• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Overclocking potential nerfed after failure

S

Silverhawk

n00b
Joined
Nov 14, 2005
Messages
3
Hi there, i've had a weird problem happen to me when overclocking hope you people can help me out.

I was overclocking my A64 3000+(Venice), i got it up to 275*9 without much difficulty. Through the test i ran my HT Multiplier at 3 and ram was set to DDR333. So with the 275*9 stable i pushed it to 280*9 to see if that would be stable. After about an hour of Prime95, windows suddenly gave an error and i had to reboot.

So I lowered it back down to 275*9. It wouldn't boot. Trying the previous stable overclocks and it didn't boot either. Now i can't get an overclock higher than 220*9. What could be the problem?

I have done the following, with no luck
  • Repositioned RAM sticks
  • Use only 1 RAM stick, and test each slot
  • Increase RAM voltage up to 2.8 (i didn't dare try higher)
  • Increase CPU Voltage up to 1.7
  • Ran memtest86 (i suspected the RAM, but memtest didnt turn out any errors after 2 hours)
  • Set RAM to CL3 instead of 2.5
  • Reflashed BIOS to older version and test
  • Flashed BIOS back to latest version to test again

The overclocking potential being nerfed doesn't really bother me that much, though it is rather dissapointing. What worries me more is that the overclocking might have damaged something and i'm not sure what.

What i'm using:
Athlon64 3000+ (venice)
Abit A8N Fatal1ty SLI
2x1GB Corsair Value Select DDR400
 
OC is a dangerous hobby and occassionally you smoke one. Sounds like you need a new CPU. What were you using to cool your CPU before you toasted it? In the future you might want to try multiples of one (1) to sneak up on a OC (275 > 276 > 277 > etc).
 
Top Nurse: The CPU isn't toasted :p It is just unable to clock as high as it did before. I was and still am using a Thermaltake Bigwater SE cooling kit. Some may laugh at it, but its doing a pretty good job.

Current temps: (222*9)
Room temp : 31C
idle Temp : 32C
Load Temp:40C

Previous temp: (275*9)
Room temp: 31C
Idle temp: 32C
Load Temp: 50C (yea a lil high, but still within acceptable range)

If the CPU is slightly defective, then i guess i'll just stick to it for a while. Its just a placeholder while i wait for the dual-cores to drop in price :)

Ice Czar: just read that really long post @_@. So you're saying my RAM/sumthing got messed up by ESD?

Thx for the replies
 
electromigration
first degrades then kills IC chips
(CPUs\GPUs\RAM ect)

Black's Equation roughly translated states that Electromigration mechanisms are accelerated by current density as well as temperature, thats a perfect description of what overclocking is about, actually determining what state a given chip is at defys most manufacturers mass production testing, meaning the chip ships if it reaches a single threshold they test at, whereas you are trying to find the upper limit of a given chip and test it repeatedly, only problem is that if you push too far with either the voltage or heat your in very real danger of degrading the chip.

which is why overclocking is as much an art form as a science
you have to know how far you can get for a stable overclock without degrading the chip
and any given chip is in an unknown state of pristine circuity untill you actually try
it could be a dog that just barely passed QC at Intel\AMD or it could be a golden chip destined for fame

http://www.odyseus.nildram.co.uk/Systems_And_Devices_Files/Component Reliability Tutorial.pdf
 
my 3000venice runs @ 35C tops 31C idle at 2.7 [9x295] on the stock hsf no liquid cooling...yet [1/2" in the works] wait, bigwater = 1/4" hee hee hee big 1/4"
 
electromigratition is constantly occuring
its merely the rate which changes and its exponential as either temperature or voltage increases
what a CPU is "rated" for in temperature is really a tradeoff between a reasonable lifespan and a reasonably maintained operating environment, keeping computer rooms cold isnt very reasonable for an enterprise or even an end user

as a general rule of thumb you have the Arrhenius equation

roughly translating to this rule of thumb
Each 10°C (18°F) temperature rise reduces component life by 50%
Conversely, each 10°C (18°F) temperature reduction increases component life by 100%.

its modified by other factors into Black's equation

generaly speaking staying under 50C at full load is a good mark to shoot for,
but the lower the better some try to stay under 40C (I for one)
others opt for whatever temperature they can get a stable overclock,
but the ones able to stay far lower will last far longer everything else being equal

there are other caveates, first do you really need the full rated lifespan before you upgrade?
do you even need half that lifespan before the upgrade?

set off against too high a temperature and or voltage could make a latent defect suddenly get much worse, as in failing now
or as what might be the case here, you can't reach a previous overclock do to degradation

largely it depends on where your starting from, and you can't really determine that without trying to overclock
so either way your playing the odds ;)
 
You must log in or register to reply here.
Back
Top