AMD Adds Clarification to "Max Boost Clocks" on Ryzen Product Pages

[Tsing]

The complaints of Ryzen 3000 chips not hitting their advertised boost speeds have evidently gotten severe enough to force AMD to clearly define what "Max Boost Clock" actually means on its product pages. There is now an "MROM-001" link that offers the following clarification: "Max Boost Clock is the maximum single-core frequency at which the processor is capable of operating under nominal conditions."

That language isn't sitting well with owners who feel they were lied to by the marketing department. Some allege that the wording (e.g., "nominal conditions") is purposefully vague so AMD has an easier time dodging potential lawsuits.

…AMD processors spread their boost frequency progressively across cores during a multi-threaded workload that scales across all cores. At any given time, only one of the cores is awarded the highest boost clock, and while the other cores too get boosted beyond the nominal clock-speeds, they are in slight decrements of 25-50 MHz.

 

[Grimlakin]

I think people are over reacting in a HUGE degree about this.

'I'm not getting all the GHZ on all he cores at all the times!'

AMD: 'Correct'

'I should be getting all the GHZ on all the cores at all the times!'

AMD: 'Incorrect, you will get all the GHZ (up to the numbers we have given) on any of the cores at any of the times under load provided cooling and power are both in good supply. It will not be all the cores at all the times or even some of the times. While all cores may see an speed increase under demand it will not be to peak values by design.'

'That's unfair I expected far better automated automatic overclocking!'

AMD '...'

 

[Dan_D]

Hrmm.......if only you could have read a product review before you went out and bought the CPU.

 

[Grimlakin]

Hrmm.......if only you could have read a product review before you went out and bought the CPU.

Exactly. It's like people are claiming they went in with their eyes closed when this is the way it's always been. I simply don't see the case here.

 

[Brent_Justice]

Marketing speak doesn't change the performance of your CPU. If you were happy with the performance of your CPU when you purchased it, nothing has changed. I just don't understand people sometimes.

 

[Dan_D]

The way boost clocks have worked for years can pretty much be applied to the Ryzen 3000 series with one exception. While Intel CPU's have been capable of reaching those clocks on any core, it wouldn't do it on more than one at a time on it's own. Ryzen 3000 may not be able to reach those clocks on every core, but again, it can do so on one. The net result is more or less the same.

The biggest difference is that Intel being able to achieve that clock on every core allows you to potentially overclock all of the cores to that speed at once if you can keep it cool and don't give a crap about power consumption. Which frankly, most of us don't. On the AMD side, you can't do that as all core overclocking gets you to about 4.3GHz on a 3900X and that's it. However, overclocking, is, was, and never has been guaranteed.

I don't get while people are bitching. I get why people are mad about not hitting the boost clocks that are advertised, even though the difference was small enough that it was no impact to performance. However, no one knew that until we got the data.

 

[Grimlakin]

It's OVER CLOCKING. When has OVER CLOCKING ever been a guaranteed thing? Be glad they have the formula to do this to the point that you don't need to bother with it other than putting in good freaking components.

People complaining about crap like this will be why companies like AMD STOP WORRYING ABOUT OVERCLOCKING and just put a good stable base clock on the chip then say.. "If you over clock it and it breaks that's you're problem. Bye Felicia."

 

[Mjz_5]

Marketing speak doesn't change the performance of your CPU. If you were happy with the performance of your CPU when you purchased it, nothing has changed. I just don't understand people sometimes.

Immature people ruin everything. Lol

 

[Brian_B]

In "average person's" defense...

if it says 4.6 on the box, I expect my computer to be able to hit 4.6

Sure, you need adequate cooling, and adequate power, and so on and so forth. I think most reasonable people who are buying bare CPUs are installing typical solutions which I would consider adequate. Not all, there are some idiots - but most are pretty vanilla and standard.

I wouldn't expect 4.6 all the time, or even most of the time. But my understanding is that a lot of people are never seeing it. I see a lot getting close... and 4.55 is close, but it's still not 4.6, unless you count rounding. And there are reports of some folks not even getting that close. - seems like I recall Dan_D only hit 4.4 - 4.3 on his first review here.

Now, I expect this all gets resolved out - BIOS updates, AEGIA updates, driver updates, etc. I'd say Dan_D's 4.57 that he saw in his BIOS update review would be close enough for me - I don't think it's a big deal right now. But I can see where some of the static is coming from.

 

[noko]

I've seen random 4.63ghz basic desktop, something like CineBench single 4.52 and that is it. What is funny is I get lower boost clocks if I select AutoOC in Ryzen Master - when everything is default it works best. All core when using multithreading in Cinebench is around 4.1ghz. Which to push all core OCing to 4.2 to 4.3ghz - have not tried with increase voltages etc and loss of higher Boost clock does seem utterly pointless. For the most part I could care less what MHz it is as long as the performance in the end is top notch or what I expected.

 

[Dan_D]

That's been my experience as well. When you start pulling the PPT, EDC and TDC values from the motherboard or the user, performance suffers. Oddly, the values used by the CPU seem to offer the best results right now.