Sign in with
Sign up | Sign in
Your question

OC gone wrong - post here if you've screwed your CPU

Last response: in Overclocking
Share
December 31, 2005 7:08:48 AM

I've got an AMD X2 4200, and recently worked up the courage to OC it from stock 2.2 Ghz to 2.4 Ghz. I'm stingy and paranoid (no doubt traits I got from my mom), so I'm afraid to take it to 2.6 Ghz. Therefore, I'm curious to know the *failure* rate of people's overclocking efforts. Perhaps we can get a laugh (or tear!) out of this thread too.

1 unrelated thing: I read through tutorials for how to OC, and most of them say to turn down the RAM clock speed and Hyper Transport speed before increasing the speed of the CPU. Can you bring these values back up after you're through with the CPU? Why or why not?

More about : wrong post screwed cpu

a b à CPUs
December 31, 2005 7:38:28 AM

Only a year ago i experimented on how easy it was to kill a cpu - tried an AMD K6-2 500, P1 233mmx and a Cyrix M2 PR333 (250/83) - basically i overclocked to the max and pushed the vcore to the limits to get it.

The Intel (233/66@2.8v) the max of 292 and vcore from a default 2.8v to the max at 3.5v and it didnt die and worked for days till i gave up on killing the thing (didnt even get too much hotter either).

I clocked the K6-2 (stock 500/100@2.2v) to 550 (bad sample, wouldnt reach 600 dammit) and pushed the vcore to 2.8v and it locked up and never posted again! - i didnt push it as far vcore wise cause it was a 250nm based chip vs 350nm P1's (Cyrix i aint sure about).

Finally the IBM/Cyrix M2 PR333 (stock 250/83@2.8v) very strangly overclocked to an amazing 350/100 at 3.5v (or was it more vcore?) and managed to stay there at a blazing hot temprature and didnt die either - the heat was so extreme the die and chip package got beyond 60c, and at that speed it still wasnt a match for any P2 and still got beaten by a P1 233mmx.

I admit i feel safer overclocking an Intel CPU and more sus bout overclocking an AMD but it was prolly just a bad sample and it was extremely overvolted (2.2v -> 2.8v) and that was way back then, now intel prescotts are more like cyrix's - slow and hot.
December 31, 2005 9:16:23 AM

Quote:
I read through tutorials for how to OC, and most of them say to turn down the RAM clock speed and Hyper Transport speed before increasing the speed of the CPU. Can you bring these values back up after you're through with the CPU? Why or why not?



The reason behind lowering the ram speed is because if you leave it at ddr400 and raise the fsb the ram is running faster and the ram may not be stable. The best way to find stable cpu OC is to lower the ram and eliminate the ram and find your best speed then do ram after. Also the htt is best to keep at 1000 or lower. HTT is based off the fsb so 300fsb you would set the htt at x3 so it would = 900 htt which WILL be stable htt (900 to 1000 is no performance difference so dont worry about it) going over 1000htt could make your comp unstable.


Edit: by the way you cant screw your cpu if u try to go to 2.6 from 2.4 or 2.8 from 2.6 if its not stable it will not boot, or will get to windows and crash or it will just not be prime stable, it wont fry your cpu. The only way you can fry your cpu is if you overvolt it.
Related resources
December 31, 2005 10:42:46 AM

Everywhere I have read states that all venice/venus core CPUs and thier variants OC pretty good. All of them also state you may not achieve the same results they did. Anandtech (or maybe it was somewhere else) got the X2 3800 to 2.7 but it was not all the way stable. From 2 ghz to 2.2 ghz there was nothing to note. They used all stock settings and there was not much difference in temps. At 2.4 ghz there was a difference in temps and it required a voltage boost. I think 2.6 is where you would want to stop. All major components needed votage boosts and there was like a 15 degrees temp boost. I noticed it cuz 2.6 was where I wanted to go. The most interesting thing in the whole damned article was a statement that you can overclock and stress your system and it may seem to work fine. Then after 6 months, you could be screwed. I overclocked an XP-m from 1.8 ghz to 2.35 (most people said they had theirs up to 2.5 ghz easily and regularly). My system became more and more unstable by the day. Finally after few weeks I was down to 2.1 ghz before i sold the whole thing. I mean, its still a 300 mhz boost but why cant I get 2.5 ghz like everyone else? Probobly cuz alot of people are one uppers and lie a little. The way i see it, if the XP platform was capable of 2.5 ghz why did the XP line stop at 2.2 ghz (and even that chip was rare and ran VERY hot) I offer for you to be happy with 2.4 ghz and push it to 2.6 when your system starts to get old. If it breaks, the oh well. it was time to buy again anyways right?
December 31, 2005 4:09:32 PM

I overclock everything I've got and the only CPU I ever killed was my old Athlon 1400 T-bird :x (sob!) Those chips were real burners to begin with and I was havign trouble keeping it cool at 1550 MHz.

So I took off my Dual Fan Global Win HSF and lapped it bit. When I remounted it I must have had it up on one side, cause it put a hardly visible crack in one corder of the chip. It never booted again but it made a nice conversation piece on my desk!

As they say, "I learned about computers from that!"
December 31, 2005 4:22:45 PM

I have an X2-4400 (factory clocked at 2.2 GHZ... the same as your 4200) and I have clocked it up to 2.6 GHZ with an upgraded air cooler and a slight voltage bump... stability has been excellent over the past 6 months. Several attempts (newbie mistakes) to get to 2.6 GHZ failed miserably, but a little reading on the internet straightened me out. I personally believe if you don't push too much voltage through one and you don't let the temps get out of control, you're free to experiment with minimal risk involved.

My understanding is that modern CPUs have some sort of built-in thermal protection... nothing like the old Athlons (pre-XP) that would go up in smoke the moment the heatsink was removed.
January 1, 2006 4:52:47 PM

Don't even know how i killed a crappy msi mobo with via 478 chipset... oc my proc 3.0 prescott to 3.3 and it never posted again... (though the proc was still good...)
a b à CPUs
January 1, 2006 10:34:29 PM

BAH MSI and VIA - theres the issues - both companies are crap - iv owned stacks of boards with VIA chipsets and MSI branded boards and THEY ALL DIED, had a MSI burner that burned a cd and had to restart windows to burn another (and it died a month later), a BX MSI board and it had a dead IDE channel and died aswell, countless other MSI boards and VIA - i havnt seen a VIA board thats lived till the next upgrade!!!!!!! And then theres my MSI GeForce4 Ti4200 8x - it overheated with that piece of crap "copper" cooler (the copper was glued to an aluminium plate) - replaced the cooler with a reference thermaltake cooler 1/4 the size and its now cool.

MSI and VIA - dont touch em, and together, thats even worse.
January 4, 2006 4:59:17 PM

Had a mobile 2500+ on an Abit NF7 v.2
Ended up running 2.4v through the poor thing, then @ 2340mhz...had the hardest lockup I've ever seen. --Short and sweet -- FRIED.
The motherboard wasnt in great shape either...but I really dont want to go their.


Remember my friends....

"With great power, comes great responsibility"
January 4, 2006 7:29:58 PM

I had a MSI mobo and it died suddenly not more than 9 months. I got 2 VIA mobo and it died also not more than 1 year and 4 months. I got a MSI DVD-Rom and it died within half a year. I got 3 Maxtor Diamond Max Plus 8 and they all died without lasting for more than 1 and a half year.I got an Intel celeron 2.0 Ghz my friend came to my house I left it in the mobo with the fan but the CPU sticked on the fan he pulled it out had a look and without lifting the lever he just stick it back in and put the fan back also and then I lost my celeron =.=
January 5, 2006 9:58:23 PM

erm never lost a cpu through overclocking before, but i have lost a whole machine by accidently plugging a cable in the wrong way, long story short it went up in flames.

i learnt to pay more attention to the manual...

lucky it was an old pentium 2 with a 2GB HDD and some onboard crappy stuff i was gonna use it for dos games lol.
January 6, 2006 8:00:19 PM

A long long time ago now I killed one of the original Pentium (60MHz) chips getting it out of the socket, this was before zif sockets were in use and you had to use chip pulling tools (or a flat head screwdriver)...

Needless to say I could not find a chip pulling tool that day :(  I cracked the whole corner off, along with about 10 pins still attached..

Ouch!

Overclocking wise my 486 33mhz ran for years at 50mhz... Don't get overclocking like that these days :p 
January 6, 2006 8:39:56 PM

i OC a P1 233MMX to 266MHz i believe(raised the voltage from 2.8-3.2 and upped multiplier). worked great for a couple of days. then it decided to crash, and permanently <i>underclock</i> itself. its now stuck at 175MHz, and nothing will change it.
i must say, it is an old piece of machinary, and outdated and useless, but still disappointing nonetheless.
January 6, 2006 10:17:27 PM

1. i've overclocked my P1 100 phz, needed to overclock with jumpers on the mobo itselve (damn bios...) 'ive got it to 120mhz, but it always turned off, then i've ripped a cooler from a torch that i've got some years ago from teh "santa" and putted it above the heat sink with wire strips, and now it keeps on!! still a crappy thing, but usual for the old good games :p 

2. thes P4 2.4ghz northwood overclocked till 3.0 ghz, i can get it stable at 3120mhz, but then i would have some heat problems with my crappy stock cooling... i'd just upped the fsb from 133 mhz till 167hmz and gave the Vcore +5%Volt.
January 25, 2006 7:02:36 PM

I'm using a VIA PM800 North bridge and VT8237 South Bridge. And I'm overclocking to 3.48Ghz from 3.0Ghz with absolutely no problem. I've never had a problem with this system.

Yes its a Prescott, and it runs about 23 to 25C idle. 35C at max load. Stock cooling. Ambient 70F. :) 
a b à CPUs
January 25, 2006 10:29:11 PM

Quote:
1. i've overclocked my P1 100 phz, needed to overclock with jumpers on the mobo itselve (damn bios...) 'ive got it to 120mhz, but it always turned off, then i've ripped a cooler from a torch that i've got some years ago from teh "santa" and putted it above the heat sink with wire strips, and now it keeps on!! still a crappy thing, but usual for the old good games :p 

2. thes P4 2.4ghz northwood overclocked till 3.0 ghz, i can get it stable at 3120mhz, but then i would have some heat problems with my crappy stock cooling... i'd just upped the fsb from 133 mhz till 167hmz and gave the Vcore +5%Volt.


My P4c 2600 hits the wall at 3.120ghz aswell, even at 1.8v it wont get to the magical 3250/1000.
a b à CPUs
January 25, 2006 10:32:43 PM

Quote:
I'm using a VIA PM800 North bridge and VT8237 South Bridge. And I'm overclocking to 3.48Ghz from 3.0Ghz with absolutely no problem. I've never had a problem with this system.

Yes its a Prescott, and it runs about 23 to 25C idle. 35C at max load. Stock cooling. Ambient 70F. :) 


Thats the coldest prescott iv ever heard of on stock cooling.

VIA chipsets - ill never trust em again like many, if that was one of my systems that would be dead in a few months (the board)
January 25, 2006 11:19:52 PM

its stock cooling... voltage is 1.300 to 1.324 (CPU-Z)
January 26, 2006 7:47:59 AM

Quote:
Overclocking wise my 486 33mhz ran for years at 50mhz... Don't get overclocking like that these days :p 
50%? Sure you do... The Northwood A 1.6Ghz was well known as virtually guaranteed to do 2.4Ghz, and my XP1700+ (1.466Ghz) would (will) do 2.3Ghz.

A lot of these insane Opterons seem to be getting to near 50% oc too.
January 26, 2006 7:55:43 AM

I'm sorry, but there's simply no way those temps are from a scotty with stock cooling.

I'll give you the benefit of the doubt and assume your board sensors/monitoring software are spouting crap...

Try to get hold of a temp. probe or something to verify that. No review I've ever seen has managed an idle temp of 35C with the stock cooler... let alone an overclocked under load one....

A good water system might manage, but even then I'm not too sure.
January 26, 2006 7:08:18 PM

read in the thread I created... it might explain things a little better :p 
January 29, 2006 9:40:30 PM

my venice gets 44% oc
February 5, 2006 6:02:43 PM

I overclocked a K6-2 amd 333 to 400. I had one of those boards you had to change jumpers on board. I messed up and read the Manual wrong somehow got it to boot up at 600megs but would never boot up the OS. After a Few years later like 2 months ago i went back to my old system was going to try to get it to actually run stable at 600 with new coolers and such. Still couldnt get it to boot up but something went wrong and it will not even boot up. I think its that whole Y2K thing. It gives me a Cmos error.

I had one of those Cyrix chips 266 was a pretty good chip for its time i had it OCed to 300 but the board i had didnt have Dimm slots still had simms.

I also had a Duron 600Mhz OCed to 650 it ran stable for a long time but then i guess it got week it would keep rebooting so i put it back down to 600 it would never run stable at 600 either it keep rebooting every 4 or 5 hours I would look at the temp in Bios but it always said 80 i thought it was set on F but it was on C i was like hmmm that could be why so i took off the HSF and it did not have any past it just had that pink stuff on the bottom of it that comes iwth it but where the chip was on the HSF there was nothing.
!