Overclocking my CPU for this. Worth it?

[LittleMe]

Guys, ManoWar2 and TruXter are intentionally posting BS because they love the drama and attention. Its called ego and pain body feeding. It sucks to see someone intentionally and forcefully post flat out false information but they are going to do it anyway. Read ManoWar2's English and his education level is clear. Just ignore their BS please.

[Vicetopia]

Studying electronics != studying microprocessors.

There's a difference, and that difference is what makes Manowar2 wrong.

 

I don't feel like joining the argument, but whoever is disagreeing with that moron is right.

[LittleMe]

 

I do not need internet or any proof for that

You and George Bush don't need evidence.

[TruXter]

I do not need internet or any proof for that

You and George Bush don't need evidence.

I know, the wtc was a figment of our imaginations.

and all the other suicide bombings too.

 

and solder lasts for ever at greater than 43 degrees Celsius.

[Rokz0r]

I get a few kicks reading some of the "educated" posts on these forums - i have been a computer tech for 7 years and currently studying games graphics programming and let me tell you the amount of crapola going on on these forums and the random claims about tech make me laugh/cry so hard i feel like taking my pants off and singing!

[dirty-dog]

I overclocked my 2.4Ghz Qaud to, 3.24Ghz and it makes a lot of FPS difference as my GFX card was out performing my CPU, I used overclockers.co.uk forums for help and of course I have an upgraded CPU fan, it was a lot easier than I thought, and the jump in FPS surprised me, although this won't always be the case, it cured my problem but might not cure yours

 

 

Free speed, why not 

I agree.

 

@dirty-dog

Certaily made a big difference even going from 3.7 to 4.0 with 1920x1200 resolution for me, saying it deos nothing is not true.

You just aren't clocking it high enough.

I did say that's how it is on MY system at the resolution *I* use. If I used a higher resolution or a different graphics card, perhaps it would be different. You do understand the principle that if a game is GPU limited on a given machine, then overclocking the CPU further won't improve performance, right?

[ManOWar2]

 

If your CPU is 30c at 3ghz is it faster than if it is 40c and 3ghz?

I have to say yes to that.

 

 

Where is your data to support that claim? 0.01 percent is not worth mentioning. A cpu at 30c will computationally perform significantly better than at 40c? I call bs, but I've been mistaken before. I don't recall seeing performance graphs to back that claim in the years of computer and cpu reviews I've read.

 

If the performance benefit isn't worth mentioning, then its certainly not worth arguing about so I presume you are saying it is. Enlighten us.

I have nothing to prove.

 

Go and speak with a doctor in electronic if you dont beleive me. He's going to tell you the same. The cooler your CPU, the better the performance.

 

Period.

 

I do not need internet or any proof for that, I studied electronics, specifically computers.

 

You learn that in your first years btw. I'm not being arrogant, what is arrogant is telling things when you're not sure about it.

 

 

 

 

 

 

 

Lot of informatician in there, and not enough people studied electronic to really understand that Overclocking a processor do not boost your power as much as you think. Otherway all company would devellop processor already overclocked, what do you think ?

 

Overclocking is only to make things sell more.

 

 

 

What is true is that on a long term basis, overclocked processor will lose performance faster.

 

In the end, your E6600 2.4 (like mine) that you overclock to 3.0GHz will effectively run 3.0GHz couple month and after that the surplus of electricity needed to do it will damage your processor AND your motherboard. It will be functionnal again, but effectivelly, even if it show in windows and the bios that you are running to 3.0 GHZ, if you test your CPU with the proper equipement you will see that it may be even down to less than it was BEFORE you overclock it.

 

Have fun ruining your 300$ processor guys. Thinkings you're running fast while you are not !

 

Only amateur and kids do overclock their processor like that.

 

 

When you need to overclock, it's time for you to change your CPU. Period. If the CPU was designed at fisrt to be overclocked.

Really? I have several certificates as well, been working in the field of computers (almost everything to do with them, mostly diagnosis and repair and building machines for everything from servers for companies to full blown PC gaming rigs for consumers) and I have NEVER seen any literature suggesting that a safe and stable overclock (staying well within specs for the hardware) has any effect on the life of a unit. Unless of course you are planning on keeping the same piece of hardware for many many years, in which case I would guess you were building a business PC in which one would question the mentality of an overclock in the first place.

 

Temperature having an effect on performance? Are you high or just ignorant? As long as the processor is in good working order and well within safe temperature parameters, a 3ghz processor is a 3ghz processor. Makes no difference if its at 20c or 40c. What temperature does effect is efficiency, a cooler processor will use less power, not become faster or slower.

 

Companies don't sell overclocked processors not because its "not safe" but because its a business. They make more selling an E8500 or 8600 than they do an 8400. Wouldn't be a very wise business decision to set the E8400 to 3.3ghz to match the E8600 when one cost significantly more than the other.

 

"After a couple of months and all that surplus electricity needed to do it will damage your hardware" - NOT TRUE. Sorry but your wrong. Peroid. But since you seem to think yourself all knowing, feel free to prove me wrong.

 

"Only amateur and kids do overclock their processor like that" - Ok we'll us amateur's and kids will keep enjoying our hardware at there fullest potential for as long as it takes them to become very obsolete and us to upgrade to newer technology, while you milk your system for the next 8 years.

 

Idiocy is strong with this one.

Well I'm sorry if you fail to understand that overclocking your processor bring more energy to all your components that are not made to accept that for most of them. I've built myself electronics chips and I know how they work.

 

Yes a bright person would try to keep his cpu as long as he can. Buying for the sake of buying is stupid. You're probably thoses who make prices go higher. Anyway I do not care if you dont believe, stay in your dream of benchmark result and virtual performance.

 

As I said, tested with proper equipement you will find out that your cpu is yes overclocked and improving some performance, while decreasing others particulary the components of your motherboard. However you may have a good motherboard that allow overclocking, but in the long term the result will be the same, decrease in performance and operability all for a short gain in performance.

 

In the end, light overclocking has minimal consequence, but generally people overclock more than 200 mhz.

 

You could also use logic to understand that if a processor was designed to be a 3.0ghz the company that sell it would have an interest in promoting it. There is a reason why a processor is not garanted after someone overclock it.

 

[ManOWar2]

I do not need internet or any proof for that

You and George Bush don't need evidence.

Dude I'm sorry if I cant copy paste my books. I would also need to translate them. That would be insane.

 

I will not make an internet research for you neither.

 

Comparing me to georges bush becomes irrelevant and show your lack or skills to answer properly to what I posted.

[Vicetopia]

 

You could also use logic to understand that if a processor was designed to be a 3.0ghz the company that sell it would have an interest in promoting it. There is a reason why a processor is not garanted after someone overclock it.

They do.

The Q6600, Q6700, and QX6850 are identical, for example.

The only differences are the multipliers. The Q6600 is stock 2.4, Q6700 stock 2.66, and QX6850 stock 3.0.

 

This is called "binning".

They built the series around the QX6850, and then forcefully lower the multiplier to sell the chips for cheaper. It's an easier way for them to profit.

 

Which is why you can overclock a stock Q6600 to 3 GHz by using the default voltage setting.

Anyway, like most electronic components, all processors have spec sheets:

 

http://processorfinder.intel.com/details.aspx?sSpec=SL9UM

 

Which also include voltage limitations.

[ManOWar2]

Guys, ManoWar2 and TruXter are intentionally posting BS because they love the drama and attention. Its called ego and pain body feeding. It sucks to see someone intentionally and forcefully post flat out false information but they are going to do it anyway. Read ManoWar2's English and his education level is clear. Just ignore their BS please.

Assuming that english is my native langage is also a bit stupid. Not everyones speak english as a native langage. In fact most of people on this planet do not.

 

I'm sure you cant do any better in a second langage.

[Jer1cho]

I have had my processor running at a 1Ghz overclock for over 2 years... It's still fine. Probably because it's running at a 1.25Voltage. So i don't see what the problem is...

[ManOWar2]

You could also use logic to understand that if a processor was designed to be a 3.0ghz the company that sell it would have an interest in promoting it. There is a reason why a processor is not garanted after someone overclock it.

They do.

The Q6600, Q6700, and QX6850 are identical, for example.

The only differences are the multipliers. The Q6600 is stock 2.4, Q6700 stock 2.66, and QX6850 stock 3.0.

 

This is called "binning".

They built the series around the QX6850, and then forcefully lower the multiplier to sell the chips for cheaper. It's an easier way for them to profit.

 

Which is why you can overclock a stock Q6600 to 3 GHz by using the default voltage setting.

Anyway, like most electronic components, all processors have spec sheets:

 

http://processorfinder.intel.com/details.aspx?sSpec=SL9UM

 

Which also include voltage limitations.

You cannot only use thoses basic informations to determine the nature of the chips inside the processor. Of course they are almost the same, but are made differently, that's why they dont call it overclocked version but they use another brand name. You will also notice that all thoses chips have different core stepping version wich is mainly about the power usage, a part that play a crucial part in your chip performance.

 

 

Again, go to the nearest university that has an electronic lab or devellopement and ask some chips programmers about it. I'm pretty sure you will find all your answers.

 

 

[Vicetopia]

 

You cannot only use thoses basic informations to determine the nature of the chips inside the processor. Of course they are almost the same, but are made differently, that's why they dont call it overclocked version but they use another brand name. You will also notice that all thoses chips have different core stepping version wich is mainly about the power usage, a part that play a crucial part in your chip performance.

 

 

Again, go to the nearest university that has an electronic lab or devellopement and ask some chips programmers about it. I'm pretty sure you will find all your answers.

Sorry, but yes, they're all the same except for the multiplier.

Although the G0 stepping Q6600 came out later, the Q6700 and QX6850 are all G0 stepping.

 

I don't need to go to my nearest university electronics lab, since I already go there daily. I suppose I'd be talking to myself... And agreeing.

[ManOWar2]

I have had my processor running at a 1Ghz overclock for over 2 years... It's still fine. Probably because it's running at a 1.25Voltage. So i don't see what the problem is...

Well if you think about changing your system soon there is no problem.

 

There is also no garantee that a component will fail !!! I've just said that generally the result in overclocking is not has high as people think it is and it may affect component and/or reduce performance of other component.

 

That is mainly due because of the softwares we use today and how they are designed to recognize wich CPU is in the motherboard. (software=OS)

 

Having set this one to 1.25 volt is also a good thing if you overclock, of course. That's some kind of way to turn around your OS.

[ManOWar2]

 

You cannot only use thoses basic informations to determine the nature of the chips inside the processor. Of course they are almost the same, but are made differently, that's why they dont call it overclocked version but they use another brand name. You will also notice that all thoses chips have different core stepping version wich is mainly about the power usage, a part that play a crucial part in your chip performance.

 

 

Again, go to the nearest university that has an electronic lab or devellopement and ask some chips programmers about it. I'm pretty sure you will find all your answers.

Sorry, but yes, they're all the same except for the multiplier.

Although the G0 stepping Q6600 came out later, the Q6700 and QX6850 are all G0 stepping.

 

I don't need to go to my nearest university electronics lab, since I already go there daily. I suppose I'd be talking to myself... And agreeing.

Ok then I dont understand why you're trying to prove something useless. I could also say that most of processor are all the same since they all acomplish the same role. You also understand that core stepping is not only about that they show on their boxes. I guess I dont use the good word or may be confuse about what you are refering too, but well I dont care. I just dont like being insulted by young ignorants that never finished school.

 

I also never said that overclocking is impossible or dangerous. It is not always safe and the result are exagerated.

 

a Q6600 3.0 Ghz will never run faster than a Q7600 3.0 Ghz

 

The Q6700 for example, compared to the Q6600 is easier to overclock and will drain less power. So no it is not the same chipset.

 

My partner is building an artificial intelligence right now in his Polytechnics school. I'm willing to trust him when he is telling me things about overclocking I tried to explain. Specially when I got some class in electronic myself and I've seen how electricity work in a chip and most important, the small difference that can exist in two chips that can change the nature of the chip in question.

 

Well if overclocking make you feel good, then go ahead. but for most of people it is not recommended as most of people do not have certifications in electronics or computing engineering. I'm not going anyfurther in this debate neither

[Vicetopia]

 

The Q6700 for example, compared to the Q6600 is easier to overclock and will drain less power. So no it is not the same chipset.

Q6600 has multiplier of 9x, Q6700 has multiplier of 10x.

As long as the Q6600 and Q6700 are running at speeds where they share the same multi/bus, they will be identical. The only differences being the properties of the individual chips. Not every processor is the same of course. They will be "identical" in the same way that one Q6600 is "identical" to another Q6600. Zero differences.

 

 

Specially when I got some class in electronic myself and I've seen how electricity work in a chip and most important, the small difference that can exist in two chips that can change the nature of the chip in question.

We're not talking about how electricity works in a chip, we're talking about the chips themselves.

You might think you know something about electronics, but you clearly know very little about microprocessors. And I already told you, that difference is why you are wrong. I already showed you the spec sheets, they're certified to operate within those voltage limits by Intel themselves. If you have problems with that, then call Intel and tell them about how electronics work. I'm sure they'd enjoy listening to you.

[eagle016]

If you have a Q6600 and a good video card I'd say you're fine. I have both of those and I can run the game fine with all high settings (except for medium on textures) with good fps even when I have explosions going on all over Star Junction (this is with a modified 'popcycles.dat' file where I get about 2-3 times the amount of cars you normally would).

[ManOWar2]

 

The Q6700 for example, compared to the Q6600 is easier to overclock and will drain less power. So no it is not the same chipset.

Q6600 has multiplier of 9x, Q6700 has multiplier of 10x.

As long as the Q6600 and Q6700 are running at speeds where they share the same multi/bus, they will be identical. The only differences being the properties of the individual chips. Not every processor is the same of course. They will be "identical" in the same way that one Q6600 is "identical" to another Q6600. Zero differences.

 

 

Specially when I got some class in electronic myself and I've seen how electricity work in a chip and most important, the small difference that can exist in two chips that can change the nature of the chip in question.

We're not talking about how electricity works in a chip, we're talking about the chips themselves.

You might think you know something about electronics, but you clearly know very little about microprocessors. And I already told you, that difference is why you are wrong. I already showed you the spec sheets, they're certified to operate within those voltage limits by Intel themselves. If you have problems with that, then call Intel and tell them about how electronics work. I'm sure they'd enjoy listening to you.

I might give you that one about thoses microprocessor. I'm not telling I'm a doctor nor a master in this, but I know what I know. They are almost identical. But if you go further in details, you will notice the q6700 do not use the same ammount of energy mainly in watt and therefore it has a consequence on your system.

 

I wont try to explain you theory about this as I am not a master in this. I'll let someone else who want to do it, do it.

 

Just for fun, even someguy tried to explain it on tigerdirect comments to people arguing about this. He was also being called ignorant and pricks because of this. I have no desire to come to that point.

 

 

Peace

Edited December 9, 2008 by ManOWar2

[dirty-dog]

a Q6600 3.0 Ghz will never run faster than a Q7600 3.0 Ghz

 

The Q6700 for example, compared to the Q6600 is easier to overclock and will drain less power. So no it is not the same chipset.

Not the same chipset? Eh?

 

Vicetopia is correct in what he said. Assuming G0 stepping, the Q6600, Q6700 and QX6850 are the same chip, just with a different multi (in the case of the QX6850, it is fully unlocked).

 

Not much of your post makes sense or demonstrates much of an understanding of the issue.

[Galdere]

The Q6600, Q6700 and QX6850 are made from the same materials in the same production.

 

However, making CPUs is a hit and miss thing. Many will have to be thrown away and those that meet the standard are tested and branded according to what they can withstand, hence the different clock speeds for cheaper processors with a lower multiplier but are otherwise created in the same way.

 

AMD will even sell their partly faulty CPUs as "triple" core phenoms, that are actually quad cores, with one of the cores that has failed the testing and so is disabled.

 

CPUs are made to last at least ten years. If you overclock by about 10% this will be within the allowance the manufacturer has tested for. If you exceed this you will shorten the lifespan of it. You can counter this somewhat with better cooling, but it will still be the case, and it will likely still last for many years of prolonged use.

 

If you overclock you also risk affecting the stabilty of your system. This could take a while to occur, but again it will often never be a problem if you are sensible in how you do it and how long you expect to use your computer.

 

So in this clash of much misinformation, there are correct sentiments from both sides. To add my opinion for anyone confused, overclocking by as much as a third is no big deal these days, particulary Intel chips which allow a great deal more with minimal risk. But remember there is a risk, and don't rely on forum posts as to what they are with this, do your own research.

Edited December 9, 2008 by Galdere

[cobzz]

 

 

If your CPU is 30c at 3ghz is it faster than if it is 40c and 3ghz?

I have to say yes to that.

 

if all is good and no lag then well it does not matter.

but if all is kinda good, then when you get into

high intensity moments you will see what the heat does to performance.

 

 

everything , even mechanical items work better when colder

 

the whole part of overclocking is making sure you cool it well.

otherwise performance will stink bad, if not break.

 

we are going to get screamed at by the mods

for carrying on a tech topic in the wrong forum section

they moved us to here

yesterday.

NO.

 

A processor operating at a specific frequency always equals a specific amount of instructions per second. I.E. Performance. Heat does NOT change time, or else, well, people near the equator would be in slow motion. HOWEVER, heat DOES change the resistance of a processor; the hotter; the more resistance. So a 30 degree Q6600 may overclock slightly further than a 70 degree processor. I can run tests if you wish, showing no correlation between temperature and performance.

 

Intel CPU's have referance temperatures - Tjunction (Tj.) and Tjunction MAX (Tj.max). Tj. Max is where your processor goes bye bye. Tj. is the temperature at which the CPU will automatically slow itself down to prevent damage. Tj.Max on the Q6600 (G0 stepping, it's LOWER on a B3 stepping, find out which yours is) is 100 degrees, Tj. is about 75 degrees. You can go right up to around 75 degrees with no problems, much over it and it WILL become slower as the clock speed will SLOW ITSELF down. To conclude; there is a differance between running hot, and OVER heating. You cannot kill the chip by high temperature as it will scale itself back and shut itself off as it climbs above Tj. Make sure Intel Thermal Monitor 2 AND PECI are enabled in BIOS.

 

Once you get to higher frequencies you need to bump up the voltage. This is what kills the chip and is the main thing that causes heat, it's not necessarily frequency. The maximum voltage for a newer G0 Q6600 chip is 1.5 volts, you typically will not need even close untill you get to around 3.6ghz; which you cannot get without awesome cooling. CPU's have something called a VID. It stands for Voltage Identification, or something like that. It's what intel recommends for the chip to run at stock speed. Most chips are different. I have an absolute worst case Q6600 with a VID of 1.3250 volts, which means this puts out the most heat and overclocks the worst of all Q6600s.

 

I am running at 1.24 volts, far below default voltage, at 3ghz - a 25% overclock. Temperature wise it's doing damned well, only 62 degrees C under load, which is infact, about the same as STOCK; and far lower than Tj. This is because the default voltage is so high (on all chips), that you can lower it so much while overclocking and end up cooler or at the same temperature. Keep in mind, all chips are differant, what works with me, may not work with you. Keep in mind, however, my CPU is a bad overclocker as indicated by its high VID. Much past 3ghz I need dramatically more voltage to even get to 3.08ghz - not worth it on my stock cooling.

 

The ONLY reason they run these things so slow stock, in my opinion, is for product differentiation. I am running my $300 (AUD) Q6600 identical to a $950 QX6850. Do it, it's worth it. It's so easy. 25% faster for no increase in temperature, and completely stable. I've ran Prime95 for 24 hours, OCCT for 2 hours, and other programmes for hours on end - with no errors. Before you start though, make sure you KNOW, EXACTLY, what you're doing. http://forums.pcper.com/showthread.php?t=441458 Once you know that, read some more. Use OCCT to check if you're stable at a clock. Then learn about tRD ("Performance Level or Static T read"), and memory timings so you can squeeze some more performance out of your CPU for no increase in heat.

Edited December 9, 2008 by cobzz

[dirty-dog]

Many Q6600s can reach the same clock speed as the Q6700 or Q6850. The latter two are preferable only in that you can achieve a high clock speed with a lower FSB than you would need on the Q6600. It is true that all chips are slightly different and overclocking ability will vary, but most G0 Q6600s can overclock by around 50% on air.

[dirty-dog]

 

I am running at 1.24 volts, far below default voltage, at 3ghz - a 25% overclock. Temperature wise it's doing damned well, only 62 degrees C under load, which is infact, about the same as STOCK; and far lower than Tj. This is because the default voltage is so high (on all chips), that you can lower it so much while overclocking and end up cooler or at the same temperature. Keep in mind, all chips are differant, what works with me, may not work with you. Keep in mind, however, my CPU is a bad overclocker as indicated by its high VID. Much past 3ghz I need dramatically more voltage to even get to 3.08ghz - not worth it on my stock cooling.

 

 

To overclock quad CPUs you need to adjust more than just the vcore. If you want to go beyond 3.0GHz or so, you may need to adjust the GTLREF setting for example. If I leave that on default, the most I can get is around 3.1GHz stable. Adjusting it I got to 3.6GHz stable (I didn't try higher).

 

 

The ONLY reason they run these things so slow stock, in my opinion, is for product differentiation. I am running my $300 (AUD) Q6600 identical to a $950 QX6850. Do it, it's worth it. It's so easy. 25% faster for no increase in temperature, and completely stable. I've ran Prime95 for 24 hours, OCCT for 2 hours, and other programmes for hours on end - with no errors. Before you start though, make sure you KNOW, EXACTLY, what you're doing. http://forums.pcper.com/showthread.php?t=441458 Use OCCT for stress testing.

I run at default 2.4GHz. Why? Because it uses quite a lot less power at that speed, especially with speedstep which reduces not just the clock speed but the vcore when the CPU isn't under load. Q6600s use a lot of power so the savings are worthwhile. Also all my games run great with it clocked to that speed so overclocking is rather pointless, especially with the extra heat and power consumption.

[cobzz]

 

a Q6600 3.0 Ghz will never run faster than a Q6700 3.0 Ghz

 

The Q6700 for example, compared to the Q6600 is easier to overclock and will drain less power. So no it is not the same chipset.

A Q6600 has a multiplyer of 9. A Q6600 has a multiplyer of 10. Therefore, a Q6600 will require a 333 FSB to get to 3ghz, a Q6700 will require a 300 FSB to get to 3ghz. Higher FSB = faster. Q6600 will be faster than a Q6700 unless you lower the multiplyer on the Q6700 to 9 and raise the FSB to 333, where the performance will be identical.

 

The Q6600 and Q6700 are made in the same way, on the same lines. They test them, high quality chips become QX6850s, lower quality chips become Q6600's. It's called binning.

 

 

I run at default 2.4GHz. Why? Because it uses quite a lot less power at that speed, especially with speedstep which reduces not just the clock speed but the vcore when the CPU isn't under load. Q6600s use a lot of power so the savings are worthwhile. Also all my games run great with it clocked to that speed so overclocking is rather pointless, especially with the extra heat and power consumption.

I'm running 3ghz with EIST enabled (lowers multiplyer but NOT voltage WHEN overclocked). I am using a similar amount of power as default as I lowered the voltage with no problems. My mobo will go to 400 FSB with no voltage adjustments.

[Chevyboy]

Manofwar your wrong on EVERYTHING you wrote this entire thread. I attempted to put all of your BS to pass and yet you ignored logic and continued to spew random bullsh*t.

 

Thankfully it seems everyone knows your wrong and nothing but an attention grabbing tool, ignorance is bliss I guess.

 

Back on topic, run the benchmark, if your CPU is nearly maxed out then you will benefit from overclocking. If your computer does nothing but email and forum browsing then again overclocking may not be worthwhile.

 

However if you game and own any "high end" processor (especially intel) than overclock it, most of the chips come factory with lower multipliers than the identical "higher end" chips of the same series. You can usually get 400-800Mhz out of almost every chip with no bump in voltage and almost no change in temperature if any at all. And contrary to 1 misinformed know-nothing, it does NOT degrade the life of your processor at all until you start to significantly up the voltage. And contrary to what our friend was spewing earlier, 3Ghz will always be 3Ghz. It will not "become slow" in 2 months time (not even in several years time if properly taken care of, same goes for all computer hardware) and about all that can go wrong from an OC assume you keep voltage within spec's and temps within spec's is it will fail stability test and need to be readjusted.

 

My old athlon 3800 was overclocked by 40% since I bought it about 4 years ago... still going just as strong as when it was new. *According to manofwar this is not possible*

Edited December 9, 2008 by Chevyboy

[dirty-dog]

A Q6600 has a multiplyer of 9. A Q6600 has a multiplyer of 10. Therefore, a Q6600 will require a 333 FSB to get to 3ghz, a Q6700 will require a 300 FSB to get to 3ghz. Higher FSB = faster.

 

 

Not necessarily. There have been benchmarks done on the same chip at a given clock speed at both 266 and 333 FSB, and it made no difference. Try it yourself. Higher FSB doesn't have the same impact on Core2 systems that it did on P4 systems for example.

 

 

I'm running 3ghz with EIST enabled (lowers multiplyer but NOT voltage WHEN overclocked). I am using a similar amount of power as default as I lowered the voltage with no problems. My mobo will go to 400 FSB with no voltage adjustments.

Indeed, it doesn't lower the voltage so you aren't making any power savings. You should disable EIST when you overclock for stability anyway.

[JohnnyK77]

 

Your idle temp seems high.

Yeah, it is definitely a bit on the high side. I didn't even mention my 250mm case fan on the side panel in the first post.

I assume my issue is the thermal paste (used the stock paste on the CPU cooler). Also, I did have to increase the voltage for my OC to be stable, which could also have a part in the higher temps.

As long as it doesn't go much over 50 (I think highest I ever registered was 52° under full load) I am fine though.

 

Do you know about the E8x00 series having a different temperature variant and thus, many programs read the temp incorrectly?

Yeah, I was aware of that. I also confirmed my temp readings with both RealTemp and SpeedFan.

 

Also, when you first put your cooler on, it helps to run the temp high for a minute, then power off the system and let it cool all the way down, and do that a few times and the paste will bond better for doing that, apparently.  Not sure about that but I read it somewhere, forget where now.

That hardly makes a difference with thermal paste AFAIK; it is a must when using pads though, you have to "burn" those in.

[Chevyboy]

 

A Q6600 has a multiplyer of 9. A Q6600 has a multiplyer of 10. Therefore, a Q6600 will require a 333 FSB to get to 3ghz, a Q6700 will require a 300 FSB to get to 3ghz. Higher FSB = faster.

 

 

Not necessarily. There have been benchmarks done on the same chip at a given clock speed at both 266 and 333 FSB, and it made no difference. Try it yourself. Higher FSB doesn't have the same impact on Core2 systems that it did on P4 systems for example.

 

 

I'm running 3ghz with EIST enabled (lowers multiplyer but NOT voltage WHEN overclocked). I am using a similar amount of power as default as I lowered the voltage with no problems. My mobo will go to 400 FSB with no voltage adjustments.

Indeed, it doesn't lower the voltage so you aren't making any power savings. You should disable EIST when you overclock for stability anyway.

In addition to what you already said, a high FSB CAN (not usually anymore, that seemed to end in the days of P4) produce artificial benchmark results higher than a same clocked processor with a lower FPS frequency, but it was proven many many times that a higher FSB frequency has NO EFFECT WHATSOEVER on real world applications and overall performance when comparing 2 same clocked CPU's.

 

The only real reason these days to go with a lower multiplier-higher FSB is because it CAN lead to an overclock that requires less vcore on SOME processors.

 

So as dirtydog implied, most of the fsb - multiplier stuff you read or hear about ended in the P4 days and most if not all of it does NOT apply to modern processors.

 

 

 

 

@JonnyK77 - Just make sure whatever program you use has the TJmax set to 100c, this was confirmed by intel a few months ago to be the official TJmax for the E8** series. Most programs you can edit or just get the newest coretemp. Again tho nothing wrong with high 60's low 70's. Prime95 pushes my E8500 (Zalman 9500A cooler) to about 52-55c depending on the temp outside but intelburntest pushes it to 65c. Anything over 75-80 and the chip will take action but anything under 75 and you should be fine. That said the cooler the better. *And if you use AS5 (arctic silver 5 paste) it is suggested to let it heat then cool several times over 24 - 200 hours to fully seat and bond.

Edited December 9, 2008 by Chevyboy

[cobzz]

Correct - actual FSB performance changes with memory divider settings and the strap... etc..

 

1. 266 strap, 9 x 356, 4:5. 3 ghz.

 

will be faster than

 

2. 266 strap 8 x 400, 1:1. 3 ghz.

 

As the memory is faster and the MCH read delay is lower. That being said, if you know what you're doing, know what tweaks to use, then 400 FSB will always be faster than 356 FSB, and this includes real world performance.

 

Another example is. 266 FSB will be about the same as 333 FSB. However, if you tighten up MCH READ DELAY, and lower the 333 STRAP to 266, then 333 will always be faster. I did try it - you try it with Everest Ultimate Benchmarks, or any other CPU bound programme. There is a relatively small differance - but it still does give you more performance.

 

http://www.anandtech.com/mb/showdoc.aspx?i=3208&p=1

 

When overclocked EIST only changes multiplyer, which does save some power. It will not touch voltage, thus stability cannot be harmed by the multiplyer going from 9 to 6.

Edited December 9, 2008 by cobzz

[Rokz0r]

Yeah in synthetic benchmarks latency and bandwidth IS faster but in real world apps it is much more dependent on your GPU and minimal gains in games with higher clocked cpu and ram...The case where you will get a big benefit is more so with multi GPU system as that's what requires more horse power to drive a nice tri sli setup that would actually benefit from more cpu hogging - a lot of the games these days arent very cpu limited like in the old days - with core 2 etc you have enough juice to even power some nice sli setups these days - but any more graphics horsepower (quad 4870x2's etc) it's almost pointless if you dont have a decent cpu power (corei7) especially with todays graphics powerhouses - hopefully the intro of dx11 will make some real time ray tracing capabilities more of a reality (apparently intel and microsoft are working closely with the intro of dx 11+larabee) then everything changes once again and you will be cpu limited because general purpose cpu cores are the best for ray tracing atm, this is why Nvidia are trying not to be left behind and are singing about CUDA.