[iceangel89]

hi,

i am new to overclocking and the last time i tried, a loud sound seems to come from my CPU - too hot i guess.

so i read from Here (Hardwarezone)that Intel E2160 can be OCed very well..s it true? anyone tried?

what are the things i can look out for? i article mentions

Quote

Sometimes you might not even hit this hitch till months later and that's when you would find it difficult to pinpoint the issue since your overclock has been stable all this while. Thus it pays to ensure you have the right memory for the job.

so is it safe? and stable? is it ok to OC a new computer? don't want to waste $$

[puntoMX]

It’s safe, the big beep you were facing is a component that is failing at the OCed speed, it could be your RAM, CPU or VGA card if you have one.

If you start OCing you need to do it step by step and learn how far each component can go, it’s not only the CPU that you OC; When bringing up the FSB all components will be OCed.

[nmX.Memnoch]

puntoMX, on Aug 23 2007, 11:24 AM, said:

When bringing up the FSB all components will be OCed.

That's not entirely correct either. Several years ago raising the FSB would also raise the PCI and AGP buses, as well as overclocking the IDE controller (in some cases). But newer boards have features that keep the PCI, AGP and PCIe buses within specifications. Most of them even have memory dividers so you can choose not to overclock anything but the CPU.

This post has been edited by nmX.Memnoch: 23 August 2007 - 10:42 AM

[puntoMX]

Just some side text, but we are talking about some one that starts OCing, most of them don’t even know how to set FSB-divider although some options are on "auto".

It’s known that nVidia was the fist one that released chipsets that had separated settings for each bus.


Ps. nmX.Memnoch, I’m not correcting you (you know your stuff mostly), just some additional information, that’s all.

[nmX.Memnoch]

Yeah, I understand that...but with the newer boards you don't have to set the divider. The BIOS does it automatically for the PCI and PCIe buses. The only divider you have to fiddle with is the memory divider.


And as you said, I'm not trying to correct you either. I'm just trying to make sure all the information is presented.

[Zxian]

Just a personal note - I never consider overclocking safe at any time. The CPU you have is designed to run with an 800MHz FSB (actually 200MHz), and at a clock speed of 1.8GHz. Anything higher, and you're running outside the tested limits.

If you're willing to overclock your components, you're willing to break them IMO. Overclocking stresses the components more than they were designed for. If they break while under normal operation, you can get a new one. If you've been overclocking, then getting a new one is harder (and you've got an ethical dilemma to face).

[ripken204]

ya overclocking isn't all that safe but it is pretty hard to mess something up unless you arn't paying attention. just watch the temps and voltages and you will be fine. this is where after market coolers come in handy.

This post has been edited by ripken204: 23 August 2007 - 02:42 PM

[Zxian]

ripken204, on Aug 23 2007, 01:40 PM, said:

ya overclocking isn't all that safe but it is pretty hard to mess something up unless you arn't paying attention. just watch the temps and voltages and you will be fine. this is where after market coolers come in handy.

I think you missed my point...

In practice, yes, you can usually go beyond the stock limits of your system without running too much risk. However, you're still running a risk. Even if you raise that clock speed from 2.4GHz (for you and I with the Q6600) to 2.5GHz, you could break your CPU, regardless of temps or voltages.

It's like day trading on the stock market - you can gain really big, but you have to be prepared to lose everything.

[ripken204]

Zxian, on Aug 23 2007, 04:53 PM, said:

ripken204, on Aug 23 2007, 01:40 PM, said:

ya overclocking isn't all that safe but it is pretty hard to mess something up unless you arn't paying attention. just watch the temps and voltages and you will be fine. this is where after market coolers come in handy.

I think you missed my point...

In practice, yes, you can usually go beyond the stock limits of your system without running too much risk. However, you're still running a risk. Even if you raise that clock speed from 2.4GHz (for you and I with the Q6600) to 2.5GHz, you could break your CPU, regardless of temps or voltages.

It's like day trading on the stock market - you can gain really big, but you have to be prepared to lose everything.

no i knew what you meant, there are just ways to become more safe if/when you do decide to overclock.

[puntoMX]

Or just don’t use your computer, less risk .

Does any one really think it matters when a CPU works 7 or 10 years? And why people don’t buy the lowest clocked ones that are made with the same process? Lower speeds are just made for the market remember that .

So, it’s not a big risk if you try to OC that CPU when you know what you are doing, this is why I posted “It’s safe,…”. The page linked too didn’t rise the voltage at all, and someone should make a backup of his computer first that’s for sure because there could be data corruption of the hard disk in 0.09% or so…

[RJARRRPCGP]

Zxian, on Aug 23 2007, 04:36 PM, said:

Just a personal note - I never consider overclocking safe at any time. The CPU you have is designed to run with an 800MHz FSB (actually 200MHz), and at a clock speed of 1.8GHz. Anything higher, and you're running outside the tested limits.

If you're willing to overclock your components, you're willing to break them IMO. Overclocking stresses the components more than they were designed for. If they break while under normal operation, you can get a new one. If you've been overclocking, then getting a new one is harder (and you've got an ethical dilemma to face).

But, with newer chips, they may actually be sold underclocked, for example, an Athlon 64 3700+ may be an underclocked Athlon 64 4000+!
That may be why some people can raise the frequencies far beyond stock without any problems and not touching the Vcore!

It's not like the 486 days, where the chips may get killed just from a 10 mhz raise!

This post has been edited by RJARRRPCGP: 23 August 2007 - 03:43 PM

[puntoMX]

RJARRRPCGP, on Aug 23 2007, 03:40 PM, said:

It's not like the 486 days, where the chips may get killed just from a 10 mhz raise!

Never killed any by setting it on a higher speed, it just simply works or doesn’t. There are electronics that can’t take a higher switching speed, but never found this on CPUs...

EDIT: By the way, does any one know what electrical design the Pentium Pro had? Guess what, 5Volts, and were sold as 3.3volt CPUs. Those babies could do more then 300MHz .

[nmX.Memnoch]

Overclocked CPUs don't generally last 7-10 years...unless you're doing some very conservative overclocking. You do have to be careful and know what you're doing. Do you think someone asking "is it safe" would be equated to someone who knows what they're doing? I don't.

Look up electromigration too. Overclocking increases the speed at which electro-migration happens, or sometimes introduces it where it otherwise wouldn't happen at all. Also look up Sudden Northwood Death Syndrome. You could get some crazy overclocks out of the Northwood P4's, but they wouldn't last that long. I was a victim of it...twice. I had a 2.26 that I wasn't even pushing that hard.

Additionally, did you know that for roughly every 10C that you reduce the CPU temperature that you roughly double it's life?

Sure...overclocking is fun if you have the money to replace the CPU. If you upgrade your CPU once a year you probably won't even see the damage you actually did to the CPU...but you better make the next owner of that abused CPU aware of what you did to it. Personally, I opt to spend a little more on the fastest CPU I can afford at the time of purchase and then keep that for 2-3 years instead of replacing it every 6-12 months.



http://en.wikipedia....lectromigration

Quote

Overclocking of processors, especially when using higher than nominal voltage, causes electromigration between their transistors and significantly shortens the chips' lifetime.

Quote

Some of what appears to be spare margin is actually required for proper operation of a processor throughout its lifetime. As semiconductor devices age, various effects such as hot carrier injection, negative bias thermal instability and electromigration reduce circuit performance. When overclocking a new chip it is possible to take advantage of this margin, but as the chip ages this can result in situations where a processor that has operated correctly at overclocked speeds for years spontaneously fails to operate at those same speeds later. If the overclocker is not actively testing for system stability when these effects become significant, errors encountered are likely to be blamed on sources other than the overclocking.

[puntoMX]

Yes, but we are talking about the link posted above, no voltage increase, just switching speeds...

[nmX.Memnoch]

You don't have to raise the VCore for electromigration to occur. 99 times out of 100 it's going to be the L2 cache that goes first...and since there's so much of it on modern CPUs that makes the effect even worse.

He's already purchased an entry-level CPU so I'm going to assume he was on a limited budget when it was purchased. That probably means he doesn't have the money to replace it right away. And as Zxian said...sure, it's under warrantly but then you have the moral issue to deal with. It wasn't a product defect that killed it...you did it yourself by running it outside of the manufacturer's tested specifications.

Not to mention that every bit of data that is processed on the system goes through the system CPU, hence the central part of "central processing unit". If there's a cache miss and the CPU sends corrupted data to the drive controller for it to be written to the drive, then it's going to be written corrupted. The hard drive controller doesn't know or care that the data was corrupted...it just writes what it's given. Yeah, there's some ECC going on, but what do you think is doing the checking (hint: the CPU)?

Are you going to be responsible for just telling someone "yeah, it's safe, go ahead" without filling them in on all of the possibilities...and then they start getting data corruption while everything appears to be running fine?

Do you know what most game developer's first question is when you open a support ticket if you're having a problem with their game? "Are you overclocking?" They want to know because their first suggestion is going to be to try the game at the CPU's stock speeds. 90% (made up number, but probably not far off) of the time if fixes the problem.

This post has been edited by nmX.Memnoch: 23 August 2007 - 04:11 PM

[ripken204]

im just going to talk about the game part of that nmX, the ppl with those problems i think are the ones who don't test for stability

also when you talk about the -10C = 2x life, i wonder how that will convert for someone at -200C on a 4 stage phase change at well over 4GHz

[iceangel89]

thanks. i guess i don't OC 1st, until i need to or the warranty ends. but whats the performance of the E2160? anyone knows or maybe i save more to get a better CPU instead. i dont want to get something whose performance is bad...

i do have a P4 530 but that comp is spoilt can't boot now i posted http://www.msfn.org/...howtopic=103415

[ripken204]

the performance is good for most things that you want to do and its really cheap.

[Zxian]

@iceangel89 - You say "how good is the performance"... compared to what? Compared to the P4-530 you've got, the performance is much better, but compared to any other C2D chips, the E2160 isn't much to write home about at stock speeds. It's fast enough for plenty of things, but it's no match for an E6400 or anything of the sort.

Like ripken said - you can't really beat the price though. I can get one here for about $100 CAD, and for a basic system, that's pretty freakin wicked.

[nmX.Memnoch]

ripken204, on Aug 23 2007, 08:26 PM, said:

im just going to talk about the game part of that nmX, the ppl with those problems i think are the ones who don't test for stability

Again...just blanket recommending "yeah, it's safe" to someone who doesn't now anything at all about overclocking isn't something you can do. He asked if it was safe to overclock a 1.8GHz CPU to 3.33GHz, a clock speed that's over 50% faster than it's rated for, and the answer he got was "yeah, sure...go head".

Now, do you think he would go about it by incrementally adjusting the FSB and other settings? No...he's going to go straight to 3.33GHz, thereby taking a chance on data corruption or even worse, burning something up. And who's going to be to blame?

ripken204, on Aug 23 2007, 08:26 PM, said:

also when you talk about the -10C = 2x life, i wonder how that will convert for someone at -200C on a 4 stage phase change at well over 4GHz

And exactly how many people out of the millions that use a computer every day are running phase change? Check that...how many out of those that are overclocking are running phase change? I think you'll find the percentage is still rather small. But yeah, their CPUs will last for a while.

Lemme ask you this as well...would you spend a bunch of money on a phase change system to only put a $100 CPU under it?

This post has been edited by nmX.Memnoch: 23 August 2007 - 08:01 PM