Overclocking?

[Pixulated]

Okay so a friend of mine is always bragging to me about how he overclocked this to that etc.

I really don't know anything about it and what it does... could someone just explain really quickly and tell me how to go about it?

Thanks

[Brad]

Basically you change some settings in the bios or via software which increases the speed of the said part your overclocking which in turn gives better performance.

[Pixulated]

Ahh thank you, I'll leave the rest upto my mate google then :P

[Brad]

Well if you want any advice then would be good if you stick your spec in your sig :)

[Helior]

Advice can certainly be offered but tbh your much better off googling and reading about it first. There are tons of good guides out there which explain what overclocking is, how it works, why it is a good idea etc.

[coursemyhorse]

Overclocking gives you a "free" performance boost, in that you run components at a rate higher than "stock". Stock being the intended base rate if you like. i.e. CPU advertised with clock speed of 2.5ghz, can be overclocked to 3.0ghz. if you select the right components and configure them well, you can save yourself some money by maxing out their potential.

[Andrew Moore]

Overclocking gives you a "free" performance boost, in that you run components at a rate higher than "stock". Stock being the intended base rate if you like. i.e. CPU advertised with clock speed of 2.5ghz, can be overclocked to 3.0ghz. if you select the right components and configure them well, you can save yourself some money by maxing out their potential.

"free" in the sense getting a 2ghz cpu to hit 3ghz would save X cash from buying the faster part. however "free" isnt free when it comes to your electricity bills. Overclocking a cpu by 50% will mean it will need more power to run. This power may be more than the power that is used to run the same part that is at the higher stock speed. So after a year or so you may even be out of pocket in terms of paying for the extra leccy it used.

It is run and rewarding to overclock though and mind over matter makes you believe that its cheaper to do this but logical thinking would in some cases state otherwise.


Andy

[Helior]

It will be longer than a year before you start losing out, a lot more in fact, if thats a concern just counteract it with a better PSU!

[coursemyhorse]

"free" in the sense getting a 2ghz cpu to hit 3ghz would save X cash from buying the faster part. however "free" isnt free when it comes to your electricity bills. Overclocking a cpu by 50% will mean it will need more power to run. This power may be more than the power that is used to run the same part that is at the higher stock speed. So after a year or so you may even be out of pocket in terms of paying for the extra leccy it used.

It is run and rewarding to overclock though and mind over matter makes you believe that its cheaper to do this but logical thinking would in some cases state otherwise.


Andy

I'd like to see some real world examples of your claims rather than the majority of speculative posts on the subject of electricity. I see way too many posts like this with no proof/links to how overclocking a CPU would really increase your electricity bill annually. Given that you can use speedstep (CPU speed/power is reduced when idle) with good reliability these days even when overclocking, it's not really an issue.

In fact your post is a little confusing. What are you actually saying?

Person A buys a CPU @ 3.6ghz for £300
Person B buys a CPU @ 3ghz and overclocks to 3.6ghz for £200

Person B has saved £100 and lets assume it's running at roughly the same power useage as person A since the CPU is the same type and clock speed is matched.

How is that NOT saving money? I think what you mean is purely limited to just running higher powered components generally right?

[Helior]

Typically the overclocked one would use more power since high FSB is slightly more power hungry than high multi, its not significant though.

[PMM]

CPU power consumption is tied into clock frequency, the faster it goes the more power used that is fact.

However the cost to run is mute. worst case senerio on a E8400 24hrs a day * 365 days a year on a flat rate 30p per Kilowatt with an CPU wattage gain of 20 watts over std =175Kilowatts / 30p

= £52 a years 24/7.

However you have to take into account the lower rates / economy 7 night units etc so infact the electricity cost is substancially lower than savings made getting a lower spec cpu and clocking it and the fact no doubt you wont have it running 24/7 & 365.

[GeorgeStorm]

A lower end chip will require higher volts to reach higher clock speeds than a higher end chip,
IE, buying an e8200 and overclocking it to 3.33 will require more volts than the e8600 at stock
So yes, while you will save money, if you keep it running 24/7 for a year or two, it could end up costing you more.

[gurusan]

^actually I've had several E8200s and all of them could do 3.6GHz on stock voltage of 1.25V

If you guys want to save on power bills than ditch the 65nm tech. My E6750 @ 3.8ghz used about 25W idle and 105-110W load....my E3120 at 4ghz uses about 8-10W idle and 70W full load.

[GeorgeStorm]

It was only an example, maybe not a good one
What are you measuring with to get just the CPU power usage?
8-10W idle? Seems a little unrealistic,?

[gurusan]

the P5Q epu chip. For reference my old q6600 @ 3.6ghz idled at about 60-75w so it seems realistic.

[PMM]

It was only an example, maybe not a good one
What are you measuring with to get just the CPU power usage?
8-10W idle? Seems a little unrealistic,?

not unrealistic 45nm architecture idles as about 4watts per core when referring to actual CPU power consumption and not overall system consumption you will find that documented on the web from reliable resources.

Equally the fact that a lower spec chip requires more power than a higher chip is not true - it all dependant as you should know on the quality of the silicon and its characteristics as defined when etched on the wafer.

Binning is not fool proof and equally supply issues may result in higher grade chips being down graded to make the shortfall.

I have brought lower spec'ed CPU's that have surpassed both in speed and required voltage that of there bigger brother.

Equally traits of a chip being as odd as they can be... undervolting sometime provides performance in excess of overvolting in a small number of circumstances so the whole argument is somewhat flawed.

that's why alot of peeps in the benchmarking game do so much cherry picking of CPU's to find the right one with the right characteristic to make a certain speed.

[GeorgeStorm]

Yeah, checked it out, im surprised they use so little power tbh.
I know benchers cherry pick, however, i dont know of a case where an e8400 will beat an e8500, if both have been cherry picked, the best e8500 beats the best e8400 etc.
Not to say that an e8400 cant beat an e8500, as I have known some that do, easily.
However, if you take the chips in general, an r8600 will require less volts to hit 4, or 4.5 than an e8400,

[PMM]

You have to remember that essentially they all have the same potential having come from the same 300mm wafer its only in-house grading that defines its rating and as mentioned in times of demand such binning maybe overlooked and high end stock put into std stock.

But one of the things to be aware of is a chips characteristic can vary over the span of the chips speed range - therefore while say a chip gets binned into a 8200 and one gets binned into an 8400 it can be down to the fact it did not meet the spec at that desired speed in relation to many factors .....inc voltage / tested temp etc.

However as you increase the speed & improve the cooling the characteristic of a chip further up the frequency band can differ so you cannot say a cheery picked 8400 would beat a cherry picked 8200 based on its denominator because electrical characteristic can differ and does not follow any linear characteristic.

Its entirely possible for the best chip to have found its way into a low ranked chip, that low ranked chip may be 0.05 volts more than its higher ranked equivalent at the same speed but at an increased speed the electrical characteristics can change enough that you can find the lower ranked chips out performing the higher ranked for a given voltage.

Its a complex world of chip selection for performance and also you have to factor differing characteristic of various motherboards to be able to drive those CPU's as well some do better than others and that can extend to voltage requirements as well - I have used the same chip in differing boards before and the required voltage has altered also (DMM confirm) and equally the ability to push to particular speed can be motherboard dependant not out of the ordinary to see as much as a 150mhz difference just on choice of board alone.

Its really a pick and mix - end of the day the speed ranking only guaranties that it meets that speed spec - but its overall ability to succeed that is one of luck.

Then there is the whole debate of burning in a chip - I have done burn ins to gain improvements some very successful some not so successful and some producing negative results...so even yourself as an end user can re-define the chips ability if the chip ends up running hot.

[gurusan]

From my experience the sweet spot of wolfies is the E8500. I've come across 3 different "golden chips"....and none of them were E8600s or E8400s. The E8200 and E8300s I've had have been mediocre clockers...most topped out around 3.8-3.9ghz with 1.4V yet they would do 3.6-3.7 on stock volts.

Could have been coincidence but I don't believe it was. Although I did have some pretty dreadful E8500s as well that topped out around 4.2-4.3ghz with 1.4V.

My current E3120 is the absolute worst E8500 I've had . I can't get the darned thing stable above 4050mhz...and it doesn't like any FSB above 450. Such a POS.

[GeorgeStorm]

Well, I still think that the best chips are binned the best
If you can show me any super chips which are lower down the range, please do, as ill buy, :P
But ofcourse im not limiting it due to voltage, when im talking about cherry picked cpus, im saying the max they can do, regardless of voltage, for benching
So then, you put whatever volts it wants through it, and you still cant get it higher, even when subzero, thats its limit, and i would be impressed if a lower range chip can beat the best of the top.

[I3R0K3N7FEET]

tbh, i think overclocking could be part of a future upgrade plan too,eg.. by a stock part that is pretty good, and later when you upgrade gfx buy an new cooler and overclock a little, and tbh if youre worried about arranties, wait till the warrantees over and then clock the **** outta it.. things that can be overclocked< as its not been said are generally, CPU< GPU< RAM also ive only been overclocking for about a year, it is definately best to do a lot of research before doing it and make sure you know the parameters of the componants youre going to clock, otherwise a little elightenment is all you need and nothing beats actually doing it.. use common sense and dont try to run 7 volts through something that says 1.250V, tbh you would have to be particulary special to trash componants anyway as crashing is actually a part of a PC's protection system, but if youre not comfortable with seeing your pc crash and starting over again then overclocking is not for you.

also george is right, except for some 'gems' such as the q6600, in general if you took a range of cpus from low end to high end tehn clock them all, the results will usually be the same the low end will still be the lowest performer and the highest will usually still be the highest performer, but as i said, there are some 'gems' as many things a batch produces some chips are better than others of the same type..

[erphy100]

It will be longer than a year before you start losing out, a lot more in fact, if thats a concern just counteract it with a better PSU!
Turn off the driveway lights 15minutes earlier.

[I3R0K3N7FEET]

replace all the bulbs in your house for A rated halogen bulbs

+1
lets all go energy efficient so we can afford to run our energy eating machines!! WHOO!!

[erphy100]

replace all the bulbs in your house for A rated halogen bulbs

+1
lets all go energy efficient so we can afford to run our energy eating machines!! WHOO!!


Got eye strain from these bl**dy candles, I've got to use candles so I can afford to run my PC. A friend works in a candle factory so I get them free, before anyone asks how much I save.

[Lambda]

I hope you're planting trees to offset your carbon emissions also!

[PMM]

Got eye strain from these bl**dy candles, I've got to use candles so I can afford to run my PC. A friend works in a candle factory so I get them free, before anyone asks how much I save.

now see you missing a trick there.... peltiers + candles = free energy

you see the candles create a heat source... a cool surface to the base of the petlier & the differential in heat between both sides create voltage/current that will be energy to offset you PC power consumption ;) :D :banana:

[I3R0K3N7FEET]

im planting trees to burn them.. to power my pc :D << not.. not yet anyway lol

[erphy100]

im planting trees to burn them.. to power my pc :D << not.. not yet anyway lol
You disappoint me, I feel like burning the bl***dy PC now. Maybe in the winter.

[I3R0K3N7FEET]

the technology to provide free energy is there theoretically, its just the evils of man wont allow such an act.

[Helior]

the technology to provide free energy is there theoretically, its just the evils of man wont allow such an act.

The technology to provide some free energy is, not nearly enough of it though.

[I3R0K3N7FEET]

no<< more than enough

im wondering if obama has the balls to release the technology... im glad he released the anti gravity technology stuff :)

[Helior]

tbh, i think overclocking could be part of a future upgrade plan too,eg.. by a stock part that is pretty good, and later when you upgrade gfx buy an new cooler and overclock a little, and tbh if youre worried about arranties, wait till the warrantees over and then clock the **** outta it.. things that can be overclocked< as its not been said are generally, CPU< GPU< RAM also ive only been overclocking for about a year, it is definately best to do a lot of research before doing it and make sure you know the parameters of the componants youre going to clock, otherwise a little elightenment is all you need and nothing beats actually doing it.. use common sense and dont try to run 7 volts through something that says 1.250V, tbh you would have to be particulary special to trash componants anyway as crashing is actually a part of a PC's protection system, but if youre not comfortable with seeing your pc crash and starting over again then overclocking is not for you.


Althought technically you do break warranty when you overclock it is virtually impossible to prove unless your an idiot, as long as common sense is applied then it should not be a concern.

I do agree that overclocking can be a future upgrade, but i personally don't think thats sensible. While overclocking probably does reduce MTBF, the reality is that with modest or even a fairly heavy overclock, the MTBF still exceeds the useful life of the CPU, so with that in mind you may as well simply overclock it right away and get the full benefits imo.


also george is right, except for some 'gems' such as the q6600, in general if you took a range of cpus from low end to high end tehn clock them all, the results will usually be the same the low end will still be the lowest performer and the highest will usually still be the highest performer, but as i said, there are some 'gems' as many things a batch produces some chips are better than others of the same type.

On average that is probably correct but the differences are small. Unless there is a difference in stepping (eg Q9650 was E0 long before the q9550) then generally the bottom of the range is normally the best value by a long long way.

[I3R0K3N7FEET]

agreed :)^^ though i wish i was a bit more clued up about cooling<< my asus trition 78 could get 3.8 on the 940 earlier this year but as the season changed, my room is a furnace and i can barely get over 3.6 stable, though tbh i keep it at stock speeds mostly.. i went on how the cooler looks rather than performs.. i wish i bought a true..

[PMM]

Actually if you killed a cpu overclocking.. in essence you can read the voltage out of one of the smd capacitors on the chips underside and conclude what voltage it was running and potentially prove it was out of spec.

[Helior]

Actually if you killed a cpu overclocking.. in essence you can read the voltage out of one of the smd capacitors on the chips underside and conclude what voltage it was running and potentially prove it was out of spec.

Maybe if you read it straight away but surely the capacitor will discharge quickly enough to make it horribly inaccurate. The voltage specs provided by intel are pretty forgiving anyway.