mobo was off by .1 under load. It said 1.55 and it was really 1.65
 
Only one, an XP 2100+ Palmino that I had overvolted for a couple years and while the temps were ok for that series, I didnt take care of it as I would my main system and it finally croaked from it. It was quite a fair amount of voltage but I cant remember the exact amount, needless to say it took some serious abuse before it died.

stealthy123 said:
any one ever actually kept an overclocked chip long enough to ruin it?

I wonder if its all talk and "what should happen" I overclock everything in my comp and as a result of that I tend to not keep it that long.

I never had something blow up that was overclocked or stop working unless I did something stupid as hell to it. but if I had a stable overclock it worked fine till I upgraded anyways.

I just wonder real world how much it effects stuff.
Click to expand...
 
I running 1.68-1.72v daily 24/7 on a venice on water.Always 100% load too cause I fold.
42-43c load. Ill let you guys know when/if it dies :p
 
wusses...

pumped 1.66 through my venice, 1,43 through my Toledo and 1.8 through my throughbred-b and bartons.

all are still ticking over nicely.
 
Fawkes said:
wusses...

pumped 1.66 through my venice, 1,43 through my Toledo and 1.8 through my throughbred-b and bartons.

all are still ticking over nicely.
Click to expand...

Hehe qft. 1.9v for bartons :p
 
I've got a few A64's here and I'm finding that once they hit their ceiling, going much beyond 1.45v doesn't give all that much advantage.

Case in point, my 3200+ which is w/c with a DD TDX block and dual 120 rad. I've given it as much as 1.65v but still cant get it stable at much more than 2700. The thing runs sub 30C even at 1.65 (according to DFI sensor, so take that with a grain of salt). Contrast this with the fact that it runs 2660 happily all day long at 1.38v. I've had my RAM, mult, and HTT all over the place so I'm pretty sure it's actually the CPU which is limiting me.

Some here will disagree, but for me, i feel the o/c you get at 1.45v is pretty telling about the top-end of what your particular CPU is good for as far as stable day-to-day usage. I suppose this is especially true on air cooling.

You might be able to squeak a few more Mhz out of it by drastically overvolting, but IMHO going much beyond 1.55v seems dangerous unless it's getting you hundreds of Mhz in return.
 
Some here will disagree, but for me, i feel the o/c you get at 1.45v is pretty telling about the top-end of what your particular CPU is good for as far as stable day-to-day usage. I suppose this is especially true on air cooling.
Click to expand...
I have to disagree with you there, I'm afraid. My 3000+ Venice is very responsive to voltage adjustments. When I had my Asus K8N4-E Deluxe mobo with the original bios (v1003) I was limited to 1.44v because of a bios bug and my best oc speed was 250x10. I upgraded the bios to v1011-003 (beta) and the max. bios increased to 1.55v (1.52V as reported by CPU-z). This resulted in an oc of 256x10. Then I upgraded to my present mobo (the Abit NV8) which has much better voltage options. I've set the vcore in bios up to 1.62v (1.65 as reported by CPU-z) and achieved oc's of 265x10 and 294x9 with a max bench of 270x10. At 270x10 it will boot up, run apps and surf the net with ease but it won't complete the Superpi 1M test. I believe I can stabilize it at 2.7ghz and will be doing some further testing including another voltage bump to 1.64 to 1.66 in bios. I am on air cooling (see sig) and the temps reported so far are 33.5C idle and 42.5C load which are great temps for the voltages I'm running. I've seen screenshots of the 3000+ skt 754 Venice at 10x280 @1.66v. Of course, that was quite probably a suicide run but who knows? In any event I believe that more can be had from 1.55v-1.6v settings without taxing your equipment too much. Its been working for me so far. The only question one has to ask oneself is "is the risk worth it?" Is a .2v bump worth the extra speed? That's a matter of personal preference. If you are not comfortable pushing the envelope then no, it isn't. But if you are like me and want to see what the limits are, then yes, it is. I won't villify anyone who stops short and says "ok that's enough for me." We all have a comfort zone and we know where it is. I'll congratulate you and say "well done!" But as for me I'm willing to blaze the trail ahead and keep pressing on and once I reach the top I'll let you know how its working and how I did it. Maybe I'll have a setback or two but hey, its like they say, "fortune favors the bold!"
 
You must log in or register to reply here.
Back
Top