Welcome Guest ( Log In | Register )

Bump Topic Topic Closed RSS Feed

Outline · [ Standard ] · Linear+

 disadvantages of overclocking, disadvantages of overclocking

views
     
oe_kintaro
post Mar 4 2009, 08:14 AM

Look at all my stars!!
*******
Senior Member
2,222 posts

Joined: Jan 2003
From: Penang


Nowadays high performance chips are fabricated with copper used as the metalization. Copper atoms are smaller than Aluminium which is the traditional material used to fabricate metalization in ICs, and furthermore lack the capability to form an impermeable oxide layer like that in Al. That makes copper more prone to electro-migration. The increased voltages will accelerate copper electro-migration and result in early failure.

Forgive me if I seem to be throwing a fox into the chicken coop, but one AMD design engineer I knew preferred to *underclock* his PC and said overclocking as well as fancy glow-in-the-dark UV doodads is for stupid people. I tend to agree. Overclocking induces electromigration, while UV light is know to accelerate the degradation of polymers.

It's just like using NO boost on your car 24/7: you feel damn syok sendiri but will cry when the repair bills come.

However, having said all that, what I can say is that there nothing ethically wrong with overclocking as long as you know what you are getting yourself into. Just don't go whining to the manufacturer asking for a RMA when you overclock parts that are not suitable to OC and cause them to break. Overclock knowing full well that you are voiding a manufacturer's warranty

This post has been edited by oe_kintaro: Mar 4 2009, 08:16 AM
oe_kintaro
post Mar 4 2009, 09:20 AM

Look at all my stars!!
*******
Senior Member
2,222 posts

Joined: Jan 2003
From: Penang


QUOTE(edmunz @ Mar 4 2009, 08:25 AM)
im interest to know more bout UV light accelerate the degradation of polymers.... whr can we find polymer in PC / wat component icon_question.gif
underclock da pc wont cause pc to b unstable ?????
*
basically nearly EVERYTHING inside a PC that isn't metal or silicon is some sort of polymer-based material. (aka plastics). We are talking about the PCBs, mold compounds, wire insulation etc. Though maybe they have varying resistance/stabilization to UV light, it is just a matter of time, and very often your system is only as good as your weakest link. (Actually, all plastics degrade given enough time and temperature, and in IMHO adding UV to the mix just helps it move along a little faster)

Not forgetting too, certain wavelengths of UV were traditionally used to ERASE nonvolatile memory ICs such as EPROMS. Though now we have migrated to flash and EEPROMS for our PCs' BIOSes and it is exceedingly rare you will find any piece of IC in your modern PC board which intentionally exposes the silicon die via a transparent plastic window (e.g. in the old CERDIP type packages), you should be aware that UV rays (as well as Xrays) are generally not your PC's best friends.

HOWEVER, as far as I know, no one has ever done a serious scientific study of how pimping up one's PC with UV lights will affect the lifetime of the components tongue.gif

edit: some general info on UV degradation of polymers:

http://www.gcrio.org/UNEP1998/UNEP98p62.html

This post has been edited by oe_kintaro: Mar 4 2009, 09:34 AM
oe_kintaro
post Mar 4 2009, 10:24 AM

Look at all my stars!!
*******
Senior Member
2,222 posts

Joined: Jan 2003
From: Penang


QUOTE(piumiu @ Mar 4 2009, 09:38 AM)
I have seen many people overclock for 4 years and more and still their PC is alive and kicking. Most of it are the famous case are the AMD Barton 2500.

With the new overclock technology that been provided by manufacturer as Black Edittion, Oc Mainboard and GPU, I don't see it as stupid.

People like us with little money tend to achieve best computing experience with our limited budget and hardware.

Your friend don't have the interest, requirement or limitation of budget that kind to understand and yet condemn millions of people stupidity in overclocking.

So are we in overclockers united lame and stupid?
We are not, we are smarter than AMD engineer.

Please provide what sort of engineer he is as I also know few fabrication specialist from Intel US.
*
You don't have to get so defensive. I'm sure he meant it in half-jest but his reasoning is not unsound. Like I said, I'm aware that I'm throwing the proverbial fox into the chicken coop, but there's nothing like healthy debate to enhance everyone's understanding of the issues involved. I'm not condemning OCing or pimping one's rig per se, but just that would like to do some sharing so that everyone knows what they are doing and the potential risks involved.

As long as you are happy with your results it is fine.

The example you gave me is like saying that you know someone's grandpa who smoked like a chimney and lived until he was 80: it does not mean that smoking will not do harm, certainly not in the face of overwhelming scientific evidence.

All manufacturers incorporate a certain amount of safety factors, and one chip binned at 2.5GHz is often exactly the same as another one binned at 3.0GHz, except for tolerance related process variance which results in one chip being faster than the other. Secondary manufacturers often take the cream of the crop and do their own testing so it is generally safe. Just remember that it is the 2nd manufacturer that guarantees you the product not the primary one.
oe_kintaro
post Mar 4 2009, 10:46 AM

Look at all my stars!!
*******
Senior Member
2,222 posts

Joined: Jan 2003
From: Penang


QUOTE(sakuraboo @ Mar 4 2009, 10:29 AM)
so if i dun install the funnee lights so no uv so OC ok?!
*
actually they are two completely different but related potential issues because of the demographic (OCers tend to be the ones more likely to install funny lights) tongue.gif sweat.gif

most polymers inside a PC may already have various stabilizers which may aid resistance to UV but I don't know of any serious scientific study at this point which investigates the effects of constant UV exposure within an operating PC enclosure. My concern about this is mainly relating to the potential for structural integrity failure of load bearing components.

OCing is associated with a different set of problems, such as electromigration (at the physical level) and system instability and unreliability (at the application level). If you are just running a PC at home, chances are most people won't mind any risks involved, but if you are running critical enterprise stuff or operating a space shuttle, chances are you wouldn't wanna risk your life plugging in an overclocked PC to do the heavy lifting. To put it simply, my stand that everyone can choose their own poison as long as you can accept it smile.gif

This post has been edited by oe_kintaro: Mar 4 2009, 10:48 AM
oe_kintaro
post Mar 4 2009, 11:21 AM

Look at all my stars!!
*******
Senior Member
2,222 posts

Joined: Jan 2003
From: Penang


QUOTE(OlgaC4 @ Mar 4 2009, 11:15 AM)
If your honda cup C70 can run 70Km/hr you modify it to run 90Km/hr the seal and casket in the engine will fail very soon because of the pressure. Furthermore the speed will not stable, some times 80km sometimes 85. I don't want when playing a games my computer sometimes go fast sometimes go very slow. Unstable.
*
A very good metaphor! I hope it makes it clear for everyone here.


ps: forgive me for being OT here, but are u a intellier or ex-intellier? the alphabet soup in your nick brings back memories for me (I interned in Intel around the time they were working on the Deschutes)
hehehe smile.gif
oe_kintaro
post Mar 4 2009, 11:55 AM

Look at all my stars!!
*******
Senior Member
2,222 posts

Joined: Jan 2003
From: Penang


QUOTE(lichyetan @ Mar 4 2009, 11:35 AM)
lol... it actually depends la.. for personal computer, OC stress test tweak for fun lo... benchmark to test how far the PC can go etc... maximize the potential CPU power, i believe Intel & AMD's R&D department also got OC inside their lab for benchmark or testing purpose whenever a new product come out... just like those DIY'ers like to mod their car, same goes to OC'er, tweak have fun. experience the speed etc...

for server usage, who wan OC ? stability is the highest priority.... for space shuttle server etc they might be using multi processor already, 10 cores 12 cores mayb... full buffered dram etc... supercomputers mayb... gotto ask NASA...

ur fren AMD might be right... but mention ppl who OC is stupid is not right... if OC'er is stupid, all DELL XPS engineers are stupid, Intel's and AMD R&D department (partly) are stupid... ur fren are the best engineer in AMD..  whistling.gif (just a joke, since i dont like calling other ppl stupid's ppl).

btw, in my opinion, by the time ur proc die(10 years? 6 years?), ur PC might be keep in the shelf d... average life cycle for a personal computer is around 2 years (thanks to microsoft and those hardware manufacturer who keeps on introducing new products)  tongue.gif

server diff business la... dont ever compare industrial use with personal use...

UV bling bling diff business... i do OC, i dun like UV lights or others... my pc is just plain black with non-transparent sidepanel...
*
Like I said, he meant in jest, but it was a memorable statement designed to provoke an interesting debate. No need to take it personally or have hard feelings over it sweat.gif

Btw oftentimes the R&D people push their parts to the limits so that the marketing people can safely quote you specs in the datasheets that they know their process can meet with one eye closed tongue.gif That's the safety factor involved in the design and manufacturing process. It helps minimize any potential law suit or compensation as a result of parts not meeting specs.

Perhaps that is also why those 2 rovers on Mars are still up and running for so long despite their initial "designed" lifespan of 6 months.

This post has been edited by oe_kintaro: Mar 4 2009, 11:59 AM

Topic ClosedOptions
 

Change to:
| Lo-Fi Version
0.0208sec    0.49    6 queries    GZIP Disabled
Time is now: 9th December 2025 - 11:12 AM