PDA

View Full Version : Why would I Over Clock my Sempron ?



MooPi
November 9th, 2010, 03:31 AM
I just OC'd my Sempron dual core. I know, why do you OC such a wimp of a CPU ? I have to much time on my hands probably. Well It is overclocked 23% to 2.46Ghz and it hasn't even broke a sweat yet. I don't know how high I'll take it but it's always fun to tinker.

Give me your OC'd cpu info. How many Linux geeks OC ?

Schrute Farms
November 9th, 2010, 03:53 AM
Why not overclock it? Because it is so "wimpy", it is a perfect chip to try to squeeze every bit out of it you can.
I say see what it'll take.

amauk
November 9th, 2010, 03:56 AM
Never understood the whole overclocking thing.
Running a component above it's tested tolerance does wonders for it's life-span...

MooPi
November 9th, 2010, 03:58 AM
Never understood the whole overclocking thing.
Running a component above it's tested tolerance does wonders for it's life-span...

Your no fun.......:)

simpleblue
November 9th, 2010, 04:18 AM
Never understood the whole overclocking thing.
Running a component above it's tested tolerance does wonders for it's life-span...

Life-span vs Voltage

Here's a graph for those interested:

http://images.anandtech.com/reviews/cpu/intel/2008/e8500-overclocking/Degradation.png


Keep in mind that this is only a few percent increase in voltage. For the small, maybe 10% increase in performance, you could easily decrease your CPU life by many years or worse. As well, the hotter electronics get, the more they conduct, and the more they conduct, the hotter they get. It feeds into itself.

If you do decide to do this I would recommend using a really good zinc-oxide cream to help dissipate the extra heat. And you'll need to apply the cream properly because if you leave any air holes, the chip is sure to blow within minutes or hours. As well a good heatsink is a must!! A powerful fan is needed, but is NOT enough by itself!

Good luck!

Windows Nerd
November 9th, 2010, 04:52 AM
Never understood the whole overclocking thing.
Running a component above it's tested tolerance does wonders for it's life-span...

If you overclock the processor without increasing the voltage, then you technically are not notching down it's lifespan at all.

The heat increase and reduced lifespan associated with overclocking comes with increasing voltage, which is necessary to do when trying to do the higher overclocks and push the limits of the hardware.

That being said, the decrease in lifespan really depends on the chip, as older Pentium III's and IV's are heat-generating monsters, and I myself have several old machines that still run good as ever with those processors in them.

With the right forethought as to cooling, the lifespan is not nearly reduced as much. Often why some chips have such good overclocking headroom than others (even from the same make and model of chip)is because companies who sell the same processor clockedat 2.4Ghz, 2.8 GHz, and 3.4 GHz often sell much more of the one clocked at 2.8. So sometimes if they need to keep up with the demand, some companies take the 3.4 chips, and underclock them to 2.8 and then sell them, when in reality the chip is made to run at 3.4 for the same lifespan.

Scott

Brent0
November 9th, 2010, 05:33 AM
I overclocked my Athlon II x3 to 3.4Ghz and thats how I have it at all times. Still have low temps, too. (19-35C) I have gotten it up to 4Ghz but anything above 3.4Ghz is unstable. I will try upping the voltage a bit to see if that helps.

I also overclocked my HD5770 a few hundred Megahertz and the boost in speed is very noticeable.

Khakilang
November 9th, 2010, 05:34 AM
It's like computer on steroid. How long will it last? Yes, it is fun to tinker with it. Unless you have the budget to change your mobo every 3 years. Than by all means have all the fun.

Oxwivi
November 9th, 2010, 07:32 AM
Got a P4 2.4GHz, how do I overclock?

Spice Weasel
November 9th, 2010, 09:38 AM
I underclock. Metered electricity isn't cheap, but I am. :P

NCLI
November 9th, 2010, 02:00 PM
I've overclocked the dual-core Intel E8500 in my desktop from 3,16 GHz to 4,20 GHz. No problems at all :D

MooPi
November 9th, 2010, 02:32 PM
I've overclocked the dual-core Intel E8500 in my desktop from 3,16 GHz to 4,20 GHz. No problems at all :D

So is heat an issue ? That is some awesome speed. I've never clocked any cpu past 4 Ghz, but have bumped my Phenom 965 to 3.7 and really pushed an old socket 939 Athlon to 3.6Ghz, but had to really watch the heat.

|{urse
November 9th, 2010, 02:34 PM
Why? Because AMD made those to be overclocked. Just like the Athlon XP's. I remember a few (probably more like 5) yrs ago when people bought AMD just because they were easily Oc'd.

czr114
November 9th, 2010, 10:10 PM
The argument about CPU life is largely theoretical if the machine's useful life expectancy is shorter than the supposedly reduced lifespan of the chip.

If the processor is slow, to the lower bound of usability, then overclocking can extend a machine's useful life and avoid a new purchase.

LowSky
November 9th, 2010, 10:40 PM
Phenom II x2 550 @ 3.1 GHz

Overclocked to 3.5 Ghz and unlocked two additonal cores.

Say hello to my little friend the Phenom II x4 B50

Dustin2128
November 9th, 2010, 11:01 PM
This one overclocks to 11.

Mr Bean
November 9th, 2010, 11:28 PM
I killed a 2.8Ghz P4 running it at about 3.2Ghz I think it was. Worked fine and within tolerable temps but one day soon (a matter of months I think it was) the PC crashed and wouldn't turn on again. One dead CPU. Replaced with a 3Ghz and ran it at 3.4GHz. So lesson not entirely learned.

I don't overclock any of my systems now though. The performance boost you get (if you even notice it at all) is, in my experience, never worth the (usually hours) worth of time it takes to find the right balance of settings. Then the first time your PC does anything even remotely iffy you have to revert back to make sure it's not a hardware issue. Complete waste of time imho.

If you find it entertaining to O/C crappy hardware, or if someone is paying you to O/C top of the line hardware, or if you've found one of those rare CPUs/GPUs that is secretly a much more powerful product underclocked, then that's great. I don't see why you'd bother in any other situation though.

cammin
November 10th, 2010, 12:13 AM
Phenom II x2 550 @ 3.1 GHz

Overclocked to 3.5 Ghz and unlocked two additonal cores.

Say hello to my little friend the Phenom II x4 B50

I did that with mine, but the unlocked cores don't always like being unlocked. They especially don't like flash videos.

The problem is they don't like staying locked, either. So I randomly get a quad core when I boot up. I've had it randomly reboot and become a quad core, once or twice, as well.

As far as overclocking goes, I didn't really need it. So I let the power saving features do their thing and it spends most of it's time at 800mhz.

forrestcupp
November 10th, 2010, 03:02 AM
you could easily decrease your CPU life by many years or worse.

Computers are obsolete after a couple of years anyway.

|{urse
November 10th, 2010, 04:08 AM
What I was told by an actual Cray systems engineer is this:

"If you've ran a cpu stably for one fourth of it's life ( assuming it's going to last 5 years ) at it's default clock then the electron paths are well-worn and will gladly accept a minor increase in core voltage periodically at fourths in lifespan."

I asked him what the acceptable relative core voltage increase would be for every lifespan fourth. My memory is rusty so I'm paraphrasing a bit.

"You can figure that if it is a 1.4ghz duron then it is safe to raise the core voltage ~120mhz every 8 months or so without diminishing the life of the processor."

If i can find the irc log I'll paste it here, but suffice to say that immediately overclocking the crap out of a brand new or a very old cpu is a bad idea.

cascade9
November 12th, 2010, 08:45 AM
Life-span vs Voltage

Here's a graph for those interested:

http://images.anandtech.com/reviews/cpu/intel/2008/e8500-overclocking/Degradation.png


Keep in mind that this is only a few percent increase in voltage. For the small, maybe 10% increase in performance, you could easily decrease your CPU life by many years or worse. As well, the hotter electronics get, the more they conduct, and the more they conduct, the hotter they get. It feeds into itself.

If you do decide to do this I would recommend using a really good zinc-oxide cream to help dissipate the extra heat. And you'll need to apply the cream properly because if you leave any air holes, the chip is sure to blow within minutes or hours. As well a good heatsink is a must!! A powerful fan is needed, but is NOT enough by itself!

Good luck!

That graph could be misleading (in a lot of cases you dont need to bump voltage to overclock), so heres the article-

http://www.anandtech.com/show/2468/1

Not all overclocking is dangerous. For example, II ran a (mild) overclock undervoltaged for years, and so it was running cooler and safer than stock.


If you overclock the processor without increasing the voltage, then you technically are not notching down it's lifespan at all.

The heat increase and reduced lifespan associated with overclocking comes with increasing voltage, which is necessary to do when trying to do the higher overclocks and push the limits of the hardware.

Niot quite...frequency also impacts. Same voltage at higher Mhz = more heat.


That being said, the decrease in lifespan really depends on the chip, as older Pentium III's and IV's are heat-generating monsters, and I myself have several old machines that still run good as ever with those processors in them.

Some of the P3s ran hot (best eamples, the 600 'katamai' and original 1GHz 'coppermine') but in gernal P3s are fairly cool. P4s are hot and nasty.


With the right forethought as to cooling, the lifespan is not nearly reduced as much. Often why some chips have such good overclocking headroom than others (even from the same make and model of chip)is because companies who sell the same processor clockedat 2.4Ghz, 2.8 GHz, and 3.4 GHz often sell much more of the one clocked at 2.8. So sometimes if they need to keep up with the demand, some companies take the 3.4 chips, and underclock them to 2.8 and then sell them, when in reality the chip is made to run at 3.4 for the same lifespan.

Scott

Close.

I'm just going on x86 CPOUs, I cant say that all CPUs are like this. What happens is that Intel or AMD (or cyrix, IDT before they got bought out) design a chip to run at X Mhz, and I'll use 1GHz as an example.

When the chips are finished but unmarked, they are tested. If the chip passes its 1GHz test, its sold as a 1GHz chip. If it fails, they retest at 933. If that fails, 866. If that fails, 800. Etc.
There is a point after which the chip is just discarded as faulty, but when that happens with depend on a lot of factors.

A chip may well have passed its 1GHz test, but due to pricing there might not be much demand for expensive 'top of the line' 1GHz chips, so its marked as 933, 866, 800 etc. Its not really 'underclocked'.

This is one way that you can get a faster chip for less money. Quite often, the slower chips have passed for higher MHz.


Got a P4 2.4GHz, how do I overclock?

Depends on your motherboard.


I underclock. Metered electricity isn't cheap, but I am. :razz:

If your BIOS supports it, try undervolting as well. ;)


So is heat an issue ? That is some awesome speed. I've never clocked any cpu past 4 Ghz, but have bumped my Phenom 965 to 3.7 and really pushed an old socket 939 Athlon to 3.6Ghz, but had to really watch the heat.

You got a Socket 939 to 3.6GHz? Thats a 50% overclock, I'm not suprised you need to watch it.

MooPi
November 14th, 2010, 07:48 PM
You got a Socket 939 to 3.6GHz? Thats a 50% overclock, I'm not suprised you need to watch it.

No that's closer to 28.5% OC. The original clock was 2.8 Ghz I believe. 800 MHz bumb

drawkcab
November 14th, 2010, 08:07 PM
You should OC so that you can school chumps in Crysis deathmatch! Yah!

czr114
November 14th, 2010, 08:36 PM
Modifying the stock clock speed may result in a loss of dynamic frequency scaling.

For those overclocking even one step, your processor wear may increase dramatically due to no lower idling. Going from, for example, 2.66 to 3 is more than just that 13% increase, as the stock 2.66 would be idling much lower most of the time (assuming you're not folding or anything like that).

Likewise, a slight manual underclock can disrupt the system's ability to underclock itself further during the majority of idle time. The supposed savings in electricity or wear might never materialize.

The Real Dave
November 16th, 2010, 01:24 AM
My secondary desktop is a monocore Pentium 4 overclocked from 2.8GHz to 3.4GHz which works out to 21% more if I'm not mistaken. It's stable as a rock, and the OC really makes quite a difference, especially in video playback.

The computer I regularly use in school (http://valid.canardpc.com/show_oc.php?id=1484275) is also overclocked, from 2.53Ghz to about 3.2Ghz

cascade9
November 16th, 2010, 09:08 AM
Modifying the stock clock speed may result in a loss of dynamic frequency scaling.

For those overclocking even one step, your processor wear may increase dramatically due to no lower idling. Going from, for example, 2.66 to 3 is more than just that 13% increase, as the stock 2.66 would be idling much lower most of the time (assuming you're not folding or anything like that).

Likewise, a slight manual underclock can disrupt the system's ability to underclock itself further during the majority of idle time. The supposed savings in electricity or wear might never materialize.

That can happen, its just one of the things you need to check on, not just hope it works.


No that's closer to 28.5% OC. The original clock was 2.8 Ghz I believe. 800 MHz bumb

You must have had a A64-FX-57 then. Not a common CPU, and most of the time I see people with FX CPUs they have tended to be pretty proud of them (so they tend to put Atloh 64-FX or somethign like that).

Still, 3.6GHz from any S939 CPU is getting pretty close to the edge.....