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

Wow, 7 pages already and I am still loving my 1600, 1700 and 2600. :) Upgrading to the 2600 will be not worth it, since the games are play, like Shadow of the Tomb Raider, are GPU bound games.
 
NKD said:
meh, my max boost is around 4575 as of now. Either way I am running all core OC at 4350 with 1.31v and memory at 3733 with below 65ns latency. While everyone has the right to bitch, I am not complaining since it blows away anything intel has at that price point. So AMD has given me a great bang for buck, I am okay with losing 25mhz on max boost for now, I am pretty sure future bioses will do just fine as the platform matures.

Feel free to return your processors event though it doesn't make shit of a difference as you are getting the same performance as the reviews. if single core boost is what you are after on a 12 core beast all power to you.
Click to expand...

What is great is the 4350 at 1.31v, none of the prior to 3000 series processors will get close to that.
 
ChadD said:
I have been considering that. Where I am the local good part shop has Cool Master 240L AIOs for $80 Canadian. A D15 is $140... main reason I was leaning toward the AIO. Having said that they do have the NH-U14s for $100. Your point is well taken though... for 20 bucks probably be quitter and leaks won't ever be an issue.
Click to expand...

Well, I use the NH-U14s on my 1600 running at 4.0 GHz and 1.4v and it works great. My Corsair H100i is much more noisy, fan wise, than the Noctua coolers.
 
  • Like
Reactions: ChadD
like this
SilverSliver said:
So, excuses I've heard about Ryzen 2 not reaching stock speeds now.

It's new tech!
Boost is close enough!
User errorrrrrrrrr
Moon phases
<insert Intel product that sucks, or may potentially suck here>
Moar fans
Dark matter
hAvE yOu uPdAtEd uR bIoS?
"cutting edge"
what difference, at this point, does it make?
BLENDER!
Click to expand...

Boost speeds are not stock speeds, please use the proper terminology.
 
ManofGod said:
Well, I use the NH-U14s on my 1600 running at 4.0 GHz and 1.4v and it works great. My Corsair H100i is much more noisy, fan wise, than the Noctua coolers.
Click to expand...

I’ve had so many AIOs fail from so many different sources I like to keep it simple in a lot of cases with air, and agree with you on the NH’s.
 
He said that the results are unscientific and that chances are that only those who are having issues reported in.

Because there was absolutely no scientific method utilized. No confounding variables mitigated, no research plan, no hypothesis to disprove, no ability to render the experiment over again, no repeatability, no falsifiability, no methods, no equipment, no peer review, it's essentially pure FUD.

What 2000 online surveys were submitted out of hundreds of thousands of chips sold?

That's akin to ONLY 1000 people surveyed that called in who had cancer
Therefore if 1000 people had cancer out 1000 surveyed then 100% of the population has cancer.

I'm sure Derbauer has a college degree. And if he does certainly he learned how to apply the scientific method. Until he uses it and is properly peer reviewed to rule out violations of the method this article means


NOTHING

I however commend him for his work but we have methods for reasons to prevent fud. Just like the anti vaxxer fud.
 
So are the 3700x/3800x 8 cores at say 4.2ghz faster than a 5960x at 4.6ghz?
 
SilverSliver said:
So, excuses I've heard about Ryzen 2 not reaching stock speeds now.

It's new tech!
Boost is close enough!
User errorrrrrrrrr
Moon phases
<insert Intel product that sucks, or may potentially suck here>
Moar fans
Dark matter
hAvE yOu uPdAtEd uR bIoS?
"cutting edge"
what difference, at this point, does it make?
BLENDER!
Click to expand...


The thing is, there are hundreds of thousands sold around the world, and yet, not a peep about boost problems from the majority in the past 3 months... Until this bogus survey.

If it was actually true, we would have heard about it long before now.
 
ManofGod said:
Please just return your 3000 series CPU and get a refund. Then you can go buy an Intel 9900KFC and we will all be happier for it. Good luck.
Click to expand...

I was outside the return window before the bios bringing the boost clock up from 4.3ghz to 4.575ghz was released. AMD didn't respond to my RMA request for almost a month and a half. I've had major reconstructive wrist surgery in the mean time and it would be physically impossible for me to rebuild my system, at this point, for months now.

NWRMidnight said:
The thing is, there are hundreds of thousands sold around the world, and yet, not a peep about boost problems from the majority in the past 3 months... Until this bogus survey.

If it was actually true, we would have heard about it long before now.
Click to expand...

It's been a well documented problem on reddit, overclock.net, tweaktown and guru3d forums. There are Google sheets running around with people submitting data and trying to figure out what's going on.

Just because you have not had a problem, or are ignorant of the problem existing, does not mean a problem does not exist.
 
Let's say I bought the 3800x for $70 more than the 3700x for the extra boost clock. But my 3800x won't ever hit 4.5GHz and stay stuck at 4.4GHz. Would that count as AMD having scammed me of $70?
 
trandoanhung1991 said:
Let's say I bought the 3800x for $70 more than the 3700x for the extra boost clock. But my 3800x won't ever hit 4.5GHz and stay stuck at 4.4GHz. Would that count as AMD having scammed me of $70?
Click to expand...

Here is the problem. If you are buying this processor based off boost clock dont do it. I mean its a bad purchase even if it was hitting the boost clock. Unless you were mistaken in the first place that it was all core boost, which it isn't. Now ask yourself this, are you buying 3800x for single core boost? If so you just made a bad purchase decision to begin with. Only reason to buy 3800x is to have higher chances of getting a higher binned chip for manual OC.

Now i would never recommend 3800x over 3700x but I have seen people get better all core OC on average out of 3800x.
 
pillagenburn said:
Anyone unhappy with their Ryzen 3000 series CPU should immediately return it for a full refund and purchase a comparable Intel product immediately.

...................Instead of filing a class-action lawsuit.
Click to expand...


No, actually we really do need a class-action lawsuit and I hope there is one.

With millions in fines. This is the only way AMD can learn their lesson. There has to be severe damages until they realize they need to be legit at all times.

I know a lot of people have lost an incredible amount of their new found respect for AMD.
 
trandoanhung1991 said:
Let's say I bought the 3800x for $70 more than the 3700x for the extra boost clock. But my 3800x won't ever hit 4.5GHz and stay stuck at 4.4GHz. Would that count as AMD having scammed me of $70?
Click to expand...

Nope because, from what I understand, that is not the only differences between those two processors.
 
The thing that everyone is missing, higher clocks do not always mean higher performance in gaming.

AMD has tuned Ryzen with the latest AGESA to have increased performance even *with* slightly lower clocks.

If you think about it, AMD has given us *more* performance than the launch reviews stated, not less. Forget about the GHz, it's not an accurate determination of real-world performance.
 
Last edited:
https://www.techpowerup.com/forums/...ue-whats-your-take.258696/page-4#post-4108389
and if nobody wants to click the link

(He was talking about all-core overclock level, however the automatic boost ALSO depends totally on temperature!)

---
AMD's mistake was omitting this aspect in their presentations

This is what it should say on the slide:
Ryzen 9 3900X - Max boost 4600Mhz at 20 degrees ambient with fan curve on "Turbo" (or aftermarket cooler capable of 150W TDP)

---
Unfortunately derb8er completely neglected this aspect in his poll.
With added question: What is your ambient temp during testing?
- under 15 °C (I live in Siberia)
- 19 °C (AC on - I like it cold)
- 23 °C (Nice and pleasant)
- 28 °C (I'm poor and no AC)
- 35+ °C (Please kill me!)

I'm 99% convinced everyone with lower boosts have higher ambient, or they simply like a warm room where they use their PC, while those that did boost at advertised or higher live in colder climates.
And the reason most CPU's in the poll did not boost at advertised, is because probably most results come from Europe or USA, where right now it's a hot summer with plenty of heatwaves.
 
cybereality said:
The thing that everyone is missing (maybe no one reads my posts): higher clocks do not always mean higher performance in gaming.

If that was true, going from 3.2 GHz to 4.3 GHz would almost be a 30% boost in FPS, but this doesn't happen.

See here, for example, the performance is the same basically:

View attachment 184745

https://www.pcgamesn.com/amd/ryzen-7-3700x-overclocking-benchmarks-performance

AMD has tuned Ryzen with the latest AGESA to have increased performance even *with* slightly lower clocks.

If you think about it, AMD has given us *more* performance than the launch reviews stated, not less. Forget about the GHz, it's not an accurate determination of real-world performance.
Click to expand...

They OC’d the memory ... that’s what the 3.2 vs 3.733 was.
 
cybereality said:
The thing that everyone is missing (maybe no one reads my posts): higher clocks do not always mean higher performance in gaming.

If that was true, going from 3.2 GHz to 4.3 GHz would almost be a 30% boost in FPS, but this doesn't happen.

See here, for example, the performance is the same basically:

View attachment 184745

https://www.pcgamesn.com/amd/ryzen-7-3700x-overclocking-benchmarks-performance

AMD has tuned Ryzen with the latest AGESA to have increased performance even *with* slightly lower clocks.

If you think about it, AMD has given us *more* performance than the launch reviews stated, not less. Forget about the GHz, it's not an accurate determination of real-world performance.
Click to expand...

I don't think anyone is arguing about the performance received versus what was advertised. What people have a problem with is AMD advertised certain things - boost clocks and, to a lesser extent, overclocking expectations and are falling short of both in the vast majority of cases being reported. Considering PBO doesn't even work, on any board, I don't know why it's listed on this graph. Manual OC on the 3900x, using per CCX, nets around a 7% multicore improvement if you can get into the 4.3 to 4.45 range. Won't affect single core, really.
 
SixFootDuo said:
No, actually we really do need a class-action lawsuit and I hope there is one.

With millions in fines. This is the only way AMD can learn their lesson. There has to be severe damages until they realize they need to be legit at all times.

I know a lot of people have lost an incredible amount of their new found respect for AMD.
Click to expand...

So organize a "let's all return our Ryzens at the same time" group on reddit or something and set it for a specific date - then return them all at the same time.

Instead of giving Lawyers more undeserved money.
 
Seems like the buying public does not care as AMD is selling almost 4 to 1 at Mindfactory.de

mPSx5QI.jpg


https://1.bp.blogspot.com/-bJwBYgRm...aueYvYh2hcJapWgeAQ34QCLcBGAs/s640/mPSx5QI.jpg
 
ManofGod said:
Well, I use the NH-U14s on my 1600 running at 4.0 GHz and 1.4v and it works great. My Corsair H100i is much more noisy, fan wise, than the Noctua coolers.
Click to expand...

Everyone that has one says the same. Pretty sure my mind is made up. Going to go Noctura, never have had one and I guess it's time. Now I'll have to figure out exactly which to order. The NH-15 on newegg for 110 is tempting. Looks like if you take the one fan off its basically a NH-15S, which means you pay 10 bucks more for an extra noctura fan, not a bad deal.

Now I just have to decide if I go all out and get a cool master silencio or a be quiet case or something.
 
SilverSliver said:
I was outside the return window before the bios bringing the boost clock up from 4.3ghz to 4.575ghz was released. AMD didn't respond to my RMA request for almost a month and a half. I've had major reconstructive wrist surgery in the mean time and it would be physically impossible for me to rebuild my system, at this point, for months now.



It's been a well documented problem on reddit, overclock.net, tweaktown and guru3d forums. There are Google sheets running around with people submitting data and trying to figure out what's going on.

Just because you have not had a problem, or are ignorant of the problem existing, does not mean a problem does not exist.
Click to expand...

Well documented from a hand full of various people is not a wide spread issue, regardless of what platform is used (reddit, tweakdown, etc) Even here, we have a few who have posted with problems. Hell, you can go look up any product in the world, and you will find people having issues. Even the best products ever made. There is no getting around it, as every product has it's share of defects and problems, even your faithful intel. People who are not having issues, don't talk about it, all you usually ever see is those that have issues, so you are assuming that it is a wide spread problem. It isn't. All this bogus survey did was give you a false sense of "look, it's happening to everyone"... when in fact it's not.

Now, on release, there where boost issues, due to the beta bios's that most MB manufactures used, which was corrected shortly after. Beyond the first couple weeks of bios issues, the amount of people having issues is normal for any product. This doesn't mean we are saying you, and other's are not having issues, but the amount of those issues is not any higher than any others. If it there was indeed an actual issue, specially with boost clocks, or the chips not hitting their rates speeds, we would have heard about it long before this bogus survey came out.
 
Last edited:
Even if we ignore that many CPUs aren't reaching the boost clock, it's worth noting that we accepted the fact that the CPU's box advertised frequency mostly only relates to a single golden core out of 6-12.

Again, the Ryzen 3000 is performing very well, but just because we are happy AMD is competitive again, doesn't mean we have to put up with their ever changing definitions and shitty marketing standards.

"Oh, it's actually single core boost only, oh, it's actually only specific single core only, oh, it's actually only specific single core only milliseconds at a time under LN2 but even then for select lucky units only, so technically we aren't lying."
 
Mode13 said:
Beyond wishful thinking, but he's right.. It's what was explained to consumers in that video in plain English as within reason given ample cooling and a good vrm. I posted the SS on the prior page where he showed PBO taking a 4550 boost core up to 4750.

Seeing it written down that way makes it even more hilarious with the 3950x.

This is straight from AMD themselves and still posted on their youtube. I can also bet precisely 0 people without LN2 have seen 4.75ghz on any ryzen 3xxx core.
Click to expand...

I don't disagree. This is what AMD told reviewers, although it was clear to us during testing that we would be unable to hit those frequencies or at the very least, it would be extremely rare. AMD was also clear in its reviewers guide that adding the 200MHz offset wasn't guaranteed.
 
Meeho said:
I guess that's their euphemism for "doesn't work at all".
Click to expand...

to be perfectly honest, I never expected it to work knowing how close to the edge these things were by the time I read that in the guide. Essentially, it's PBO+AutoOC with a 200MHz offset. When the damn things barely boost to their maximum stated boost clocks, it just didn't seem likely. But I get why people are so upset about it. When you read that you can set the CPU to do that, you expect that it should be possible.

Personally, I think AMD over promised and under delivered. These chips can often do 4.3GHz on all cores. Had these things been set at 4.4GHz max boost and been good for 4.3GHz all core, I don't think anyone would be complaining. That said, these frequent micro-boosts into the 4.4GHz+ range (and higher on the 3900X) are what give AMD the single-threaded scores that we see in all the benchmarks. AMD was simply too desparate to stick it to Intel that they lost sight of the fact that they were over extending themselves on what they could promise.
 
Last edited:
Dan_D said:
to be perfectly honest, I never expected it to work knowing how close to the edge these things were by the time I read that in the guide. Essentially, it's PBO+AutoOC with a 200MHz offset. When the damn things barely boost to their maximum stated boost clocks, it just didn't seem likely. But I get why people are so upset about it. When you read that you can set the CPU to do that, you expect that it should be possible.

Personally, I think AMD over promised and under delivered. These chips can often do 4.3GHz on all cores. Had these things been set at 4.4GHz max boost and been good for 4.3GHz all core, I don't think anyone would be complaining. That said, these frequent micro-boosts into the 4.4GHz+ range (and higher on the 3900X) are what give AMD the single-threaded scores that we see in all the benchmarks. AMD was simply too desparate to stick it to Intel that they lost sight of the fact that they were over extending themselves on what they could promise.
Click to expand...
Pretty much this.

If there is one lesson AMD should have learned the past years, it's that in the long run it's better to underpromise and overdeliver. I appreciate this is somewhat of a shift in the way CPUs are operating, bringing them closer to GPUs, thus not all of the traditional specification conventions apply, but even then Nvidia [of all companies] managed to do a better job handling the marketing and actual results of stated opportunistic boost frequencies.
 
Dan_D said:
to be perfectly honest, I never expected it to work knowing how close to the edge these things were by the time I read that in the guide. Essentially, it's PBO+AutoOC with a 200MHz offset. When the damn things barely boost to their maximum stated boost clocks, it just didn't seem likely.
Click to expand...
This bears highlighting. AMD must have known this as well! Why in the world would they release that video?

Apparently this is also a thing:

mmwaf1992nj31.png



And who could forget the Fury X overclocker's dream gem:
 
illli said:
I'm 99% convinced everyone with lower boosts have higher ambient, or they simply like a warm room where they use their PC, while those that did boost at advertised or higher live in colder climates.
And the reason most CPU's in the poll did not boost at advertised, is because probably most results come from Europe or USA, where right now it's a hot summer with plenty of heatwaves.
Click to expand...

I live in the tropics and my 3700X with the ABB AGESA only does 4367 mhz max on only 2-3 cores even with AC going.

With the older 1.0.0.2 AGESA it did 4399 mhz on multiple cores (4.4 for all intents and purposes).

Not that I can feel the difference but those are the numbers for everyone.
 
Really wish that amd could create a personalized agesa that allows you to achieve those extra few mhz while reverting the performance gain, so you get nothing effectively and stay happy watching the meaningless higher single core boost.

Because yeap that's how it seems to be, the latest agesa actually boosting effective single core performance is meaningless because the number in hz is slightly smaller.

Ps: I really hope that you guys are contacting a class action expert firm with all the information, including that small tidbit of the lower number hz not being directly proportional to lower performance but the other way around instead increasing the performance vs release day.
 
Oh I wanna see them try and file it, I already chose the moniker "The Spinal Tap Class Action Lawsuit"
Because why doesn't their amp go all the way to 11 anymore? What is that thing about it still having higher dB? It needs to go to 11!
 
You must log in or register to reply here.
Back
Top