Monday, August 12th 2019

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

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.
Source: squidz0rz (Reddit)
Add your own comment

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

#101
hzy4
Mine 3700x runs 1.23v at 4.1Ghz but could not get stable at 1.35v at 4.3Ghz, so I figured to run auto until the drivers get tweaked and more info about safe voltage for 24/7 is available
Posted on Reply
#102
Xpect
londiste, post: 4096847, member: 169790"
.... for a CPU where spec says 4400...
Well, no, Spec says 3600 MHz. Boost is Just an added Bonus if the CPU evaluates it can go higher.
Posted on Reply
#103
EarthDog
Xpect, post: 4098238, member: 172697"
Well, no, Spec says 3600 MHz. Boost is Just an added Bonus if the CPU evaluates it can go higher.
That isn't how it works though. If it says on a package that it should boost to X.XX GHz, then it should boost to that value. Many are finding it does not, even with 'nominal' conditions.
Posted on Reply
#104
TheMadDutchDude
Mine break into the mid 70s or so, but I’m on custom water.

The only time I’ll ever see the full boost is if I get it under the cold stuff, which won’t happen.

My buddy in the UK has one that does 4.35 all core at 1.3v.
Posted on Reply
#105
Xpect
EarthDog, post: 4098240, member: 79836"
That isn't how it works though. If it says on a package that it should boost to X.XX GHz, then it should boost to that value. Many are finding it does not, even with 'nominal' conditions.
It says on every IT Shop here in Germany "Clock 3.6GHz, Boost up to 4.4GHz"

In my opinion and if you read the exact wording for those products, Boost Clock is Just a Bonus. Advertising, Sure, but always as "up to". Guaranteed is only Stock Clock.
Posted on Reply
#106
EarthDog
Xpect, post: 4098397, member: 172697"
It says on every IT Shop here in Germany "Clock 3.6GHz, Boost up to 4.4GHz"

In my opinion and if you read the exact wording for those products, Boost Clock is Just a Bonus. Advertising, Sure, but always as "up to". Guaranteed is only Stock Clock.
It says MAXIMUM on their product page and box. And people with 'nominal' setups (as AMD described it themselves) are not able to reach it. It really seems that people who are reaching the max boost, there isn't a rhyme or reason to it. Great motheboard and cooling... still doesn't reach it... MEH parts, a couple have... it makes no sense. THAT is mainly my issue with it.

Why can't I reach my 3700X's max? I have a 3x120mm AIO, tried it across 6 motherboards ranging from entry level to mid-range and still, it's short.
Posted on Reply
#107
hzy4
All x570 MB are garantied to support the 12 core chips so its unlikekly tha mb fault that 3700x cannot hit 4.4 boost
Posted on Reply
#108
Nordic
TheMadDutchDude, post: 4098169, member: 185664"
I haven’t seen mine hit it’s max frequency, and it’s under custom water cooling.
What waterblock and radiator do you use? I have been thinking about installing my custom watercooling loop again. I was getting temps of 40c on my 4.8ghz@1.5v 2500k and my 7970 overclocked to 1425mhz. I was thinking that the same setup could potentially cool my 3900x to about the same temperature.
Posted on Reply
#109
EarthDog
hzy4, post: 4098416, member: 186895"
All x570 MB are garantied to support the 12 core chips so its unlikekly tha mb fault that 3700x cannot hit 4.4 boost
Sure.

I don't think it's a power issue in the first place though.
Nordic, post: 4098458, member: 96457"
I was thinking that the same setup could potentially cool my 3900x to about the same temperature.
Heh... not going to happen, bud. 40C at full load, even on a 2500K (but at 4.8 GHz and 1.5V) sounds impossible or your load was something jenky like gaming and not a proper stress test.
Posted on Reply
#110
TheMadDutchDude
Nordic, post: 4098458, member: 96457"
What waterblock and radiator do you use? I have been thinking about installing my custom watercooling loop again. I was getting temps of 40c on my 4.8ghz@1.5v 2500k and my 7970 overclocked to 1425mhz. I was thinking that the same setup could potentially cool my 3900x to about the same temperature.
I’m running an EK Supremacy EVO, Hardware Labs 360 slim something or other and a 240 EK PE rad.
Posted on Reply
#111
bug
Xpect, post: 4098397, member: 172697"
It says on every IT Shop here in Germany "Clock 3.6GHz, Boost up to 4.4GHz"

In my opinion and if you read the exact wording for those products, Boost Clock is Just a Bonus. Advertising, Sure, but always as "up to". Guaranteed is only Stock Clock.
You have to realize this is an opinion shared only by you and probably a few other people that have never seen a CPU.
Posted on Reply
#112
Nordic
EarthDog, post: 4098464, member: 79836"
Heh... not going to happen, bud. 40C at full load, even on a 2500K (but at 4.8 GHz and 1.5V) sounds impossible or your load was something jenky like gaming and not a proper stress test.
I stress tested it with prime95 and running the gpu full bore. I also was running boinc and folding@home 24/7 then, at least not when not using the computer. 40c max load was what I got.

The same cooling loop with only a 3930k couldn't keep that cpu below 80c when overclocked to 4.6ghz.

I had a monsta 240mm long 80mm thick radiator with 32mm panaflows and 25mm spacers between the fans and the radiator. The fans ran at full speed. Swiftech waterpump and pump with an alphacool full cover gpu waterblock. The radiator is so thick that I used it as the resivor. I have pictures somewhere. I bet I even made a thread about it back then around 2013.

A 3900x doesn't use the amount of power as a 3930k, but does have a smaller node. Since I am running it at stock, and letting it overclock itself I should get some great temps.
Posted on Reply
#113
EarthDog
Nordic, post: 4098557, member: 96457"
I stress tested it with prime95 and running the gpu full bore. I also was running boinc and folding@home 24/7 then, at least not when not using the computer. 40c max load was what I got.

The same cooling loop with only a 3930k couldn't keep that cpu below 80c when overclocked to 4.6ghz.

I had a monsta 240mm long 80mm thick radiator with 32mm panaflows and 25mm spacers between the fans and the radiator. The fans ran at full speed. Swiftech waterpump and pump with an alphacool full cover gpu waterblock. The radiator is so thick that I used it as the resivor. I have pictures somewhere. I bet I even made a thread about it back then around 2013.

A 3900x doesn't use the amount of power as a 3930k, but does have a smaller node. Since I am running it at stock, and letting it overclock itself I should get some great temps.
Sure.

Anyway, there's no chance for these chips to get anywhere close to those temps. Soooooooo many different variables involved, but no way.
Posted on Reply
#114
Nordic
bug, post: 4097215, member: 157434"
Ok, I get it, you're bitter.
Try this: fire up SuperPi (single threaded) and see whether you reach the advertised speeds. If you don't, then yes, AMD has shafted you.
I apologize that it took me this long to get back to you, but here are my results. I reset hwinfo's sensor results right before running superpi, and took the screenshot as soon as it was over. Looks like during this run I peaked at 4400.3mhz on core 5. Core 5 is my fastest core in my first CCX.
Posted on Reply
#115
EarthDog
Nordic, post: 4098955, member: 96457"
Looks like during this run I peaked at 4400.3mhz on core 5. Core 5 is my fastest core in my first CCX.
Doesn't the 3900x boost to 4.6GHz?
Posted on Reply
#116
Nordic
EarthDog, post: 4098965, member: 79836"
Doesn't the 3900x boost to 4.6GHz?
It is supposed to. I have seen my 3900x boost as high as 4525mhz max, and it doesn't sustain that speed for more than a second. I have only gotten to 4525mhz while idle that I have seen. Bug specifically called out superpi so I shared my results under that workload.
Posted on Reply
#117
londiste
Isn't highest on that screenshot 4450.3 MHz on Core 3?
Posted on Reply
#118
Nordic
I didn't see that. The results are better then I thought.
Posted on Reply
#119
EarthDog
londiste, post: 4099053, member: 169790"
Isn't highest on that screenshot 4450.3 MHz on Core 3?
Ohhhh, now only 150 Mhz away from what it should be doing!
Posted on Reply
#120
Nordic
Even if I was getting 4.6ghz, it doesn't sustain those speed very long. I averaged about 4080mhz on my fastest core while running superpi. I reset hwinfos recorded values, immediately clicked start on superpi. As soon as superpi was done, I took the screenshot.

The problem isn't just the inability to reach the boost clock but the inability to sustain a higher clockspeed than 4.2ghz.
Posted on Reply
#121
bug
Nordic, post: 4098955, member: 96457"
I apologize that it took me this long to get back to you, but here are my results. I reset hwinfo's sensor results right before running superpi, and took the screenshot as soon as it was over. Looks like during this run I peaked at 4400.3mhz on core 5. Core 5 is my fastest core in my first CCX.

First of all, no apology needed, I wasn't tasking anyone, I was just offering a way to test boost under the best circumstances.
Second of all, bummer. It seems like AMD really pulled numbers from where the Sun don't shine :(
Posted on Reply
#122
medi01
Turmania, post: 4097706, member: 182201"
I'm an old fashioned simple lad.when I buy something according to the box, I expect it to perform like it. You write 4.5 ghz boost I expect it to do so with all cores.I don't buy an 8 core processor just so 1 core reaches advertised speeds.it's shame and scam.
Well, then so is Intel's CPU (and even more so, as frequency drops much sharper with more threads)
Posted on Reply
#123
Nordic
medi01, post: 4108881, member: 158537"
Well, then so is Intel's CPU (and even more so, as frequency drops much sharper with more threads)
Given proper cooling, I don't think there isn't an intel CPU that won't do the boost frequency on all cores if you set it to do that in the bios. That is technically overclocking, but it does work. AMD's boost is like this magical thing that sometimes boosts the cpu higher than we can overclock it ourselves. Intel's boost is deterministic. AMD's is something else.

EDIT: I think it was debauer who said he is unsure if AMD even knows why their boost is behaving the way it is.
Posted on Reply
#124
medi01
Nordic, post: 4108930, member: 96457"
Given proper cooling, I don't think there isn't an intel CPU that won't do the boost frequency on all cores if you set it to do that in the bios. That is technically overclocking, but it does work. AMD's boost is like this magical thing that sometimes boosts the cpu higher than we can overclock it ourselves. Intel's boost is deterministic. AMD's is something else.

EDIT: I think it was debauer who said he is unsure if AMD even knows why their boost is behaving the way it is.
By default (and without excessive power consumption) AMD boosts to 96% of the advertised max on all cores.
While Intel boosts to 89%.

Cough.
Posted on Reply
#125
R0H1T
Nordic, post: 4108930, member: 96457"
Given proper cooling, I don't think there isn't an intel CPU that won't do the boost frequency on all cores if you set it to do that in the bios. That is technically overclocking, but it does work. AMD's boost is like this magical thing that sometimes boosts the cpu higher than we can overclock it ourselves. Intel's boost is deterministic. AMD's is something else.

EDIT: I think it was debauer who said he is unsure if AMD even knows why their boost is behaving the way it is.
Should be clear why that is, AMD doesn't have the luxury to sell locked chips with 2GHz base clocks as all their chips are unlocked. So while Intel could throw away their less than ideal chips even at uncompetitive prices, as locked variants, in retail or OEM channels. AMD cannot do that, now whether they tried or not is up for debate.

Continuing from the other thread, there is indeed a problem with AMD's advertised or rated boost clocks because it shouldn't be this random. It would greatly helpful if they put a reference (spec) PC with workloads & BIOS+other settings listed so that users could actually see how/when/where these chips clock the best. Now I also mentioned that the sample set from that survey is insignificant, hence drawing any conclusions for the rest of the market is guessing at best/dubious at worst. So yeah it'd be so much better if AMD came out much cleaner than just issuing statements about the fix.

I will again reiterate though that the boost behavior is dependent on a huge number of variables, so it isn't all their fault because they had to keep TDP in check as well as maintain backwards compatibility!
Posted on Reply
Add your own comment