Real meaning behind Turbo and TDP

Since apparently there are people here that still do not understand what intel means with TDP and turbo speed I thought I would make this thread to inform people, especially reseller should know this when offering systems that are designed for specific TDP's.

I was quite shocked to find out that even resellers didn't even know about this and implied that for instance the 8700 non K was a CPU that would run on 65W at full speed.

This is important because it is not. This is intels way of making their processors seem more effiecent than they really are.

Now lets get right into it, when intel talks about max turbo frequency, they mean the max turbo frequency that the CPU can run on, the turbo is higher when you run only 1 core, than all cores, this makes it very different from what it actually runs on on all cores.

Let's take the 8700K for instance, the CPU is specified at 4.7ghz max turbo clock, but in reality without overclock, it is running at 4.3ghz when all cores are used, this is a 400mhz difference, which is quite a lot.

Now lets talk about TDP, when intel refers to TDP, they actually talk about the average power the processors dissapates in base frequency.


Now lets see how much a 8700K vs a 8700 non k would actually run at full stock turbo speeds.(both run at 4.3ghz when all cores are used)


As you can see, both CPU's use equal amount of Watts (the 8700 is slightly higher due to slighlty worse silicon). Both CPU's are well above their specified TDP limits, especially the 8700 non k. Sure this is the hardest stresstest you can do, so obviously this is the maxium stress you can put om them, but both will need around 100-110 watts to run at turbo speeds without throttle on games that require some CPU power or rendering.

I hope you guys learned something, and please don't ever again refer the 8700 non k ever again as a 65W CPU, stop it.

 

By intels definition of TDP, the average for the i7-8700 running at 3.2GHz would be around 65W, right? and the K at 3.7GHz at more or less 95W?

 

A discussion of turbo without a mention of its TIME LIMITED NATURE is negligent. Turboboost has had TIMERS since its first implementation that dictate what power limit is active for how long. The flowchart goes like this:

e.g. 65W TDP CPU; 81.25W turbo power limit, 28 second turbo limit;

Idle
*heavy load started*
CPU runs at 81.25W with whatever multipliers (subject to maximum allowed) that permits.
*28 seconds passes*
CPU downclocks/de-turbos itself down to 65W power limit.
*runs at 65W power limit*
load ends
idle load causes a turbo timer reset

*MANY* Coffee Lake reviewers - and now you too - were stung by an Asus motherboard feature (Multi Core Enhancement) - and other OEM boards with similar features - that HACKED the time limit to be PERMANENT but that is a MOTHERBOARD FEATURE RUNNING THE CPUS BEYOND INTEL'S TURBO SPECIFICATIONS and not a CPU FEATURE. On the K series and Zx70 chipsets these limits are NOT FIXED and can be altered by the user in this way too. What the ASUS MCE feature does is that it does this on NON-K CPUs on Zx70 AS WELL and what you see is a result of edited turbo power + timer PLUS an unlocked chipset

The *exact same mechanism* that was used on my 2012 P170EM's i7-3720QM to run well beyond CPU spec, at up to 3.9Ghz, 75W load, with an unlocked Prema BIOS. The TDP and internal specs are not altered; the CPU is running at unlimited turbo.

The only wrong thing going on here is that Asus didn't publicise having this on BY DEFAULT well enough!

ComputerBase where you got that power draw chart from used ...
*drumroll*
an Asus Z370 board and a Gigabyte with similar function

This video explains it:

Others caught are shown @ 2:13

The 8700 is NOT using less power because it's "slightly worse silicon", its MCE maximum is 4.6GHz, cf/ the 8700K's 4.7.

This MCE feature is present in my Z97+ Mitx board I figured out what was going on with a couple googles after I noticed the i7-5775C was running at 3.7 all core all the time, I really don't know how this was missed, it's not a new thing...

What you are complaining about is an on-by-default but toggleable intentional OEM motherboard OVERCLOCKING HACK making a CPU run hotter ***OMGWTFBBQ***

And now you can go apologise to all those you've insulted now that you know what is really behind an 8700 running beyond 65W.

@Papusan add this to your list
@Meaker@Sager @XMG @clevo-extreme this lack of understanding is why the guy blasted you in that P950 thread earlier on

 

The image posted by Danishblunt, but deleted, shows this in action.....where the turbo speeds of 4.3GHz were reached at 81.25W, then settled down to 65W, although running a bit below the base clock for the CPU (3.0GHz instead of 3.2GHz)
So I don't really understand where the issue is at, or the point trying to be made.... pretty standard stuff as far I know.

 

My all means go to town if you want to explain TDP and TurboBoost 2.0, but don't under any circumstances do this with the pretence that I've been working in the industry for the last 15+ years with a cardboard box on my head.

If you could just:

1/ stop miss-representing things that I do say
2/ stop telling people I said things that I didn't
3/ stop projecting your hate for Intel's marketing department onto me for no justifiable reason whatsoever
4/ stop confusing the hell out of everyone on here by suggesting that I and some other company representatives are idiots and ignorant - which I find both insulting and hilarious at the same time
5/ accept that you have no understanding of how the industry refers to component and platform matching, or how pre- mass production and sample testing works with ODMs
6/ even when I correct your false accusations to me, you still stick to the same blunt statement with nothing to back it up (i.e. the golden chassis sample argument for instance)
7/ stop making me waste my time on a Sunday convincing other nbr members that I'm not an ignorant idiot that works for an incompetent company.

If you could just manage that please it would be pretty peachy.

Cheers.

 

I get sad when I see useless posts like this one in OP. No wonder people get the wrong info and and taking their choices on wrong promises when all they got is like this.. This will never ever stop. I'm just waiting for the next one

 

maybe he didn't even read what he posted

comedy gold LMAO so your whole agenda now shift towards sh!lling for AMD? or whats going on LOL

- no, prime 95 isnt the heaviest bench, though it is already unrealistically heavy for consumer workload.
- p95 has different versions and has different settings for different workloads/temps
- intel's TDP though not openly stated, it is for SSE legacy workload under restriction with power limit set in bios, and when AVX is involved it goes out the window
- intel can list 65w or even 15w if they wanted to, they practically own the market, its consumers falling for it not doing enough research.
- since you like a CPU that is called BGA 99.9% of the chips are restricted by intel's set TDP so you don't have to worry, intel will make sure you dont go over those value if you buy their BGA LMAO. up to 3 layers of protection!! intel's own CPU, OEM bios, physical heatsink, come in a complete package so you pay more than what desktop user pay while getting less performance.