Advertisement
If you have a new account but are having problems posting or verifying your account, please email us on hello@boards.ie for help. Thanks :)
Hello all! Please ensure that you are posting a new thread or question in the appropriate forum. The Feedback forum is overwhelmed with questions that are having to be moved elsewhere. If you need help to verify your account contact hello@boards.ie

CPU voltage query

Options
  • 16-07-2004 4:09pm
    #1
    Registered Users Posts: 8,048 ✭✭✭


    I know AMD dont recommend you going over 1.7-1.75v but im wondering are there any side affects apart from extra heat? Will it severely shorten the life of your chip or cause some other sort of damage?

    I got a vapo so heat aint an issue, im at 1.75v atm and my mobo would let me go up to 1.8 as max. Just wonderin if i go up to 1.8 would it caus emore damage than at 1.75 or 1.7. Load temps atm are about -15 with idle about -23.

    (all other components should be able to handle it ;) )


Comments

  • Registered Users Posts: 2,754 ✭✭✭Col_Loki


    Generally with XP chips its a mix of voltage and Heat that effects CPU life. Even if you have low temps the voltage will effect the chip but not to the same degree if you have it well cooled.

    The P4 northwoods . prescotts arent that good with voltage and many say an upper limit of 1.7v as you can get sudden deaths of processors, but this generally happens on 1.8v+ from what ive seen.

    On the AMD64 im not sure, its a pretty new chip and its a hard one to call when it comes to voltage TBH. Alot of people treat them the same as the P4's but give them a little more voltage, ie 1.7v wouldnt be strange. With your cooling you definetly take the heat out of the equation but im not sure just how well there protected from high voltage.........

    [edit] PS are you sure that its your CPU limiting your overclock? Ie can you achieve the same speeds on a lower voltage? Have you tried 3x HT?


  • Moderators, Recreation & Hobbies Moderators, Science, Health & Environment Moderators, Technology & Internet Moderators Posts: 91,566 Mod ✭✭✭✭Capt'n Midnight


    Originally posted by Col_Loki
    Generally with XP chips its a mix of voltage and Heat that effects CPU life. Even if you have low temps the voltage will effect the chip but not to the same degree if you have it well cooled.

    With light bulbs (comparing like with totally unalike) a 5% over voltage will half the life and 5% under voltage will double the life. But then doubling the life of a light bulb means less light and to get the same amount of light with a bigger bulb will use more electricty and so cost more resulting in no savings.

    I reckon the same is probably true of CPU's - how many hours a day do you use it - how many years do you want it to last. It's a case of live faster die younger - but if you are upgrading again in the future that might not be such a big issue.

    Anyone know if you can change the voltage as well as cpu speed on the fly - sorta like ye olde turbo switch - so when surfing you are ticking over and then kick in the after(burner) for games ?


  • Registered Users Posts: 8,048 ✭✭✭BKtje


    The amd 64's have a feature called cool n quiet which will lower the voltage and cpu speed when not in use.

    There are programs that that allow you to change it from inside windows.

    Nope im not sure about CPU limiting it. Ill try HTT x3 now actually.

    All i know is that at
    1.75v it wont load windows at 2.5ghz and 2.45v is unstable.
    1.8v it will partially load windows and 2.45 is fairly stable.

    Ive tried 4 bios's now. 3 of em wont allow me to hit even 240fsb. it just won't post. The original 1.1 bios dos it nps.

    I was hopinh to upgrade again ina years time to whatever the best 757 processor is (by that time 757 will probably have the best that it will get).

    Im still undecided wether to leave at 1.8 or 1.75. (i also noticed that the nforce 3 isnt very good with voltages. ie i had it at 1.75 and cpu-z says thgat the voltage is 1.78 -1.75-1.76. ie fluctiates a lot. Apparently thats normal for this board/nforce.

    Infact just looking now, even tho i stuck it to 1.8v its still the same voltage as 1.75. Weird for sure :p

    Gonna try with HTT x3 now.


  • Registered Users Posts: 8,048 ✭✭✭BKtje


    Right it just loaded windows no problem at htt x3.

    Ill see how far i can push this now just outta curiosity.,


  • Registered Users Posts: 8,048 ✭✭✭BKtje


    Right it went as far as 2.55 with httx3. At 2.6 it refused to post at all (similar to the 2.4 problem of other bios's).

    I dunno now if the cpu is at its limit or not. Guess ill have to wait and see if they ever make a decent BIOS. :(


  • Advertisement
  • Registered Users Posts: 2,754 ✭✭✭Col_Loki


    Ok well thats a nice improvement. Can you try at 2xHTT? That seems to be alot of peoples problems.

    From the looks of things your CPU wasent the limiting factor which is good. There are a few tricks etc with the board, what are you using to overclock with? BIOS or Windows?


  • Registered Users Posts: 8,048 ✭✭✭BKtje


    Can you try at 2xHTT?
    I can give it a go tho im loosing a fair amount of performance aint i by going that low which wont be returned by a higher core speed.
    There are a few tricks etc with the board, what are you using to overclock with? BIOS or Windows?

    Im using Bios, im not actually a fan of the windows way as it tends to crash half the time and even if it is stable it just reverts later (ie pain in the ass, when i overclock i see if it will load windows and if it does then ill stress test it for a bit).

    Ill give HTT x2 a go. Can't change the htt setting from windows.


  • Registered Users Posts: 8,048 ✭✭✭BKtje


    No change, still wont post at 260fsb even with htt x2. (i m sure this is a bios issue rather than hardware as the same happens at a later bios but at 240fsb instead of 260).


  • Registered Users Posts: 2,754 ✭✭✭Col_Loki


    You should see almost no difference performance wise between 3xHTT and 4xHTT especially since its not Dual DDR board.

    I would say at this point 2.55ghz is prob the max with BIOS overclocking. Its a real pitty your having problems with ClockGen as it seems to give great results when compared to a BIOS overclock...... and also aparently allows for tighter memory timings at higher speeds.

    2.55ghz Still isint bad, very fast.

    [edit] Try a benchmark between 3xHTT and 4xHTT and see if im correct or not on that one..... Remember HTT is different from FSB.


  • Registered Users Posts: 8,048 ✭✭✭BKtje


    okie ill give 3dmarks another whirl.


  • Advertisement
  • Registered Users Posts: 8,048 ✭✭✭BKtje


    You were spot on loki. Got about 1000 more 2001 3dmarks at 2.55 httx3 than 2.4 httx4.

    Cheers :)


  • Registered Users Posts: 2,754 ✭✭✭Col_Loki


    lol, thats good to hear!!


  • Registered Users Posts: 423 ✭✭Digi_Tilmitt


    Athlon XP's are pratically invincible, I couldn't kill one with volts if i wanted to. As long as you keep the temps down anything goes. As for everything else....


  • Registered Users Posts: 9,216 ✭✭✭MrVestek


    This is kinda along the same lines. I bought an athlon xp 2500+ today only to find that my mobo dont support the barton core so the most i can get out of it is 1700+ (about 1.4ghz) is there any way to push it to it's proper speed or even beyond? When i push it to a host frequency of over 245 it wont boot. Anyone?


  • Registered Users Posts: 447 ✭✭cerebus


    Just in case anyone is interested, high voltages have all sorts of reliability/lifetime impacts on transistors especially at feature sizes under 0.13u.

    Two effects you might want to google are hot-carrier injection (HCI) and negative-bias temperature instability (NBTI). N-type devices suffer more from HCI, while NBTI is more of a problem for P-type transistors.

    These phenomena slow down transistor operation through various mechanisms - pretty complex physics, but you can think of them as increasing threshold voltages (Vt) and reducing saturation currents (Idsat).

    These effects tend to be exponential with voltage, so a small increase in your core voltage can have a large impact on device lifetime (I can probably dig up an example showing change in device lifetime for various voltage deltas if you are interested). Capt'n Midnight's lightbulb example was pretty close to what happens with silicon as well.

    Most designers will build some margin into their circuits, but I guess there's always the risk that you might have a uP part that just about made it into a particular bin (and so might be particularly sensitive - i.e. a small degradation might cause it to drop out of the bin).

    Having said all this, it is very unlikely that you'll see much impact from the tweaks you are talking about - just thought you might be interested more than anything else...

    [size=-2]<edit>Whoops! I said exponential with temperature when I should really have said exponential with voltage</edit>[/size]


  • Moderators, Recreation & Hobbies Moderators, Science, Health & Environment Moderators, Technology & Internet Moderators Posts: 91,566 Mod ✭✭✭✭Capt'n Midnight


    According to FOCUS August edition page 42 (why are there more pages in the Italian / German editions ?) the current predicted limit is 5 atoms with an ETA of 2012.

    ( Since it's FOUCS they say that DNA will be smaller - it won't - though biomolecues are faster. )


  • Registered Users Posts: 8,048 ✭✭✭BKtje


    hehe i have no idea wtf you talkign about :D

    Ive had to reduce the speed to 2480 as i was getting errors very early on in prime95 at anything above that. Think it ran for 9hrs last night without any errors at 2480 tho i was locked whe i got back so dont remember to well.

    Ill find out soon enough :D

    Just below 500mhz overclock makes me happy tho :)


Advertisement