max voltage for 'b' Thoroughbred Athlon

[Jake the Dog]

I want to start seeing just how far I can push my current 1800+ 'b' Athlon. currently I have it running @ 2071MHz (166MHz FSB x 12.5) on 1.725v max temp is 46C. I'm going to start raising the FSB next and presumably as I do so I'll need to up the voltage on the CPU. with regards to voltage; is there a known limit these CPU's can take or is it more a matter of they can take as much voltage as running temps will allow? eg if I clock it to a max temp of 60C but this requires say 1.85v, is this workable? I know I'm shorting the lifespan of the CPU but that is a risk I'm prepared to take as long as I can get 12 months out of it.

 

[time]

I've seen someone mention an AMD rated limit of one-point-nine-something (1.92?), but I haven't downloaded the AMD processor PDF to check, so it could be just their imagination.

Abit limited their boards to 1.85V, for whatever reason, although other boards can go much higher.

Anandtech suggested that the high one-point-sevens was enough for most people.

I did see one guy pushing one to 1.95V, in his words "22% overvoltage". His temps were getting high, but that could be due to other factors. Personally, I wouldn't rely on temperature alone as a warning.

I don't think you'll see any advantage above 1.95V anyway. You can usually get an idea where a CPU is going to top out as you ramp everything up.

 

[honold]

abit limits all their bios to some semblance of sane, but there are always hacked versions out there that enable you to increase it

i don't like anything that idles over 45c

 

[blakerwry]

i don't like anything that idles over 45c

like cars?


My CPU doesn't idle anymore anyway... personally, I think 1.8 is the fastest I'd even test with... 1.7-1.75 is the point where I'd stop. Doesnt the "b" run at 1.60v by default? which models ran at 1.65v?

 

[Jake the Dog]

OK thanks. I'll give it a go soon. looks like I'll be fine since I won't be going over 1.85v or 60C under load, which ever comes 1st.

 

[CityK]

AMD's PDF

Using the documents page numbers (as opposed to Adobe's):

Page 23 (table of electrical & thermal properties for CPUID 681) gives your processor Vcc_core = 1.60V....nothing new there, we all knew that. Note this value is later refered to as Vcc_core_Nom (nominal voltage) and the reason why is evident from the waveform diagram displayed later on page 35.

Page 34 (Vcc_core Characteristics) gives:
Vcc_core_DC_Max = max static voltage above Vcc_core_Nom = 50mV.
From this it is implied that Vcc_coreMax = 1.65V.

Page 35 gives a nice little waveform diagram, tying this stuff together.

and lastly,

Page 36 (absolute Ratings) pegs absolute Vcc_coreMax + 0.5V

So, it looks like 2.15V does it. Kind of interesting that it only takes 50mV before you start pushing it into no no territory.

Hopefully I haven't misinterpreted it.

CK

 

[blakerwry]

i would assume that Vcc_coreMax is the maximum that a motherboard can run the CPU and be AMD aproved...

This, i would guess, is based on CPU life and possibly to ensure that the CPU conforms to the spec'd thermal output figures.

 

[Jake the Dog]

I managed to get my 1800+ to 2083MHz (2600+) @ 1.825V. temp under full appears to be 49C, even on a very hot day,. I can’t see it getting much over 60C, if at all. considering that a 1800+ is less than 30% the price of a 2600+, I'm quite pleased

 

[Stereodude]

There are people running over 2.1V with peltiers and watercooling. I'm running my T-bred B at default voltage only slightly faster than it's rated.

Stereodude

 

[CityK]

There are people running over 2.1V with peltiers and watercooling.

No doubt. I guess if you look at what AMD says about their absolute rating:

The AMD Athlon XP processor model 8 should not be subjected to conditions exceeding the absolute ratings, as such conditions can adversely affect long-term reliability or result in functional damage.

one might conclude that the key word is *can*. Ironically, there is nothing absolute implied by that word - either it will or it won't. Those using an extreme cooling apparatus may very well never experience a problem. It is also likely that these more exotic type individuals are well aware of the associated risk, and either except it, or care little about it in the first place.

CK

 

[blakerwry]

There are people running over 2.1V with peltiers and watercooling.

No doubt. I guess if you look at what AMD says about their absolute rating:

The AMD Athlon XP processor model 8 should not be subjected to conditions exceeding the absolute ratings, as such conditions can adversely affect long-term reliability or result in functional damage.

one might conclude that the key word is *can*. Ironically, there is nothing absolute implied by that word - either it will or it won't. Those using an extreme cooling apparatus may very well never experience a problem. It is also likely that these more exotic type individuals are well aware of the associated risk, and either except it, or care little about it in the first place.

CK

What about migration effects?... these CPUs are going to get fried one way or another if you run them like this.

 

[Stereodude]

Yeah, but if it last 9 months to a year I don't think they really care.

Stereodude

 

[blakerwry]

i guess I have enough computers that it's worth it to me if my CPU's last "past their prime", that way I can upgrade the lower models with my hand me downs.

 

[Jan Kivar]

i guess I have enough computers that it's worth it to me if my CPU's last "past their prime", that way I can upgrade the lower models with my hand me downs.

True. It's nice to see my 4-year-old BX rig running @ 1,3 GHz. Sadly I think that my current system isn't so futureproof... The performance is ok ATM...

Jan