Review AMD's New Binning Strategy on Ryzen 3000: Core-by-Core Turbo Analysis

[Giroro]

As the article mentions, these processors can adjust their clock speed hundreds of times per second, but if a utility were to actually poll the processor for its clock speed hundreds of times per second, that would negatively impact performance, and could actually even affect those clock rates. So, the utilities are likely only updating their readings about once per second and perhaps averaging results, so any quick bursts to the full clock rate would likely not get detected. Even the test setup for this article only polled the processor's clock rate 10 times per second, which is still a fairly course measurement relative to the rate at which the clocks can change on these processors.

If there is particular test configuration that AMD is using to validate that their processors can meet the advertised specs, then AMD should be sharing that that hardware/software/polling rate with reviewers. AMD is making a marketing claim, so it's their responsibility to prove it. It's not a reviewer's job to jump through extreme hoops to defend how a profit-driven publicly-traded company's advertisements aren't technically illegal. That isn't the customer's job either. It's a reviewer's job to help protect consumers by reporting how the product holds up to real-world testing.

Until AMD can find a repeatable way to demonstrate that these processors hit the claimed boost clock, then you should err on the side of caution and assume that they do not hit the claimed boost clock.

 

[Ncogneto]

Fine. Want related? What about the lawsuit against the FX series CPUs? While I don't agree with it, the suit is happening due to AMD marketing the CPU as a 8 core part and depending on how the court defines a core it may or may not win. The biggest issue is that a dual core section of an FX CPU is not equal to what Intel or current AMD CPUs.

This is similar in that AMD lists the CPU as a turbo clock of say 4.6GHz. Product never hits 4.6GHz.

Again I am not saying it is something that anyone should sue over or that any lawsuit should be filed. I find the majority of them, my Microsoft one included, stupid. I got a settlement claim paperwork for one against Microsoft and threw it away.

What I am saying is that depending on the country and the consumer laws it is possible AMD might face a lawsuit due to this. Do it mean they should? No. Is it still possible? Of course it is as people sue over everything these days even if it doesn't directly affect them.

This is a small issue that could potentially cause AMD legal problems depending on the country and consumer laws.

Also margin of error is different than advertised specifications.

which brings us back full circle, show me an example of a COUNTRY. Your losing this argument.....badly.

 

[jimmysmitty]

which brings us back full circle, show me an example of a COUNTRY. Your losing this argument.....badly.

I think I see the issue. I never said the country. I just said its a potential lawsuit depending on the country and their consumer laws, some of which heavily favor the consumer over a corporation. And no matter how you spin it it is a potential. The FX lawsuit was approved in the US which laws do not heavily favor the consumer.

 

[Ncogneto]

I think I see the issue. I never said the country. I just said its a potential lawsuit depending on the country and their consumer laws, some of which heavily favor the consumer over a corporation. And no matter how you spin it it is a potential. The FX lawsuit was approved in the US which laws do not heavily favor the consumer.

Except you did:

"There are countries where something as little as "advertised clock speed" not meeting would be a big problem. "


And how did the FX lawsuit turnout?

 

[Arbie]

AMD should ensure that at least one core meets the turbo value. That's it. And that's all they've claimed. The more cores (actually threads) running, the lower all of them will be clocked - by design.

If you got a chip where no core will reach the turbo, then you have a complaint - though probably only a small one. But all the rest of this anti-AMD bitching is baseless. And as has been noted, the article title is unjustifiably negative and click-baity. It seems pretty hard for AMD to get a fair shake in the mainstream tech press.

 

[jimmysmitty]

Except you did:

"There are countries where something as little as "advertised clock speed" not meeting would be a big problem. "


And how did the FX lawsuit turnout?

Yes I can see how that could be misleading but I did mean in terms of consumer laws for countries not the county itself suing AMD for it.

And I will let you know when I see the results. It was started in 2015 but it was approved to move forward this year. My point of that is that a judge in the US, again a country with far less consumer leaning laws, is allowing it to proceed. In a country where the consumer laws are much more favorable it could easily become a problem for AMD to deal with. I never said it will or that I actually want it to but that it is a possible issue.

My solution would have been the best. AMD should set the speed to a universal high they can achieve and if it goes above that its a bonus.

 

[germanium]

One thing I would like to mention is that if said boost clocks are measured through windows task manager they are quite often inaccurate & often read lower than they are actually. CPU-Z measures each core individually & is more accurate than windows. My windows reads 4.97 GHz when it is actually bouncing from 4.998 to 5.002GHz on all cores according to CPU-Z

 

[gerardfraser]

Well I did a video of my 3600X because i still disagree with article.It is all in the BIOS used to achieve rated boost clocks.

Cinebench 20 all cores hit and past Maximum boost written on the box of 4400Mhz
Valley Benchmark all cores hit and past Maximum boost written on the box of 4400Mhz.

View: https://www.youtube.com/watch?v=AHdzllmzY7c


Screens from video if you do not want to look at the boring stuff.

Cinebench20
Core 0=4450Mhz
Core 1=4450Mhz
Core 2=4425Mhz
Core 3=4425Mhz
Core 4=4400Mhz
Core 5=4400Mhz

Valley Benchmark
Core 0=4525Mhz
Core 1=4525Mhz
Core 2=4425Mhz
Core 3=4425Mhz
Core 4=4400Mhz
Core 5=4450Mhz

Spoiler

find nearest open gas station

 

[gerardfraser]

EDIT Aug 12 2019:Let me start by saying I was wrong and I apologize for any misinformation I supplied due to my ignorance.
I was aware of Third-Party Monitoring Tools not being as accurate as Ryzen Master and I just ignored that fact all together.

I have started testing with Ryzen Master only now to see what the actual clocks on my Ryzen 3600X really is.
Although I still get over Max Boost as what is written on the box so I have no recourse to complain to AMD or refund my CPU'S.
Also they are still fast little CPU's Still happy.

I have only witness Cinebench20 hitting Max Boost on 2 cores as of testing today.I will test some more with different BIOS and if any change is observed I will post back.

 

[redgarl]

This is pathetic how right I was, me, an ignorant reader, and how wrong you were, the so call tech press. Anyone that argued with me are now looking bad... it is not even funny. Anyone with an head on their shoulder would have been able to make that assumption... but I guess Toms was too in a hurry to publish this rubbish than really investigating...

https://www.tomshardware.com/news/amd-ryzen-3000-boost-frequency-bios-fix,40308.html

 

[jimmysmitty]

This is pathetic how right I was, me, an ignorant reader, and how wrong you were, the so call tech press. Anyone that argued with me are now looking bad... it is not even funny. Anyone with an head on their shoulder would have been able to make that assumption... but I guess Toms was too in a hurry to publish this rubbish than really investigating...

https://www.tomshardware.com/news/amd-ryzen-3000-boost-frequency-bios-fix,40308.html

If the tech media and people didn't say anything about it there is a good chance that AMD might have just ignored it. It is never a bad thing for tech media to question something or present current issues with a product. It helps keep them in check.

Now we will have to see if the BIOS fix does its job and if there are any trade offs for said fix.

 

[evilhf]

Here it reaches even more than the cash boost!
I am very happy with this wonderful CPU.
And even more looking forward to the improvements that will come with new BIOS !!

https://uploaddeimagens.com.br/imagens/7-4_original-jpg