AMD Updates Ryzen Product Pages to Elaborate on "Max Boost Clocks"

[btarunr]

AMD over the weekend updated the product-pages of its Ryzen processors on the company website to be very specific about what they mean by "Max Boost Clocks," that are advertised almost as extensively as the processor's main nominal clock-speeds. AMD describes it has "the maximum single-core frequency at which the processor is capable of operating under nominal conditions." We read into this as the highest boost-clock given to one of the cores on the processor.

If you've been reading the "clock-frequency and boost analysis" charts in our processor reviews, you'll know that 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. The graph below is from our Ryzen 7 3700X review. The second graph below is from our Core i9-9900K review, which too shows only one of the cores getting the max boost frequency, and the remaining cores getting lower boost clocks, although the graph looks flatter.



View at TechPowerUp Main Site

 

[Vayra86]

The 9900K graph 'looks flatter'... yeah, if you disregard the actual numbers they've used. Intel's turbo is an ancient piece of junk compared to XFR.

Intel 'looks flatter' ... with a 500mhz frequency gap versus AMD's 150 mhz.

So what they're saying then is that it's pretty much a BS number that you might never see, as it's a "nominal" number. There's not even a percentage figure of what the chance is to reach that "nominal" frequency. I would go as far as to call that false advertising or selling a product under a false pretense, as not a single customer would've read boost speeds as something you might never achieve.
I have a feeling this is going to backfire badly.

Good point, because if they advertise 4.4 Ghz and hit 4375, that is really under

I'll still take XFR any day of the week though...

 

[TheLostSwede]

So what they're saying then is that it's pretty much a BS number that you might never see, as it's a "nominal" number. There's not even a percentage figure of what the chance is to reach that "nominal" frequency. I would go as far as to call that false advertising or selling a product under a false pretense, as not a single customer would've read boost speeds as something you might never achieve.
I have a feeling this is going to backfire badly.
Also, what's nominal conditions?

 

[Lionheart]

Not to be a negative nancy but false advertising anyone orrrr...... am I overreacting?

 

[biffzinker]

Also, what's nominal conditions?

Temperature, voltage, power consumption, and motherboard limits would be my best guess.

 

[TheLostSwede]

I'm curious, why is the other note about EPYC, when it's on the Ryzen pages?

Temperature, voltage, power consumption, and motherboard limits would be my best guess.

Yes, but maybe AMD should provide an example then? As they can't just pull something like this, without providing a reference point, no?
At this point, I really feel AMD cheated me out of extra money for nothing.
This doesn't even start to take their PBO video into account, where they claim we could expect an extra 200MHz boost as long as the motherboard could deliver more power.

 

[Bwaze]

They should state what are the conditions for achieving boost clock.

A lot of people don't achieve anything near with single core load (like Cinebench single core, SuperPi single core...) , and processor only boosts close to stated boost speed with very light single core loads for a very short periods of time.

Should the processors be able to sustain boost clock with single core load, proper motherboard and enough cooling? Or does hovering near for a milisecond in idle count as achieving stated frequency, don't look at frequencies during load?

Will AMD also disclose what happened to the PBO + Overclock, and rasing boost clocks up to 200 MHz with good enough motherboard and cooling? Or is that a forgotten marketing push that wasn't really based on anything achievable?

 

[Patriot]

They should state what are the conditions for achieving boost clock.

A lot of people don't achieve anything near with single core load (like Cinebench single core, SuperPi single core...) , and processor only boosts close to stated boost speed with very light single core loads for a very short periods of time.

Should the processors be able to sustain boost clock with single core load, proper motherboard and enough cooling? Or does hovering near for a milisecond in idle count as achieving stated frequency, don't look at frequencies during load?

Will AMD also disclose what happened to the PBO + Overclock, and rasing boost clocks up to 200 MHz with good enough motherboard and cooling? Or is that a forgotten marketing push that wasn't really based on anything achievable?

Yeah, they wrote 4.7Ghz and talked about how XFR was as many cores as can boost vs intel is single core.

 

[hzy4]

the graph for 3700X doesnt seems right, in CB R20 all core I can see 3,92Ghz max.
the graph shows 16 threads 4,2Ghz-4,25ghz.
if this would be right you would see much more higher scores at auto clock.
when I manually OC to 4,2Ghz all core the scores are higher then auto.

 

[Space Lynx]

I'm wondering if I should have done i5-9600k and 5.1ghz all 6 cores no downclocking now... a full 1ghz faster than AMD... :/ too late now, I can't return my CPU I don't think

 

[SystemMechanic]

Most people dont even know that only a few cores can even reach the max clocks in Ryzen 3000 series haha. In my 3900x The second chaplet cores dont even hit 4.5ghz, let alone 4.6. They mostly topout at 4.3 to 4.4Ghz.



Only CCX0 and CCX1 are good enough to reach 4.6Ghz.

Also funny that people thought a 12 core cpu would reach 4.6Ghz on all cores...

 

[Space Lynx]

It's just a bit shady marketing imo

 

[TheLostSwede]

the graph for 3700X doesnt seems right, in CB R20 all core I can see 3,92Ghz max.
the graph shows 16 threads 4,2Ghz-4,25ghz.
if this would be right you would see much more higher scores at auto clock.
when I manually OC to 4,2Ghz all core the scores are higher then auto.

This is the problem, for some people it works, for others, it doesn't.
The clocks are all over the place and this is why people are upset...

Most people dont even know that only a few cores can even reach the max clocks in Ryzen 3000 series haha. In my 3900x The second chaplet cores dont even hit 4.5ghz, let alone 4.6. They mostly topout at 4.3 to 4.4Ghz.

Only CCX0 and CCX1 are good enough to reach 4.6Ghz.

Also funny that people thought a 12 core cpu would reach 4.6Ghz on all cores...

I never expected that, as it was boost speed. However, I do think people expected a more consistent behaviour, as it's really quite different from CPU to CPU, not by 25-50MHz, but by 200-300MHz in some cases.

 

[Bwaze]

Some people are now waiting for a special bios that would enable their (for example) 3900X go from 4.4 GHz that they are achieving now in single core loads to 4.6, with possibility to overclock it to 4.8 GHz.

That would indeed be some Fine Wine.

AMD has to say sooner or later what are the real specifications of Ryzen 3000 processors, and if those promisses were simply false advertising.

 

[biffzinker]

I'm wondering if I should have done i5-9600k and 5.1ghz all 6 cores no downclocking now... a full 1ghz faster than AMD... :/ too late now, I can't return my CPU I don't think

Don't know if it matters but I did consider the i5-9600K with an overclock since I spend most of my time playing games. Outside of games the 3600 out runs the i5-9600K with an overclock though.

 

[Space Lynx]

Don't know if it matters but I did consider the i5-9600K with an overclock since I spend most of my time playing games. Outside of games the 3600 out runs the i5-9600K with an overclock though.

all I do is game as well. I just checked and can't return my 3700x. oh well, no big deal. I am still impressed with what AMD has accomplished, my 5700 XT only gets 10 fps slower at 1440p in sekiro shadows die twice vs RTX 2080 SUPER... at almost half the price... AMD is not getting enough love at the moment so I am fine with supporting them this round.

 

[cucker tarlson]

percentage-wise,at full load 3700x is much closer to max boost clock (96%) than 9900k (89%).

 

[Space Lynx]

percentage-wise,at full load 3700x is much closer to max boost clock (96%) than 9900k (89%).

3700x is still a great buy imo, I don't regret it. I especially love the native Linux support, I don't have to install drivers on my clean install of latest version of Linux Mint. also Intel is getting security issues (major ones) almost on a bi-weekly basis, before its all said and done intel will be patched so many damn times the 3700x will surpass even 9900k in gaming. also security is important for me, so i'd rather have 5-10 fps slower in games than risk it.

 

[cucker tarlson]

3700x is still a great buy imo, I don't regret it. I especially love the native Linux support, I don't have to install drivers on my clean install of latest version of Linux Mint. also Intel is getting security issues (major ones) almost on a bi-weekly basis, before its all said and done intel will be patched so many damn times the 3700x will surpass even 9900k in gaming. also security is important for me, so i'd rather have 5-10 fps slower in games than risk it.

well if you're concerned about security (or not) 3700x is a good choice ragardless,but can we please stop the patched pefrormance penalty misinformation or at least limit it to red trolls ? I mean we've heard it a thousand times before zen 2 launch but in the end old ass 8700k still leaves any zen 2 behind.go on gamersnexus review and see how big the gap actually is.

it's nice xfr works better than turbo boost,but still zen 2 overclocks like a turd compared to intel's ridiculed 14++++++++++++

 

[hzy4]

This is the problem, for some people it works, for others, it doesn't.
The clocks are all over the place and this is why people are upset...

But the CB R20 scores and performance across some popular benchmarks is on point with the major reviews out there. I was checking the clock speeds in the newest Ryzen Master, they did not go higher then 4,05Ghz all core so its a mystery for me. This is pointing to a measurement error of the software.

 

[Space Lynx]

well if you're concerned about security (or not) 3700x is a good choice ragardless,but can we please stop the patched pefrormance penalty misinformation or at least limit it to red trolls ? I mean we've heard it a thousand times before zen 2 launch but in the end old ass 8700k still leaves any zen 2 behind.go on gamersnexus review and see how big the gap actually is.

it's nice xfr works better than turbo boost,but still zen 2 overclocks like a turd compared to intel's ridiculed 14++++++++++++

While performance was only affected 1% or so of meltdown and spectre patches, keep in mind there are still loads of unpatched security issues with Intel, the last major Intel security issue was reported last week and is just as dangerous as Meltdown/Spectre was, and that's not including the several others... or Intel recommending users disable HT... it's just too much imo...

even the games a 9900k beats 3700x in at 1440p, its only by 5 fps or so a lot of the times. and at 4k its tied.

 

[bug]

Also, what's nominal conditions?

I have a feeling defining that would require AMD to spell out you also have to get lucky enough that your CPU will include a "good bin" CCX.
That doesn't detract from Zen's value per se, but it looks like it unnecessarily exposes AMD to a class action lawsuit (we all know there's a lawyer somewhere with some time to kill).

 

[cucker tarlson]

While performance was only affected 1% or so of meltdown and spectre patches, keep in mind there are still loads of unpatched security issues with Intel, the last major Intel security issue was reported last week and is just as dangerous as Meltdown/Spectre was, and that's not including the several others... or Intel recommending users disable HT... it's just too much imo...

even the games a 9900k beats 3700x in at 1440p, its only by 5 fps or so a lot of the times. and at 4k its tied.

well specter and meltdown weren't dangerous for a home user to begin with,same as the one with ht.but if you feel like it's too much for you then it's fine.

when you compare mostly gpu-bound scenarios then even a 9400f will match 3700x so those kind of comparisons aren't really going anywhere.

 

[londiste]

even the games a 9900k beats 3700x in at 1440p, its only by 5 fps or so a lot of the times. and at 4k its tied.

At 4K (and largely at 1440p) you are fine with 3600 or 9400F.

 

[biffzinker]

but still zen 2 overclocks like a turd compared to intel's ridiculed 14++++++++++++

Zen2 does overclock like a turd but the IPC improvement from Zen/Zen+ neutralizes the clockspeed difference going from Intel's high clocked Skylake core design unless you move to the Sunny Cove cores re-design. I have a feeling Sunny Cove isn't expected to clock as high on 10nm+ as Skylake on 14+++. Zen2 clocks the same as Sunny Cove.