Ryzen 3000 boost clock controversy - der8auer publishes his survey results, not a good look for AMD

Thank you. It's like people have forgotten how to read. This won't even make it to a lawsuit with 'max boost' terminology used.
People are also not realizing that most of the ryzen motherboards, all though set at a bus speed of 100, are actually running at a bus speed of 99. Which means a 3700x for example will boost to 4356, or there about when hitting the boost multiplier of 44.
 
I think my 3700x is still great value, but it also only hits 4375 boost.

4.375GHz is actually pretty much hitting the advertised boost clocks. The reason you aren't seeing 4.4GHz is because the base clock is generally below 100MHz. If you add a slight increase to your base clock you should see 4.4GHz. (Theoretically, as I haven't tested this exactly.)
 
  • Like
Reactions: N4CR
like this
His results are based on over 2000 results sent in, with screenshots and free-form text fields - he put a lot of work into this analysis.

But "my chip does it, so its FUD". Very meaningful counteranalysis, nicely done.

You got it! What's right for me is right for me.

Mathematics knowledge no longer exists.

Bus x multi = clock speed.

Blame the fucking boards.

Der8auer should know this
 
I think it’s also important to note the recent 1.0.0.3ABB AGESA actually lowered boost clocks for many people as well.

As a 3700X owner, I’ve never seen the 4.4 GHz core speed hit on a single core, and it’s not because of a 99MHz bus clock. Pretty frustrating to get away from “evil” Intel and go over to AMD just advertising made up numbers as fact. Sure it’s just a few %, however I’d like to get what I paid for. This is on a $330 motherboard as well.

My previous 4.4GHz 7700K hit boost clocks all day long, every day when needed. My RTX 2070 is exceeding the reported boost clocks probably 99% of the time it’s under load. People expect the same of AMD.
 
That's like complaining about 4gb vram vs 3.5gb vram.


Guys, it's not Intel or nVidia, it's okay.


The NVIDIA GTX 970 went to trial because the ccard was advertised as 64 ROPs and 4GB ram. That 56 ROPs and 3.5GB VRAM meant you were missing 13% of the performance lifetime (or 26% if you went SLI, like many people did).

Of course people started returning the cards, or requesting a price reduction. And Then came the well-deserved lawsuit. Just because games of the day did not challenge the architecture, didn't mean you coulld just gloss over 13% lower bandwidth and VRAM per-card.

The Pentium III 1.13 GHz was not stable, and was rushed to market to compete. There was no lawsuit because it was officially recalled.

We're talking about a boost clock here that's not guaranteed, one that's falling 2% short of the mark. That's hard;y worthy of screaming "Bloody Murder" over this!

The lawsuit over Bulldozer's core architecture has more merit than this bitch fest over 100 mhz.
 
  • Like
Reactions: N4CR
like this
The NVIDIA GTX 970 went to trial because the ccard was advertised as 64 ROPs and 4GB ram. That 56 ROPs and 3.5GB VRAM meant you were missing 13% of the performance lifetime (or 26% if you went SLI, like many people did).

Of course people started returning the cards, or requesting a price reduction. And Then came the well-deserved lawsuit. Just because games of the day did not challenge the architecture, didn't mean you coulld just gloss over 13% lower bandwidth and VRAM per-card.

The Pentium III 1.13 GHz was not stable, and was rushed to market to compete. There was no lawsuit because it was officially recalled.

We're talking about a boost clock here that's not guaranteed, one that's falling 2% short of the mark. That's hard;y worthy of screaming "Bloody Murder" over this!

The 970 had 4GB of VRAM (3.5+0.5 if you want to get technical about how it was allocated). Nvidia deserved the lawsuit, but can people stop spreading misinformation?
 
I'm guessing someone will eventually start a class action for this, especially considering the one that was just resolved after 5 years. Here's a list of class actions filed against team blue for their blunders.
 
The NVIDIA GTX 970 went to trial because the ccard was advertised as 64 ROPs and 4GB ram. That 56 ROPs and 3.5GB VRAM meant you were missing 13% of the performance lifetime (or 26% if you went SLI, like many people did).

Of course people started returning the cards, or requesting a price reduction. And Then came the well-deserved lawsuit. Just because games of the day did not challenge the architecture, didn't mean you coulld just gloss over 13% lower bandwidth and VRAM per-card.

The Pentium III 1.13 GHz was not stable, and was rushed to market to compete. There was no lawsuit because it was officially recalled.

We're talking about a boost clock here that's not guaranteed, one that's falling 2% short of the mark. That's hard;y worthy of screaming "Bloody Murder" over this!

The lawsuit over Bulldozer's core architecture has more merit than this bitch fest over 100 mhz.

99% of this website is millennial crybaby entitled infants. At least that's my conclusion.

Hey start your class actions. You'll feel justified when you get your 7 dollar check in the mail


Yeah justice!

Over a 50mhz clock loss because you dont cool your chip right or your motherboard quartz 10mhz base crystal is vibrating a 9.9mhz instead lmao
 
I'm guessing someone will eventually start a class action for this, especially considering the one that was just resolved after 5 years. Here's a list of class actions filed against team blue for their blunders.

Wouldn't be surprised. Whether or not it goes anywhere though is going to heavily depend on exactly what a court would consider qualifies as "up to".
 
99% of this website is millennial crybaby entitled infants. At least that's my conclusion.

Hey start your class actions. You'll feel justified when you get your 7 dollar check in the mail


Yeah justice!

Over a 50mhz clock loss because you dont cool your chip right or your motherboard quartz 10mhz base crystal is vibrating a 9.9mhz instead lmao

False advertising is false advertising. Just because you get triggered by people calling it out doesn't make it right.
 
I think it’s also important to note the recent 1.0.0.3ABB AGESA actually lowered boost clocks for many people as well.

As a 3700X owner, I’ve never seen the 4.4 GHz core speed hit on a single core, and it’s not because of a 99MHz bus clock. Pretty frustrating to get away from “evil” Intel and go over to AMD just advertising made up numbers as fact. Sure it’s just a few %, however I’d like to get what I paid for. This is on a $330 motherboard as well.

My previous 4.4GHz 7700K hit boost clocks all day long, every day when needed. My RTX 2070 is exceeding the reported boost clocks probably 99% of the time it’s under load. People expect the same of AMD.

You can always go to the AMD forum and perhaps ask for assistance, but you dont even state what your cpu is boosting to right now. My chip has never hit 4.6 most times it's at 4.5 single core and yet my chip outperforms what the reviews showed I would get. I think far too many are over fixated on a number that means nothing, rather just have my chip perform and it does and even a bit better then reviews showed it would.
 
I want my 3700 to sustain the frequencies advertized, if it doesn't I am going to feel like it is not working properly. What is on the box should be a minimum stat sustained consistently. Don't come in here saying most people will not notice xxx mhz, and that its only x%. These shills are the same people who will say that this is supposed to be a "HARD" forum. I guess when it doesn't fit into the "AMD can do no wrong" narrative we are supposed to shrug and give them credit for trying, or make excuses and blame noob consumers. If boost clock means nothing to consumers and sales then leave it off the description and the box. Seriously, this forum is starting to read like a Star Citizen sub-reddit.
 
The 970 had 4GB of VRAM (3.5+0.5 if you want to get technical about how it was allocated). Nvidia deserved the lawsuit, but can people stop spreading misinformation?

The split memory planes essentially means you have a 224-bit/1.75MB L2 cache card, since you can either access the 3.5GB region, or the 512MB region, but not simultaneous. The slow 512mb "cache ram" are less likely to get optimizations in newer games after the card has stopped being supported, which is one of the reasons the lawsuit was filed.

https://www.anandtech.com/show/8935...cting-the-specs-exploring-memory-allocation/2

You don't buy a 70-series card if you ONLY want to play today's games, you buy for longevity.
 
I want my 3700 to sustain the frequencies advertized, if it doesn't I am going to feel like it is not working properly. What is on the box should be a minimum stat sustained consistently. Don't come in here saying most people will not notice xxx mhz, and that its only x%. These shills are the same people who will say that this is supposed to be a "HARD" forum. I guess when it doesn't fit into the "AMD can do no wrong" narrative we are supposed to shrug and give them credit for trying, or make excuses and blame noob consumers. Seriously, this forum is starting to read like a Star Citizen sub-reddit.
That is not how boost clock works. Even Intel's boost clock is not garanteed to be substained as it is all realient on various factors, temperature most of all. Yes, Intel CPU's are generally capable of sustaining single core clock speeds, but it still is not garanteed.
 
You can always go to the AMD forum and perhaps ask for assistance, but you dont even state what your cpu is boosting to right now. My chip has never hit 4.6 most times it's at 4.5 single core and yet my chip outperforms what the reviews showed I would get. I think far too many are over fixated on a number that means nothing, rather just have my chip perform and it does and even a bit better then reviews showed it would.

So if instead of advertising boost clocks you would prefer AMD advertised that your cpu will outperform all the reviews you see or read by and undisclosed amount based on nebulous benchmarks not to be named. Ok, good for you, but I would prefer real measurable statistics.
 
My 3700x hit advertised boost clocks on the launch bios, but seems to be lower around 4375 max on 1003aba.

Performance is higher in this bios though
 
Max boost clock isn't advertised and guarantied. The only guarantee number for 3700X should be 3.6GHz base clock. Get over it guys, the max boost clock is now the silicon lottery (along with other variables that will affect your results).
Yes they were optimistic when they defined those numbers, I agree but you can't expect 100% of parts to hit the OEM "MAXIMUM boost clock"... It would be written minimum instead if that was the case.

I guess it's a good sign when everyone gets their panties twisted over words used like that. Must mean AMD is back in the game ;)
 
3200 stock, but higher recommended by AMD themselves:

View attachment 184263

-------
Interestingly when I first saw this chart (just ahead of launch), to me, 3200MHz c14 looks as the sweet spot - 3733MHz c17 is hard to find (memory that will hit those speeds that doesn't come at a high price).

AMD seems overly excited at times (in pr marketing side of things) - The more near term future is already very bright for them, there's little need to over promise at this point.
 
3700X Spec pulled from AMD.COM clearly states Max Boost Clock 4.4GHz so doesn't guarantee it. Why is everyone thinking otherwise ? 2700X says max boost 4.3GHz and I don't recall people going nuts from not getting it... mine does 4.2GHz on all core with 16GB 3200MHz. I prefer that than 1 core at 4.3GHz.
I tried to find a user guide which explains the "Max" part in Max boost clock but I couldn't find any from a 2minutes google fu.
So if a car's advertised maximum speed is 200mph it doesn't necessarily have to hit it? Your argument has no legs. Max means exactly that, the maximum it should boost to. If it doesn't then it is not up to specs. And since Der8auer tested in the exact scenario AMD specified you can't even argue that the road was uphill.
 
AMD fluffs clocks
Intel fluffs TDP

I guess pick your poison.

Agree that it’s deceptive and will probably end up in a settlement.
 
So if a car's advertised maximum speed is 200mph it doesn't necessarily have to hit it? Your argument has no legs. Max means exactly that, the maximum it should boost to. If it doesn't then it is not up to specs. And since Der8auer tested in the exact scenario AMD specified you can't even argue that the road was uphill.


But much like you can't get a Bugatti Veryron to it's top speed on any old road, you can't expect a variety of motherboards and cooing solutions used for these survey results to give you a clear idea of what is actually limiting the boost clock, here.

Just because it says "up to" doesn't mean that you will hit that limit, if you're running with shitty tires, and a load of groceries in the trunk, or various installations of the boost firmware.. Or that fairly disappointing stock cooler:

https://www.techspot.com/review/1875-ryzen-3900x-wraith-prism-rgb-vs-liquid-cooler/

Until you control for the other parts of the system, you can't say whether the boost clock is working improperly. It just means you can't truer this survey
 
Last edited:
Max boost clock isn't advertised and guarantied. The only guarantee number for 3700X should be 3.6GHz base clock. Get over it guys, the max boost clock is now the silicon lottery (along with other variables that will affect your results).
Yes they were optimistic when they defined those numbers, I agree but you can't expect 100% of parts to hit the OEM "MAXIMUM boost clock"... It would be written minimum instead if that was the case.

I guess it's a good sign when everyone gets their panties twisted over words used like that. Must mean AMD is back in the game ;)

It is advertised, it is guaranteed. In fact the boost clock is listed before the base clock on the box. When a car manufacturer says a car has a top speed, that is the same as saying max speed. The definitions are the same. I have a 3700x, I am not asking for money back, however this is a big deal and should be called out.
 
So if a car's advertised maximum speed is 200mph it doesn't necessarily have to hit it? Your argument has no legs. Max means exactly that, the maximum it should boost to. If it doesn't then it is not up to specs. And since Der8auer tested in the exact scenario AMD specified you can't even argue that the road was uphill.

Bah your definition I guess. I also would argue that the max speed on a car depends on a lot of factor like in this scenario... You can't expect to hit the same speed in higher altitudes or very hot temps. Not sure what your point is and I'm sure max spec is qualified as a typical characteristic and not a hard specification but you're entitled to your opinion. QA would reject those CPU if it was a hard spec... can't believe they do not test the CPU before they ship it... ?

Edit: I understand the grip and emotion around it but wording is important here. I've been burnt in the past with similar things... like 90 days isn't 3 months for XYZ claim, like wtf but yeah I was wrong.
 
Last edited:
It is advertised, it is guaranteed. In fact the boost clock is listed before the base clock on the box. When a car manufacturer says a car has a top speed, that is the same as saying max speed. The definitions are the same. I have a 3700x, I am not asking for money back, however this is a big deal and should be called out.

Could you share the box picture ? I haven't been able to see if it says boost clock or max boost from google image.
 
So if instead of advertising boost clocks you would prefer AMD advertised that your cpu will outperform all the reviews you see or read by and undisclosed amount based on nebulous benchmarks not to be named. Ok, good for you, but I would prefer real measurable statistics.

I prefer performance and the chip delivers better then reviewers got with theirs, I really could care less what it maxes at on boost speeds. I got a bit better then I paid for and those are real numbers and what matters.
 
It is advertised, it is guaranteed. In fact the boost clock is listed before the base clock on the box. When a car manufacturer says a car has a top speed, that is the same as saying max speed. The definitions are the same. I have a 3700x, I am not asking for money back, however this is a big deal and should be called out.

First what car manufacturer advertises top speed, and where? Is that speed the default speed, or the boost speed? How are those speeds effected by fuel grade, terrain, temperature, tires, etc?

Is the motor's horsepower the only influence on the top speed of a car, or does the transmission and it's gear ratio play a part? How about the electrical system?

To use a car vs a cpu, you have to isolate the motor in a car and use horsepower, and isolate the cpu to really use that analogy and to determine if it is the motor or cpu that is faulty and not the motherboard etc. Just as it can be a faulty transmission in a car that can prevent a car from performing as it should.
 
The split memory planes essentially means you have a 224-bit/1.75MB L2 cache card, since you can either access the 3.5GB region, or the 512MB region, but not simultaneous. The slow 512mb "cache ram" are less likely to get optimizations in newer games after the card has stopped being supported, which is one of the reasons the lawsuit was filed.

https://www.anandtech.com/show/8935...cting-the-specs-exploring-memory-allocation/2

You don't buy a 70-series card if you ONLY want to play today's games, you buy for longevity.

I did say the lawsuit was deserved.
 
First what car manufacturer advertises top speed, and where? Is that speed the default speed, or the boost speed? How are those speeds effected by fuel grade, terrain, temperature, tires, etc?

Is the motor's horsepower the only influence on the top speed of a car, or does the transmission and it's gear ratio play a part? How about the electrical system? To use a car vs a cpu, you have to isolate the motor in a car and use horsepower, and isolate the cpu to really use that analogy and to determine if it is the motor or cpu that is faulty and not the motherboard etc. Just as it can be a faulty transmission in a car that can prevent a car from performing as it should.

I'm sure a car "top speed" information do not qualify as a hard specification nor would it be honored by a warranty claim. That's silly lol. Tire thread alone would kill it after a few thousand km.
 
Thanks, also says max boost clock. Back to square one, the definition of it.
Typical Spec (or characteristic) can be a trap for the user. Do AMD have a published user manual / datasheet with description ?

If it hits the max clock once in a blue moon for .5 nanoseconds (obvious exaggeration, but you get the point) does it really count as hitting the advertised boost speed? The way Zen 2 handles boosting is very smart, but it also means if your chip can reach the promised max boost clock its only there for fractions of a second at a time.
 
I'm sure a car "top speed" information do not qualify as a hard specification nor would it be honored by a warranty claim. That's silly lol. Tire thread alone would kill it after a few thousand km.

Yep. I was just demonstrating how silly his whole car argument is.
 
But much like you can't get a Bugatti Veryron to it's top speed on any old road, you can't expect a variety of motherboards and cooing solutions used for these survey results to give you a clear idea of what is actually limiting the boost clock, here.

Just because it says "up to" doesn't mean that you will hit that limit, if you're running with shitty tires, and a load of groceries in the trunk, or various installations of the boost firmware.. Or that fairly disappointing stock cooler:

https://www.techspot.com/review/1875-ryzen-3900x-wraith-prism-rgb-vs-liquid-cooler/

Until you control for the other parts of the system, you can't say whether the boost clock is working improperly. It just means you can't truer this survey
The "road" and measuring method was specified by AMD did you miss that part? I even mentioned it. So you couldn't have.

Do you honestly think 95% of 3900x owners have shitty equipment?

I think you're being disingenuous if you answer yes.
 
Last edited:
Bah your definition I guess. I also would argue that the max speed on a car depends on a lot of factor like in this scenario... You can't expect to hit the same speed in higher altitudes or very hot temps. Not sure what your point is and I'm sure max spec is qualified as a typical characteristic and not a hard specification but you're entitled to your opinion. QA would reject those CPU if it was a hard spec... can't believe they do not test the CPU before they ship it... ?

Edit: I understand the grip and emotion around it but wording is important here. I've been burnt in the past with similar things... like 90 days isn't 3 months for XYZ claim, like wtf but yeah I was wrong.
AMD gave the specifications of how should it be tested. So if it doesn't hit it within their specified test conditions, when do they hit it?

AMD marketing overreached again, not the first time for that. Except for the first time in a long time they have a product where it was absolutely unnecessary. They could've specified a boost clock that every chip was guaranteed to hit, and the few that went above would be the discrepancy. This is amateur hour. I already found it strange that the higher clock count cpus had higher specified boost clocks. It wasn't necessary. THey could've said the 3900X boosts to 4500 it would still be hell of a deal and everyone who bought it with 4600 advertised boost would've gotten it with 4500 all the same.
 
The road and measuring method was specified by AMD did you miss that part? I even mentioned it. So you couldn't have.
Do you honestly think 95% of 3900x owners have shitty equipment?
I think you're being disingenuous if you answer yes.

Some do hit the "Max" boost clock so they didn't lie. I will agree that they should of used different words to prevent this confusion. I would also think that those word and target were carefully chosen to be along the "competitive" lines versus intel and to sell dream even if it doesn't change much... 1 core hitting 4.6GHz.. where will you benefit from that? In Excel ? Boost clock (Max boost here) were never a specification / characteristic for all cores anyway.

I would love to read the full datasheet / customer manual if anyone can point it out. Uncertainties & typical characteristic are a norm in the domain I work so I'm not surprised it caused confusion.
 
  • Like
Reactions: N4CR
like this
It's annoying to see the lengths people go to to excuse this shitty practice from AMD. We are now at a place where people come up with the most obtuse definitions that explain how AMD isn't actually lying. We understand, you are satisfied and the chips get the benchmark scores they get regardless of stated frequency, but let us stop with the BS and admit this is bullshittery from AMD. First we went from all core boost to single core boost to not sustained boost to occasional boost to once in a blue moon milisecond boost for 5% of CPUs.

It's downright lying at worst, dishonest marketing at best.
 
I think the 20+ hardware exploits for Intel CPUs from 1995-present are a bit more scandalous, and serious, than the lack of boost clocks.
Even without it, these are still damn good CPUs, especially for the cost.
 
Back
Top