Monday, September 2nd 2019

Der8auer: Only Small Percentage of 3rd Gen Ryzen CPUs Hit Their Advertised Speeds

World famous overclocker Der8auer published his survey of boost clocks found on 3rd generation Ryzen CPUs. Collecting data from almost 3,000 entries from people around the world, he has found out that a majority of the 3000 series Ryzen CPUs are not hitting their advertised boost speeds. Perhaps one of the worst results from the entire survey are for the 12-core Ryzen 9 3900X, for which only 5.6% of entries reported have managed to reach the boost speeds AMD advertises. However, the situation is better for lower-end SKUs, with about half of the Ryzen 5 3600 results showing that their CPU is boosting correctly and within advertised numbers.

Der8auer carefully selected the results that went into the survey, where he discarded any numbers that used either specialized cooling like water chillers, Precision Boost Overdrive - PBO or the results which were submitted by "fanboys" who wanted to game the result. Testing was purely scientific using Cinebench R15 and clock speeds were recorded using HWinfo (which got recommendation from AMD), so he could get as precise data as possible.
Der8auer comments that he still recommends Ryzen 3000 series CPUs, as they present a good value and have good performance to back. He just finds it very odd that AMD didn't specify what you need to reach the advertised boost speeds.

If you would like to see the more in depth testing, here is the English version of the video:

Add your own comment

248 Comments on Der8auer: Only Small Percentage of 3rd Gen Ryzen CPUs Hit Their Advertised Speeds

#26
Mephis
TheLostSwede, post: 4108679, member: 3382"
Care to share the rest of your hardware, as well as UEFI version? Without that, it's hard to give any suggestions.
But this is a perfect example of the problem. AMD didn't say that these speeds were dependant on specific circumstances. They advertise it as the max boost clock, with no disclaimers. This leads customers to believe that their processor should boost to that speed.

We can debate what "max" means all day long, but we shouldn't have to. They should have been really clear about this. Either in the launch presentation or at the very least on the product page on their own website.

Source:
AMD 3900x
Posted on Reply
#27
Jism
bug, post: 4108683, member: 157434"
No, that is not what boost has meant until Zen2. If that were the case, you'd see manufacturers using "1THz boost" on their boxes, because, well, it's "up to" and one of their engineers has seen one CPU reaching those speeds when no one else was looking.

And the debate is not useless, like the debate about GTX 970's VRAM wasn't useless at the time: it doesn't mean users get screwed, but it sends a clear message this is a practice we could do without.
I am on the side of AMD this time. Not because i have a AMD product but because i understand the way the boost algorithm works and how AMD implemented this into their products. Not all CPUs will reach that advertised boost clock "up to". It has a few reasons. I think the best way to describe it is being the lucky one having a piece of silicon that is capable of doing so. If the 5% reaches the advertised (Up to, lol) boost clocks then they are the happy few.

AMD doesnt do binning. They implemented a technique into their CPUs that already bins for them and you as a consumer. What more do you want? This eliminates technically the need for manual overclocking and why Sillicon lottery simply gives up for attempting to bin on AMD cpu´s in the first place.
Posted on Reply
#28
Mephis
Jism, post: 4108687, member: 91255"
I am on the side of AMD this time. Not because i have a AMD product but because i understand the way the boost algorithm works and how AMD implemented this into their products. Not all CPUs will reach that advertised boost clock "up to". It has a few reasons. I think the best way to describe it is being the lucky one having a piece of silicon that is capable of doing so. If the 5% reaches the advertised (Up to, lol) boost clocks then they are the happy few.

AMD doesnt do binning. They implemented a technique into their CPUs that already bins for them and you as a consumer. What more do you want?
You may understand the boost algorithm, but do you really think the average consumer does? It's not on the consumer to figure it out. Especially when AMD could put a footnote on the website with the product listing.

And if we are going to go with the whole "up to" thing, then why not put 5ghz? I'm sure some people can reach that.
Posted on Reply
#29
bug
Jism, post: 4108687, member: 91255"
I am on the side of AMD this time. Not because i have a AMD product but because i understand the way the boost algorithm works and how AMD implemented this into their products. Not all CPUs will reach that advertised boost clock "up to". It has a few reasons. I think the best way to describe it is being the lucky one having a piece of silicon that is capable of doing so. If the 5% reaches the advertised (Up to, lol) boost clocks then they are the happy few.

AMD doesnt do binning. They implemented a technique into their CPUs that already bins for them and you as a consumer. What more do you want? This eliminates technically the need for manual overclocking and why Sillicon lottery simply gives up for attempting to bin on AMD cpu´s in the first place.
If all the above is true, AMD should have used numbers that virtually every CPU can reach. It's that simple.
Posted on Reply
#30
Jism
2 important things are listed on ANY amd ryzen SKU:

"Base Clock"
"Boost Clock"

Where base clock is the absolute guaranteed base clock. You get what you pay for in any given workload in combination with a AMD certified cooler or third party cooler which meets the required TDP. Where Boost clock is advertised as, "Up to". Where up to simply depends on the given cooler you are using, the motherboard (VRM), the given workload and the FIT of the silicon where it is able to reach that clock and be perfectly stable. Oh wait do you want the max boost clock now and have a unstable single core? What does the majority of consumers dont understand about that?

There is a 5% of CPUs getting what its advertised for. Silicon lotterly, lol.

These kind of fud is just generating clicks, useless newsworthy articles. If you just dive a little bit more down into the technical standpoint then you and others understand.

Even a 12 core AMD monster does work perfectly fine on a cheap 50$ 350 board, even in overclocked condition with a 4 phase VRM. This is due to the requirements AMD demands on any motherboard vendor putting out boards. Your not getting a repeat of the FX era where a 8350 would throttle because board vendors would skimp out on VRM design.

I think the overall quality of the first ryzen untill now is just very very good.
Posted on Reply
#31
EarthDog
Jism, post: 4108696, member: 91255"
2 important things are listed on ANY amd ryzen SKU:

"Base Clock"
"Boost Clock"

Where base clock is the absolute guaranteed base clock. You get what you pay for in any given workload in combination with a AMD certified cooler or third party cooler which meets the required TDP. Where Boost clock is advertised as, "Up to". Where up to simply depends on the given cooler you are using, the motherboard (VRM), the given workload and the FIT of the silicon where it is able to reach that clock and be perfectly stable. Oh wait do you want the max boost clock now and have a unstable single core? What does the majority of consumers dont understand about that?

There is a 5% of CPUs getting what its advertised for. Silicon lotterly, lol.

These kind of fud is just generating clicks, useless newsworthy articles. If you just dive a little bit more down into the technical standpoint then you and others understand.

Even a 12 core AMD monster does work perfectly fine on a cheap 50$ 350 board, even in overclocked condition with a 4 phase VRM. This is due to the requirements AMD demands on any motherboard vendor putting out boards. Your not getting a repeat of the FX era where a 8350 would throttle because board vendors would skimp out on VRM design.

I think the overall quality of the first ryzen untill now is just very very good.
Yada. We know.

What you may ave missed from the other thread is people with MORE THAN CAPABLE motherboards, coolers, etc, are not able to reach the boost clock under any circumstances.

AMD didnt put any requirements in on previous chipsets for Ryzen 3. That is what x570 is for. If you'll note, their vrms are, overall, more robust than x470, b450, and especially the b320. Didnt the 3900x throttle on the b350 board testing here? You're playing with fire throwing a 3900x+ on b350...in some cases, this certainly is like fx and the vrm situation on low end boards. The biggest difference is these cant get past their own two feet overclocking as they are maxed out already.

Also, amd does do binning. How the heck do you think you have so many skus??? The bottom line is more than just a small percentage of cpus with their proper conditions should hit boost clock.

Maybe they would be better served doing it like nvidia and listing a MINIMUM boost clock in stead of something reserved for magic unicorns.
Posted on Reply
#32
springs113
My 3700x, all cores at some point or another hits 4.4ghz. I haven't tried all core manual oc but I am very satisfied with my results so far. This is an issue but I think that there's a lot at play here. Too many variables to just blame the main person(AMD).
Posted on Reply
#33
Metroid
This is the reason why I bought the midrange r5 3600, I was very certain would be the only one to hit adv boost speed. Let's face here, even with binned chips high ends chips are not getting the boost speed amd said it would. So until amd fixes the mess, I will keep only buying midrange cpus from them.

Basically on the video, 50% of 3600 hit the adv boost, 10% of 3600x hit the adv boost, 15% of 3700x hit the adv boost, 27% of 3800x hit adv speed boost, 6% of 3900x hit the adv speed boost.
Posted on Reply
#34
Vya Domus
Polls are worthless, unfortunately.

newtekie1, post: 4108658, member: 20670"
When will AMD's marketing team learn from their mistakes?
When everyone else will do the same. As far as they're concerned they just do what everyone else does, Nividia sold FE Pascal cards that were advertised to have a certain boost clock but they all ran pretty much at the base clock after 5 minutes under load.

But I posed this question on his video and I will do the same here, how should AMD advertise these ? If they say they have 100mhz lower boost clocks but some people get more than that you'll get the same problem, people will feel it's unfair that some get more.

Metroid, post: 4108713, member: 178915"
even with binned chips high ends chips are not getting the boost speed amd said it would.
AMD didn't said that they are binned, that's just a speculation on our side.
Posted on Reply
#35
EarthDog
Metroid, post: 4108713, member: 178915"
This is the reason why I bought the midrange r5 3600, I was very certain would be the only one to hit adv boost speed. Let's face here, even with binned chips high ends chips are not getting the boost speed amd said it would. So until amd fixes the mess, I will keep buying midrange cpus from them.
Cause you bought on a BS premise? I've got a 3700x that doesnt hit boost clocks...isnt that midrange being 3 products down from the top and two from the bottom?
Posted on Reply
#36
bug
Vya Domus, post: 4108714, member: 169281"
But I posed this question on his video and I will do the same here, how should AMD advertise these ? If they say they have 100mhz lower boost clocks but some people get more than that you'll get the same problem, people will feel it's unfair that some get more.
Nope, that's the way boost always worked. You get what's on the box, everything else is a bonus/luck of the draw.
Posted on Reply
#37
Jism
EarthDog, post: 4108703, member: 79836"
What you may ave missed from the other thread is people with MORE THAN CAPABLE motherboards, coolers, etc, are not able to reach the boost clock under any circumstances.
Uh yeah but you miss a crucial point completely. The boost is based on not just thermals or VRMs but also workload and power requirements needed for that specific workload. Every AMD Ryzen CPU has a FIT thing build inside that monitors the CPU´s condition and silicons reliability. You cannot, and a boost state will not, harm the CPU while going into boost. A single core can only use an X amount of current which is programmed in the FIT. If a cpu core is getting unstable at a certain boost clock in combination with a voltage it will lower it down, OR if a CPU is using too much current at a given workload in combination with a boost clock it will lower it down. Its as simple as that. If you would disable the FIT completely (as you can do with CCX hacking these days) then you might archieve the advertised clock, but risking to degrade the CPU.

Simular as going with a manual overclock. You risk losing warranty in a nutshell. AMD would be in complete shit if their boost clocks would be harming their CPU´s on long term based. Everybody is creating FUD these days while not much really take the effort to understand the boost algorithm from a technical standpoint.

No the majority is not a tech expert when buying a CPU, but AMD is right. A base clock plus a up to boost clock. Marketing wise its perfectly fine. Do you want to add a disclaimer then that the maximum boost state is based upon silicon lottery if you take good cooling / VRM and all that into account?

Derbauer is stupid. He tested on exotic cooling what XFR would do in a nutshell. From that standpoint he already knew that publishing this article (apart from generating fud) is kind of useless and that a Up to boostclock matches the technical standpoint.
Posted on Reply
#38
EarthDog
Jism, post: 4108725, member: 91255"
Uh yeah but you miss a crucial point completely. The boost is based on not just thermals or VRMs but also workload and power requirements needed for that specific workload. Every AMD Ryzen CPU has a FIT thing build inside that monitors the CPU´s condition and silicons reliability. You cannot, and a boost state will not, harm the CPU while going into boost. A single core can only use an X amount of current which is programmed in the FIT. If a cpu core is getting unstable at a certain boost clock in combination with a voltage it will lower it down, OR if a CPU is using too much current at a given workload in combination with a boost clock it will lower it down. Its as simple as that. If you would disable the FIT completely (as you can do with CCX hacking these days) then you might archieve the advertised clock, but risking to degrade the CPU.

Simular as going with a manual overclock. You risk losing warranty in a nutshell. AMD would be in complete shit if their boost clocks would be harming their CPU´s on long term based. Everybody is creating FUD these days while not much really take the effort to understand the boost algorithm from a technical standpoint.

No the majority is not a tech expert when buying a CPU, but AMD is right. A base clock plus a up to boost clock. Marketing wise its perfectly fine. Do you want to add a disclaimer then that the maximum boost state is based upon silicon lottery if you take good cooling / VRM and all that into account?

Derbauer is stupid. He tested on exotic cooling what XFR would do in a nutshell. From that standpoint he already knew that publishing this article (apart from generating fud) is kind of useless and that a Up to boostclock matches the technical standpoint.
I didnt miss a thing.

From that other thread, we tried everything. Hell, I even left hwinfo up for over a day and still it never hit that clock. Max speeds were always less. I have the cooling, the boards (6), the bios' (couple on each), the loads (from idle to spi 1m and just desktop - word, excel, chrome) neeeeeever. :(

What were disagreeing on is that you think it's ok to be marketed as X which is something few, a small minority no doubt, can achieve. I have all the ingredients to make it, yet I cant get there. That's BOLOGNA.
Posted on Reply
#39
Mephis
Vya Domus, post: 4108714, member: 169281"
Pools are worthless, unfortunately.



But I posed this question on his video and I will do the same here, how should AMD advertise these ? If they say they have 100mhz lower boost clocks but some people get more than that you'll get the same problem, people will feel it's unfair that some get more.
Are you really saying that there would be a problem if the rates them lower and some chips achieved higher than advertised clocks? That is the way it has always been. That would be the exact same has it has always been. Chips run at advertised speeds and some better samples are able to be clocked higher.
Posted on Reply
#40
Nkd
Look I am not debating about people not hitting their boost clocks. I am 25mhz below so you could count me in. But if you are after stats he should know better given how respectable he is. That is a very skewed result.

there are so many missing factors in there it’s hardly factual.
Like throwing away things he didn’t even ask for in the survey. Also what were the users using to monitor the boost clocks?

I do think thus has a lot to do with bios tweaks. Which AMD needs to sort out with mobo manufacturers.
Posted on Reply
#41
Jism
EarthDog, post: 4108726, member: 79836"
I didnt miss a thing. From that other thread, we tried everything. Hell, I even left hwinfo up for over a day and still, never.
From my experience, the boost thing starts to back down the moment the CPU or core hits 60 degrees. If it doesnt pass 60 degrees and its not hitting the max, your not having the lucky peace of silicon that could boost up to the advertised clocks. There is a 5% of users who actually do boost up to advertised clocks. We have seen a user hitting even 4.55GHz single core on a 2700x while the majority doesnt pass 4.35GHz. Do you want to start hold AMD accountable for that one user that boosts up to 4.55Ghz on single core now ?

And apart of that, who gives a flying shit really if your CPU does 4.4Ghz and not 4.45GHz lol. Like your going to miss out on that one single thread 50Mhz difference.
Posted on Reply
#42
Metroid
EarthDog, post: 4108716, member: 79836"
Cause you bought on a BS premise? I've got a 3700x that doesnt hit boost clocks...isnt that midrange being 3 products down from the top and two from the bottom?
I updated my post, internet connection here was lost for few minutes, down below is what the video stated, so is not bs as we can see, the 3600 is the one with the highest chance to hit adv speed boost.

"Basically on the video, 50% of 3600 hit the adv boost, 10% of 3600x hit the adv boost, 15% of 3700x hit the adv boost, 27% of 3800x hit adv speed boost, 6% of 3900x hit the adv speed boost. "
Posted on Reply
#43
Mephis
Jism, post: 4108725, member: 91255"
No the majority is not a tech expert when buying a CPU, but AMD is right. A base clock plus a up to boost clock. Marketing wise its perfectly fine. Do you want to add a disclaimer then that the maximum boost state is based upon silicon lottery if you take good cooling / VRM and all that into account?
Yes! That is exactly what they should have done. Especially, since they had to of known this would be an issue. If they didn't know, then they completely failed in the testing phase.
Posted on Reply
#44
Vya Domus
Mephis, post: 4108728, member: 186806"
Are you really saying that there would be a problem if the rates them lower and some chips achieved higher than advertised clocks?
So this works one way only huh ? You do realize this is literally just placebo, right ? No matter what you do some people will get less, some more, no matter which way you set the balance. This is bound to cause complaints no matter what. Arguing about this is simply a waste of time, the only thing that would make this somewhat fair is that everyone stops stating "boost clocks" on their products and only claim a certain base clock.
Posted on Reply
#45
EarthDog
Jism, post: 4108730, member: 91255"
From my experience, the boost thing starts to back down the moment the CPU or core hits 60 degrees. If it doesnt pass 60 degrees and its not hitting the max, your not having the lucky peace of silicon that could boost up to the advertised clocks. There is a 5% of users who actually do boost up to advertised clocks. We have seen a user hitting even 4.55GHz single core on a 2700x while the majority doesnt pass 4.35GHz. Do you want to start hold AMD accountable for that one user that boosts up to 4.55Ghz on single core now ?

And apart of that, who gives a flying shit really if your CPU does 4.4Ghz and not 4.45GHz lol. Like your going to miss out on that one single thread 50Mhz difference.
see edits.

You're missing the point.

The stock cooler cant keep these cpus under 60c. The type of cooling that needs to happen with isnt common.

I have no idea the relevance of one dude hitting a clock on sen+ cpu. This is about zen 2. That said, even you did say zen 2, it's still a poor talking point. If its advertised... a small minority hitting it isnt right...regardless if its 50mhz or 100mhz. If it says xx on the box and I have the conditions they set forth met, a MAJORITY should be hitting those clocks, not a handful with luck.
Posted on Reply
#46
Jism
Mephis, post: 4108733, member: 186806"
Yes! That is exactly what they should have done. Especially, since they had to of known this would be an issue. If they didn't know, then they completely failed in the testing phase.
You and others are wasting time on this subject. This whole subject is created to generate more FUD in favor of clicks, views and hits. Intel does offers CPU´s simular. A base clock and a boost clock. Always within thermals and current, a up to value. Do you want to start hitting on Intel too now?

EarthDog, post: 4108737, member: 79836"
see edits.

You're missing the point.

The stock cooler cant keep these cpus under 60c. The type of cooling that needs to happen with isnt common.

I have no idea the relevance of one dude hitting a clock on sen+ cpu. This is about zen 2. That said, even you did say zen 2, it's still a poor talking point. If its advertised... a small minority hitting it isnt right...regardless if its 50mhz or 100mhz. If it says xx on the box and I have the conditions they set forth met, a MAJORITY should be hitting those clocks, not a handful with luck.
The Zen+ works with simular tech, and pretty much the logic to that can be found in the Zen2 family as well. The XFR is a extended feature now on Zen2 family. But that doesnt mean the behaviour of Zen+ could not be applied to Zen2. My CPU on stock stops holding boost clocks once it reaches 60 degrees and starts to back down from 4.2Ghz on a IBT workload to 4.1GHz, once it starts to pass another thermal threshold it backs down to 4.05Ghz and so on. Its all temp related really. Remember my SKU is a 3.7Ghz base model. Getting 4.2GHz out of it on the long run is 500Mhz above what AMD sold me.

Once i adjusted my fan setup from going from 3 to 6 fans on a 360mm rad and lowering the voltages supplied, it start to hold the boost in this case on 4.2GHz all core IBT. Over a small period of time the water starts to heat up and brings the CPU to 4.1Ghz eventually now. But my CPU worked all this time for a longer period on 4.2GHz compared to stock settings. I am getting a overal better result then compared to a stock setting. My single core boost stick constant on 4.35GHz now with a sling once in a while to 4.2Ghz. Its doing exactly as AMD intendede the technology build inside the CPU.

So yeah this is a very stupid way of measuring how the avg 3x series hold their boost clocks. It takes some settings in bios first, it takes some measures in cooling as well, and know that a feature like FIS might be the culprit to archieving higher boost clocks. You have to dive into your CPU first to understand why its not reaching those advertised speeds. If your 50mhz under it, you belong to the few lucky. But you could, stick some time just like i did, and see if you can get a higher clock or a longer boost state.

That there is variation on different motherboard vendors, you have to find out as well.

As for the Ryzen CPU stock cooler, really its sufficient to keep the thing under a typical workload under a certain temperature. Not when you start bashing with CB or IBT for that matter on the long run. Afterall watercooling is always the better sollution for faster heat transfer compared to a heatsink.
Posted on Reply
#47
Mephis
Vya Domus, post: 4108734, member: 169281"
So this works one way only huh ? You do realize this is literally placebo, right ? No matter what you do some people will get less, some more, no matter which way you set the balance. Arguing about this is simply a waste of time, the only thing that would make this somewhat fair is that everyone stops stating "boost clocks" on their products and only claim a certain base clock.
Except every other manufacturer that has done boosts before this has had products that at least can achieve those boost clock. they may not have been at those clocks forever, but we are talking about a large percentage of a product that can't reach the boost clocks to begin with. I'd Intel or Nvidia did this, it would be just as crappy.
Posted on Reply
#48
RH92
That's proper journalism there from Der8auer kudos for his work !

I totally agree with his conclusion ... YES Ryzen 3000 are still great CPUs ... NO AMD had absolutely no need to advertise false boost clocks knowing that a BIG majority of their CPUs can't hit those advertised clocks ! Had some other companies made the same all hypocrites here going through crazy lengths to defend AMD would had roasted those other companies so yeah ...........

Their marketing team is playing dangerous games nowadays and they need to stop this !
Posted on Reply
#49
EarthDog
Mephis, post: 4108739, member: 186806"
If Nvidia did this, it would be just as crappy.
Funny because big bad NVIDIA does it RIGHT listing a GUARANTEED boost clock (minimum) and not listing a maximum!!!!!

Look at gpuz... the boost clock. And 100% of these will achieve that clock and typically 100-200MHz more when using boost (assuming no power virus running against it).

Jism, post: 4108738, member: 91255"
You and others are wasting time on this subject. This whole subject is created to generate more FUD in favor of clicks, views and hits. Intel does offers CPU´s simular. A base clock and a boost clock. Always within thermals and current, a up to value. Do you want to start hitting on Intel too now?
Dont act dense... you know Intel actually hits these boost clocks on all cpus and that is a piss poor argument, right?
Posted on Reply
#50
Vya Domus
Mephis, post: 4108739, member: 186806"
Except every other manufacturer that has done boosts before this has had products that at least can achieve those boost clock.
Then everything is alright, because AMD's CPUs can reach their boost clock too, under *asterisk conditions just like every other manufacturer stipulates. You bought one that doesn't or you feel like it should hit these clocks more often? You are free to return it, claim RMA, whatever.

Mephis, post: 4108739, member: 186806"
but we are talking about a large percentage of a product that can't reach the boost clocks to begin with.
No we aren't, we are speculating that may be the case, based on some polls.
Posted on Reply
Add your own comment