Higher BLCK means higher clock memory m8 for example

200x20=4ghx memory 1600mhz more vcore
191x21=4ghz memory 1533mhz(cant remember the exact feq) less vcore

 

100% no mate theres a test floating about with dd3 ram at 1333mhz and 1600mhz and there was something really stipid between then in games like 1fps...bech marks is another story but day to day no diff

 

I have turbo off because my machine isn't stable with it on.

 

You won't notice one tiny bit of difference between 20x200 and 21x191 appart from in synthetic memory benchmarks and thats it.

 

I am taking a quite irrational approach to turbo.

If the chip runs too hot, uses too much power or draws too much current, it is designed to throttle. It flicks back to non-turbo mode many times a second, spending more and more time on the lower multi as it gets hotter. I believe some boards can stop this throttling. However this makes turbo mode looks rather like bribing people to use speed step. A higher multi, fine. Enforced speed step, absolutely not. I want to be able to choose whether to use speed step or not.

Second, Intel are keen on these 'self overclocking chips', and I believe the i5 do it to an even greater extent. Firstly this is a cop out, "our chip will run at 3.8ghz but only if you use half the cores and don't let it get too hot" is cheating. Secondly, I predict posts in the future to the effect of "my processor is at 4ghz and its cool cos I only ever use word so it never throttles" which will enrage me.

So, I choose not to use it. Firstly to make my small point that I don't approve, and don't want Intel to go down this road. Secondly because I don't want the chip multiplier hopping around all over the place when running close to its limits, there is no way in which this can be good for stability or the lifespan of the chip.

I don't expect people to agree, but I believe in this.

 

I now prefer to o/c with turbo disabled now, for the reasons Jon mentioned above.

Though at 4.2GHz I have it enabled.

Getting 191x21 stable is a lot easier than 200x20 though.

 

I understand your point, but if the cooling is more than sufficient, which in most cases it is then the CPU doesnt get anywhere near its thermal limit and as such doesnt throttle. Anyone using the UD5 has the option of enabling or disabling speedstep, turbo and EIST.
Along the same reasoning, is it sensible to have your OCed CPU running at 4Ghz+ when its idle or you are using WORD for example. I dont think so, in this instance I think its sensible to allow the CPU to reduce its BCLK and Multiplier.
I have run numerous tests using that program you suggested and never once did the CPU throttle while under load and this is because it didnt reach the tjmax.

 

you need more QPI voltage for the higher BCLK, I'm going to try my 200x20 OC with the Skews we have been discussing. Remember your CPU will not throttle until it reaches it tjmax, which I'm sure you dont get anywhere near.

 

It does not only throttle because of temperature. It also throttles if it draws too much current, or if it uses too much power. These last two give me pause for thought.

My computer doesn't really spend time idle, when it's actually stable I run folding@home in the background.

 

Where did you get that info from. Do you have a link?
I tested mine for hours when you mentioned this before and never found one instance of throttling. Most of the d0 chips are still well within their vid range anyway. Surely you need more volts for more current. My 191x21 with everything enabled doesn't get anywhere near the 130w intel spec. At what point does it throttle.

 

Here is a nice little video from Intel explaining how turbo mode works.

http://www.intel.com/technology/product/demos/turboboost/demo.htm?iid=tech_tb+demo

I think I may have posted this on these forums a while ago, but here it is again anyway.

What Jon says is right enough, it isn't just to do with heat that these chips throttle.

Though when I tested my 4.2GHz o/c with Turbo enabled with the i7 Turbo app from unclewebb it never showed any signs of throttling, but for my 4GHz o/c, I still prefer to do it without turbo boost, I kinda get a feeling as though I am cheating when using turbo tbh...

 

Here. This is the relevant Intel whitepaper, I read through it a while back. Aimed at a different audience to Davys video, I personally prefer the technical documents.

I think xtreme found this to be a problem around 1.45V/90degrees, but the error on that is considerable since I only vaguely remember and don't have the thread to hand.

Perhaps a bit lower than 1.45, from post 7 here

Note that there is something in the bios which should disable this function, however I don't know how it is implemented.

 

Thanks guys. I'll have a read of that. I understand what you mean about cheating Davy.

 

Interesting read, but I maintain its a bit irrelevant to us for a couple of reasons,

We are not getting our cpus anywhere near the voltages and temps required to enable throttling.

Most of us do not have the C-state enabled in the bios in the first place.

But like the LLC discussion, I will begin work on my 200x20 overclock forthwith......lol

 

I thought that the the ud5 was unaffected by intels TDP limit with later bios then the f3 and c1e disabled? I had the origional asus p6t deluxe and when using huge amounts of voltage at 21X 200 on a old c0 chip I could see it throtteling downwards in small increments. Thats why i changed to the ud5 and I have never seen the ud5 doing this.

Are you saying that turbo mode will throttle even though cpuz says 21x just because its reached a power consumption limit alone?

 

That's what I'm saying with the c states disabled it doesn't throttle, so it's all academic. Still gives me an excuse to use on the Wife as to why I'm still faffing about.

 

LoL!!

I give mine excuses all the time to!!

@chiLLZ, if you download this tool you can use it to see whether you are suffering from turbo throttling, CPUZ wont tell you if your chip is throttling.