overclocking

[Magic Carpet]

Don't understand the hype about it. It's "okay" when a reviewer shows off a particular chip capabilities but when manufacturers are starting to rely on this "feature", I question their intelligence.

Except for the so called "life or death" situations, overclocking doesn't make sense to me.

Common sense. 30% effort, 70% result.

Overclocking sense. 70% effort, 30% result.

If you are after the performance, it would make more sense to build a cluster instead. Seeing the people that are getting 2011 chips and running 24/7 oc on them is just... *no comment*. I am sure as hell, Intel would be selling higher-clocked chips if there was a rational sense but they don't, for a reason. Partly AMD's fault, partly these chips just aren't good enough for those clocks. Surely, many reviewers / internet posters will beg to differ but this is just my opinion.

Some people don't really know what they need/want.

 

[bononos]

Intel would sell the absolute minimum clocks for their cpus if they could but they can't and have to rely on multiplier locking and what not. Have you any reason to believe those chips aren't good for higher clocks?

 

[Magic Carpet]

Intel would sell the absolute minimum clocks for their cpus if they could but they can't and have to rely on multiplier locking and what not. Have you any reason to believe those chips aren't good for higher clocks?

Stability issues aside, as the clock/performance goes up, so does the power consumption. To the point, where even hard-core overclockers realize it is not worth it (performance per watt). Obviously, it's the people's choice how they run their computers and what's good for them. I just hate how it's being marketed to an average Joe, nowadays.

If you think logically, if those K binned CPUs were comfortably safe running @ 5 Ghz, they would have been spec'ed appropriately. There are lots of ifs and but's that aren't mentioned in the selling pitch.

 

[antisocialmunky]

It is fun to buy a cheap chip and have it overclock more than a more expensive chip like that Athlon X4 631 that we were playing around with.

Intel's locking of the chips isn't as big of a deal as charging $100 for hyper threading while having it enabled on the i3.

 

[bononos]

...... If you think logically, if those K binned CPUs were comfortably safe running @ 5 Ghz, they would have been spec'ed appropriately. There are lots of ifs and but's that aren't mentioned in the selling pitch.

If they were spec'd higher, it would eat into the sales of higher end cpus. Thats all I'm saying. 5Ghz would be a far out dream goal for an oc'er, what about those with clocks at 3.6 or 3.8?

 

[dguy6789]

The Sandybridge chips are safe running at substantially higher clocks than stock. They don't come that way stock exclusively for the reason of sales and marketing. Same with pretty much every chip Intel has made at least since Core 2.

It works like this(all numbers are hypothetical): Intel produces processor A, it is capable of ~5Ghz. It only needs to be run at ~3.5Ghz to compete with competitor successfully. Intel sells processor at 3.5Ghz. As competitor becomes more competitive Intel can release the same processor at higher Ghz as necessary to maintain their dominance in sales without increasing cost to themselves. There is no incentive for them to throw down their ace in the hole when it is far beyond what is necessary to meet their sales goals. Not launching at higher clocks has nothing to do with safety or reliability, it is 100% nothing but strategic marketing.

 

[Magic Carpet]

It is fun to buy a cheap chip and have it overclock more than a more expensive chip like that Athlon X4 631 that we were playing around with.

Well... overclocking the lower end of the same family isn't quite the same extreme. Both driven by greed though.

Intel's locking of the chips isn't as big of a deal as charging $100 for hyper threading while having it enabled on the i3.

Pricing is incredible in this industry and often does not make sense. That's how it works though. You just don't want to be the last in this chain. End-user pays up for everything. I wonder, what is the % of people who keep up with these things, *only just* to conveniently recycle their hardware and upgrade. Which brings up another question, what is the % of people who effectively make use of their hardware. As an old saying goes, it's not what you got, but it's how you use it.

If they were spec'd higher, it would eat into the sales of higher end cpus. Thats all I'm saying. 5Ghz would be a far out dream goal for an oc'er,

They could have, but didn't. Few people would consider buying a ~300W TDP CPU with a hefty price tag. Whereas right now, overclocking can be treated as a bonus / additional value. That's what really drives people mad. Emotions take over rationality.

what about those with clocks at 3.6 or 3.8?

They could have easily spec'ed them at these clocks. I have no answer to that. Somebody at Intel could probably shed some light on these things.

Not launching at higher clocks has nothing to do with safety or reliability, it is 100% nothing but strategic marketing.

I disagree, the current thermal envelope wouldn't allow for such a chip (5 Ghz) to make into production. Regardless of marketing and stock market.

 

[MustangSVT]

are u serious?

 

[SickBeast]

OP I've gotta say, IMO your logic is flawed and you're coming across as not making any sense.

Overclocking is incredible for the consumer. You can sometimes double the speed of your CPU for free.

In my case, I added 80 watts to the TDP of my CPU by overclocking. That's not much. The performance I get out of doing it is worth it and takes my processor from being obsolete to being relatively modern.

 

[BD231]

Yeah I guess Intel charge's a premium for the K series CPU's and does anything it can to lock multipliers because they believe overclocking is a fruitless venture .... . Don't be retarded OP, there's substantial evidence pointing to the contrary.

 

[bononos]

....
They could have, but didn't. Few people would consider buying a ~300W TDP CPU with a hefty price tag. Whereas right now, overclocking can be treated as a bonus / additional value. That's what really drives people mad. Emotions take over rationality.

Many people tinker away in their garages for the same reason people fiddle around on their 4-5Ghz projects.
Then there are people who just want stable oc's and bump up clocks slightly and their behaviour is quite rational.

They could have easily spec'ed them at these clocks. I have no answer to that. Somebody at Intel could probably shed some light on these things.
....

Many posters and myself have pointed out several times that Intel has no reason to spec their chips any higher for sales and marketing reasons. Clueless everyday consumers would be content with stock, ATOTs are smart enough to realise some benefits from gentle oc'ing.

 

[Shmee]

sometimes its all about the fun of tweaking it and getting there.

 

[Rifter]

are u serious?

This

OP I've gotta say, IMO your logic is flawed and you're coming across as not making any sense.

Overclocking is incredible for the consumer. You can sometimes double the speed of your CPU for free.

In my case, I added 80 watts to the TDP of my CPU by overclocking. That's not much. The performance I get out of doing it is worth it and takes my processor from being obsolete to being relatively modern.

And this

Yeah I guess Intel charge's a premium for the K series CPU's and does anything it can to lock multipliers because they believe overclocking is a fruitless venture .... . Don't be retarded OP, there's substantial evidence pointing to the contrary.

And this.

OP what exactly are you trying to say?

I know exactly what i need and want. I need a CPU that can push a multi GPU setup and i want it as cheap as possible, hence my i7 930@4.2Ghz.

 

[Dadofamunky]

are u serious?

Second that.

 

[BrightCandle]

Every single CPU die produced is in some sense overclocked. Each of the pieces of silicon is tested for voltage and clock speed and then binned to run at a particular speed. Just because there is a CPU sold at the 35W bracket that runs at 1.8Ghz doesn't mean that the exact same design of silicon running at 95W at 4Ghz is overclocked, its simply setup differently based on how it was originally binned and what chips they need.

Right now we are getting a lot of overclocking headroom out of our chips because the manufacturers are trying to hit a power budget that is lower than the lofty heights. While PC gamers might not care all that much about 35W of extra consumption the enterprise and mobile users do, grey box users aren't the big growth opportunity. Overclockers are simply optimising performance and moving through the performance/power curve further than Intel/AMD will for a sold chip. This is not a waste of time when you consider that you can get a 50%+ increase in performance for no additional cost. Which for some of the people doing this (myself included) saves them considerable time every single day in their professional capacity. Unless you push right to the limit it wont produce an unstable or an early failure of the machine.

 

[Magic Carpet]

There is no such thing as a free lunch. You just fail to see the big picture. There are additional costs to that 50% OC in terms of components and electricity. You may not care about total cost of ownership. System enthusiasts normally do not. But you guys, are the minority. Knock, knock!


What is not to understand in my original post?

Don't understand the hype about it. It's "okay" when a reviewer shows off a particular chip capabilities but when manufacturers are starting to rely on this "feature", I question their intelligence.

Here, I meant, the FX8150 chip.

Except for the so called "life or death" situations, overclocking doesn't make sense to me.

That covers a situation when you have got to catch a plane and want to stuff up your iPhone with movies and obviously you have little time to do so. Or, you may have skipping and choppiness with sound and video playback (more applicable to older CPUs though). And the like.

Common sense. 30% effort, 70% result.

Overclocking sense. 70% effort, 30% result.

This is actually the point why we have gone multi-core in the first place. Do more with less (watts). There is no need to overclock a modern cpu for general use. Certainly, you can if you want to.

If you are after the performance, it would make more sense to build a cluster instead. Seeing the people that are getting 2011 chips and running 24/7 oc on them is just... *no comment*. I am sure as hell, Intel would be selling higher-clocked chips if there was a rational sense but they don't, for a reason. Partly AMD's fault, partly these chips just aren't good enough for those clocks. Surely, many reviewers / internet posters will beg to differ but this is just my opinion.

The point here is quite simple. If you are doing massive calculations/rendering for a living (that 50% OC won't save you anyway). It is just better to use a cloud cluster, you can remote desktop / walk to.

Some people don't really know what they need/want.

There are people that set trends and there are people that follow them. You can probably work out the percentage on your own.

I have tried my best to explain my point here. If you still fail to comprehend it, then just let it go. Thank you very much.

 

[SlowSpyder]

I got my 1090T for $150 and overclocked it to over 4GHz. I honestly spent less than an hour tweaking it, then I let the stability tests run while I was doing other things around the house. I really did not need to put a ton of effort into this, and for it I got 33%+ more clock speed than the factory clocks. What's not to like about Black Edition chips (and other overclocking friendly chips) so long as the cost is reasonable?

And besides, it's a hobby for most of us. Many hobbies don't make financial sense, it's done for the enjoyment of doing it. I don't understand how anyone could spend huge dollars on stamps, but stamp collectors do.

 

[Idontcare]

This is actually the point why we have gone multi-core in the first place. Do more with less (watts). There is no need to overclock a modern cpu for general use. Certainly, you can if you want to.

No phrase comes close to displacing the intentionally ambigious phrase "general use" when it comes to precipitating forum arguments over the relevance of computing performance at any point in time - be it overclocking, core counts, price-points, etc.

I hope you can appreciate the fool's errand you have set yourself on here by invoking this ill-defined phrase as the fundamental tenet to legitimize your assertions.

The point here is quite simple. If you are doing massive calculations/rendering for a living (that 50% OC won't save you anyway). It is just better to use a cloud cluster, you can remote desktop / walk to.

Please help me understand how my application of interest (MetaTrader 4.0), for which its time-to-answer is rate-limiting in the process of effectively using it (the Strategy Testing feature), stands to benefit from running on a cloud cluster.

The Strategy Tester routine in this application is single-threaded, and backtesting simulations can easily turn into month-long calculations.

The developer (metaquotes) has stated in no uncertain terms that it will not be releasing this app as a multi-threaded app, performance improvements for the end-user will only come by way of higher IPC and/or higher clockspeeds from our CPU's.

 

[AdamK47]

There seems to be a disconnect on how people who overclock use their system and how Mr. Carpet thinks they use it, or should use it.

 

[Majic 7]

There seems to be a disconnect on how people who overclock use their system and how Mr. Carpet thinks they use it, or should use it.

Agreed. HPZone -->