The Official X2 Overclocking Discussion Thread

[scott14]

(I put this on AMD's forum but got no reply.)

How damaging/degrading would it be to go above AMD's max recommended voltage of 1.30V to 1.35V with my new air cooled 4200+ X2 s939 ?

Is it real destructive to go higher?

The highest I'm able to go and still obey AMD's recommended max voltage is 1.35v X 104% = 1.4040 volts as calculated, but actually dithers between 1.344 & 1.360 volts as reported by cpu-z. So that's right at the max recommended voltage red line.

At that level I can only OC to htt 239 X 11 = core clock of 2629 MHz, temp 54 deg C while OCCT! stress test is running.

However, by increasing the voltage above 1.4 actual, I can get the cpu core clock up to 2706 or probably higher.

If it's not harmful, I'd like to at least run it at 1.250 X 113% = 1.4125 as calculated, or 1.376 as actually reported. Or higher, say 1.3 X 110% = 1.43 calc, 1.39 actual. Or higher.

Guidance on voltage limit will be appreciated. Thank you.

(Somewhere I read that actual voltage itself wears down a cpu, quite distinct from temperature effects.)

[Lenny_Nero]

How damaging/degrading would it be to go above AMD's max recommended voltage of 1.30V to 1.35V with my new air cooled 4200+ X2 s939 ?
Is it real destructive to go higher?
(Somewhere I read that actual voltage itself wears down a cpu, quite distinct from temperature effects.)

There are 2 sure fire ways of killing a CPU*, well any electrical component for that matter ...heat and voltage, well 3 if you want to count hitting it with a hammer as well.
As for the AMD specs I am sure the they have room for the cheepo main boards and PSU's that dont have a very stable hold on the voltage and send loads of ripple and odd order harmonics down the lines. So there is almost no doubt that a small over voltage, as long as the running temps are not pushed up will cause any problems. In fact I am under volt'ing mine (its been hot and running very long jobs without me around) and still getting ~2.4 GHz. I normally run at 1.38 to 1.4 ish (depending on what does the reporting) and that is happy to keep my 4200 at 2.6 video encoding for weeks on end, and I know of people that are happy to run at above 1.5 on air (most with optys) but TBH all they do is play games and the like so its not a big deal if there is a big spike in the supply and the chips get toasted. That said, I think its worth getting a good PSU to help smooth the line, the poster above 'd1b1' looks as if a no name PSU is getting used and I am surprised that they have done so well with it so far, the power supply is a very underrated piece of kit dont skimp on it.

I have seen some people put 1.7V thru but I am not even going over 1.5 with my chips because I use my computers for a long time, but there is some sort of myth that chips ware out with use it does not happen. I have 30 plus years with computers and my dad started working with them in the late 1950's and has never heard of a chip that just get worn out, they dont, there will always be something else that caused the chip to fail see above*

If the process goes quite a few orders of magnitude smaller the life might come down from 10,000 years to 9999. But I still think that most people will want to upgrade before they have used 1% of their CPU's lifespan.

[FlawleZ]

There are 2 sure fire ways of killing a CPU*, well any electrical component for that matter ...heat and voltage, well 3 if you want to count hitting it with a hammer as well.
As for the AMD specs I am sure the they have room for the cheepo main boards and PSU's that dont have a very stable hold on the voltage and send loads of ripple and odd order harmonics down the lines. So there is almost no doubt that a small over voltage, as long as the running temps are not pushed up will cause any problems. In fact I am under volt'ing mine (its been hot and running very long jobs without me around) and still getting ~2.4 GHz. I normally run at 1.38 to 1.4 ish (depending on what does the reporting) and that is happy to keep my 4200 at 2.6 video encoding for weeks on end, and I know of people that are happy to run at above 1.5 on air (most with optys) but TBH all they do is play games and the like so its not a big deal if there is a big spike in the supply and the chips get toasted. That said, I think its worth getting a good PSU to help smooth the line, the poster above 'd1b1' looks as if a no name PSU is getting used and I am surprised that they have done so well with it so far, the power supply is a very underrated piece of kit dont skimp on it.

I have seen some people put 1.7V thru but I am not even going over 1.5 with my chips because I use my computers for a long time, but there is some sort of myth that chips ware out with use it does not happen. I have 30 plus years with computers and my dad started working with them in the late 1950's and has never heard of a chip that just get worn out, they dont, there will always be something else that caused the chip to fail see above*

If the process goes quite a few orders of magnitude smaller the life might come down from 10,000 years to 9999. But I still think that most people will want to upgrade before they have used 1% of their CPU's lifespan.

I agree, but you have to keep in mind that the hardware is rated from the manufacturer in longevity under certain thermal exposures. The higher the duration of high thermal exposure, the shorter the life expectancy for any piece of silicon will be.

[Lenny_Nero]

I agree, but you have to keep in mind that the hardware is rated from the manufacturer in longevity under certain thermal exposures. The higher the duration of high thermal exposure, the shorter the life expectancy for any piece of silicon will be.

Yes, that is what I said, and I think that the point that most people and you miss ...is that if you dont over volt or over heat the chip it will go on for as long as you need it to. Most people will want to upgrade a very long time before they have used a few % of a CPU's life time. Processors dont wear out with a few years use.

I have an old 300B Celeron that has been running for the last 7~8 years with a 100% overclock (no core voltage change) it runs 24/7 and because of the very strong crypto I run on my network (its used as a print server/router/firewall) it runs above 85~90% CPU all of the time, I expect that it will be doing so for another 10 years without putting a dent in its longevity and that is with the standard HS/f, as long as you use good cooling and keep things clean and dust free there will be no problems.

I expect to get an easy 10 year life out of my 4200 and I very much doubt that running it at or just below a 1.4 core voltage will do anything to stop that.

[scott14]

Yes, that is what I said, and I think that the point that most people and you miss ...is that if you dont over volt or over heat the chip it will go on for as long as you need it to. Most people will want to upgrade a very long time before they have used a few % of a CPU's life time. Processors dont wear out with a few years use.

I have an old 300B Celeron that has been running for the last 7~8 years with a 100% overclock (no core voltage change) it runs 24/7 and because of the very strong crypto I run on my network (its used as a print server/router/firewall) it runs above 85~90% CPU all of the time, I expect that it will be doing so for another 10 years without putting a dent in its longevity and that is with the standard HS/f, as long as you use good cooling and keep things clean and dust free there will be no problems.

I expect to get an easy 10 year life out of my 4200 and I very much doubt that running it at or just below a 1.4 core voltage will do anything to stop that.

A) Re: Voltage to cpu
Which vcore reading do you most believe?

My DFI mobo bios says one thing, say for example 1.43, while at the same time both cpu-z and Everest say something different, like 1.408 v.

Which one would you go by?

B) Re: cpu temperatures
When there are 3 different temp readouts given for (1) CPU, and (2) CPU core 1, and (3) CPU core 2, and they're all different, which one would you go by when you're trying to keep cpu temperature below some level, like say maybe 55 deg C during stress testing max load?

[mad_skills]

anyone knows what's a probable max for a G1 steping Brisbane on a high end air cooler like Tuniq Tower or TR Ultra 120 eXtr. ?