Printable View
Interesting. My Winny's OC has IMPROVED over time. at stock volts / cooler used to be 2ghz, now 2.2ghz is fine and 2.4ghz is quite possible. Finishing my WC case before I really go at it though. Yes, I know I'm slow but end of Junior year / start of Senior in HS I have higher priorities [college]
if it's not the vdimm then could it be the motherboard killing your cpu slowly then?
I've experienced it with my Athlon-XP, over a few months time it just refused to do the same clocks..slowly degrading more and more.. I thought i had ruined the chip with vcore (i had it running 1.9v-2.0v for quite a while) And i just accepted it. Been running little over stock speeds for over a year now.. Though now i'm thinking it was either dust on the heatsink or thermal paste going bad as I just replaced my Zalman7000 cu with a TT BT and used AS5 (i couldn't resist testing it on my old gear, as i was still waiting for my other kit to be delivered) And well..it's doing it's old clocks happily again. :) I just never really wanted to reinstall the heatsink as i was scared of crushing the core ;) (i'm not scared anymore though hehe)
Perhaps it's just the vcore that is being used. My Winny ran 2.4GHz for 24/7 at 1.4v....so a little under stock, I just upped it to 2.5GHz and 1.45v. So I guess I haven't had to run a lot of vcore through it, mostly just stock or there under.
Have or had the problem. It started in yesterday evening.
My setup with 146 + XP90C with Delta 80mm 60%, DFI Ultra-D, Ballistx and OCZ 520W powersuply.
Worked in 3weeks with the same frequency and votage.
302 x 10 with 1.5V + 1.04%. Manage everything, so realy rock stable.
But at yesterday it whent down, and didnt whana do the same freq with the abow votages etc. But I just manage it do go back (I belive) after sanding the plate down, witch the aircooler holds it self down with (I'v removed the IHS).
So hopefully now with stable temps it will do the same as before.
I think it's happening to me. Used to need 1.552v for 2.7ghz, then bumped up to 1.584, now 1.584 fails prime in 10 hrs. I think I will go to 1.6v, let it sit, when/if it starts degrading, I'll lower clock speed.
I ran a 3000+ DTR claw 24/7 at near 1.9v on air 2.75ghz for months without a loss in clocks. I did have it at 2885mhz before but that was benching with water and the window wide open in the dead of winter.
I've been running 1.65v 2.8ghz (can run 1.675 2.85ghz but PWM gets HOT) on my 165 for a few weeks now so we'll see what happens with that.
Did you use your PSU in signature? Did you use just 1 board for all processors? Did you use any bh5 type memory?
I've had several cpus since the end of last year or so: s754 newcastle, winchester, venice, San Diego
Each of these at least for 3 months. I had absolutely no problems losing overclock, although they were heavily overclocked. For instance, SD was running 1.7V under XP-120 (by the way, great cooler it was :D) and vapo pe for half a year or so.
has it to do with temps are just Vcore, i am just wondering if my opteron 146 that does right now 2.75 @ 1.45 Vcore would degrade faster with a stock cooler than with a zalman 9500 (or what are you suggesting??)
I,m interested to see if this also happned to people who burned-in their cpu, maybe this loss of a OC is just a slow burn-in process.
There are such an enormous amount of factors involved in getting a good/excellent overclock.
Perhaps it was your motherboard voltage regulators, your power supply, the AS5 dried up, the memory was sucking the life out of the mem controller; point is who knows....
bottom line: something is different, most likely not the chip unless you juiced the heck out of it.
Look at all the possibilities, think about each one, then tackle the task at hand .... get your machine back into top shape.
I think that is where you should start because within the recurring slow deaths of your cpu's lies the answer.
Seconded :)Quote:
Originally Posted by phadeout
is it summer time where you live cronic maybe thats the problem hehe j/k
seriously tho all my cpus all seem to oc better with time and i mean my athlonxp systems and my newer a-64 systems. ive been running my sandy 3700 on watter @ 2700mhz for like 6 months i got it a year ago but the last 6 months it runs 2750 fine
v-core was 1.72 as thats pretty much the highest voltage that still gives a good enough boost to warrant the extra volts. personally i run super high v-core with my athlon xp too at 2.15 volts for over a year and it was oced to 2.6ghz on water now in my brothers computer at 2350 on air cooling 1.85v i think? anyways it never showed any sign of backing down.
i can imagine well enough the cpu will get tired out but i think this usually happens after they are obsolete anyways lol
and yeah the grease under the ihs on your athlon64s may be drying up
personally i remove the ihs imediatly after testing that the processor works lol
that ihs thing is retarted imo and only keeps blind grandmas from cracking the core or something
or the people who still want their warranty :stick:
Because you guys put way too much voltage in them and burn them out..
I have yet to wear out proc that I kept at stock voltages or less
Agreed. IMO, the IHS is the worst thing ever devised by mankind. But I understand why Intel and AMD had to do it. They had to account for the ever increasing number of DIY tards, rofl. :p:Quote:
Originally Posted by brandinb
I've been overclocking since the early P3 days, with too many CPUs passing through my little OC factory to count, and not one has degraded or died while under my service, (uh oh, now I've jinksed myself). Now that I think of it, I have four extensively OC'd CPUs that are still chugged along nicely for at least the past two years!
I do notice one thing that happens almost without fail though, OC'd CPU's seem to always gradually get hotter over time. That is until I get out a can of compressed air and clean out the dust. :hehe:
And I do mean all the dust, the stuff that collects at the base of the heatsink, the chipset fan, case fans, mobo components and sinks, cards, PSU, drives, that large ghost turd in the back corner. Everywhere in my cases gets a squirt of air.
Call me crazy, buy I like to keep things that are overclocked as cool as possible with high CFM fans. And I don't like filters because they quickly get choked off by dust quicker than anything anyway.
I'm in no way suggesting OC'd CPU degradation doesn't exist, but every time I see these overclocked hardware damage threads pop up they make me think I must be doing something right. :cool: :p: And I must say, compressed air has to be one of the best things ever devised by mankind. ;)
Had my san diego since may, running 2.86 ghz now, at first she'd only do 2.7.. She's only gone UP not down! My old 3000+ newcastle is still running 24/7 at 2.4 ghz, and it's been doing that since the 3000+ newcastle was released. Same with my old 3500+, which is doing 2.55 in my friends rig as we speak. Never had a problem with degration.
You guys just push far too much voltage through your chips.
Perhaps its not a good idea to run your cpu to the MAX. What I mean is, if you are not going to use what the CPU provides then don't push it.
Just because the Opteron can reach 3.0, doesnt mean you have to run it at 3.0
Save a few months/years and do say 2.6 or 2.7, I'm sure your games won't have any issues with it and its still fast.
If i am going to see downfall within a year, i rather not do 3gigz
And i will put money on it that most where with EXTREME mem voltage for utt or EXTREME htt speed 310plus on a dfi ultra d board or similur i love DFI but it killed my chip at stock speed slowly at first then it just shut off and never worked again...
Dragon
I have only had one A64 lose its overclock, which I think has more to do with the memory controller than anything. It was a Winchester - I was running 3.4 - 3.5 volts through the memory for about 2 months straight folding 24/7.
One day - it just lose 200 mhz. Bam.
I sold it - and have been a :banana::banana::banana::banana::banana: with vdimm ever since. :banana:
Lost two winnie 3500 and a venice 3200 only ran them at 2500MHz HT4 HTt250 1:1 no mad volts just a reasonable overclock both faded both where well cooled kept well below 50 both on MSI Neo2 plat both with DDR booster but only had it set to 2.9V
One off the reason I went for the opty - hopfully stronger mem controller!!!