Discussion Ryzen 3000 series benchmark thread ** Open **

[beginner99]

(for now we can call them out for bad launch BIOS support).

bascially typical AMD launch. It's even more puzzling as some of the insiders have said that the mobos have literally been availabel for 2-3 months before launch. So I guess AMD has a problem on the software / bios side of things. I suspect they did a respin pretty late for better clocks and hence it isn't all sorted out yet.

Still given that they could have just put a safe number on the package. No one would complain if it boosts higher than advertised. The 4.6 ghz on the 3900x seems like they just had to put a higher number on there.

 

[coercitiv]

No one would complain if it boosts higher than advertised.

Kinda agree. My only observation would be that single core ratios (or targets) should be removed altogether.

 

[Zucker2k]

There's a low clock ceiling of about 4.4GHz-ish which some individual cores seem to be able to breach. In my opinion, Zen2 is mostly a 4.3GHz product, if AMD had gone conservative with their boost algorithms. Zen+ would probably had exhibited the same problems we're witnessing here but AMD wisely held back at 4.35GHz. Look at the voltages they're pumping into these 7nm chips; 1.47v for the 3900x, for example. I don't know any process that'll scale up high (4.6GHz+) on those voltages on air. In light of the above, is it any wonder that the lower core products have no issues clocking up to their advertised boost clocks? I think not. The CCX design, and die layout all seem to be accessories to the heat issues as well. Hot-spotting, anyone?

 

[Kenmitch]

Ryzen Master has a new ECO option available for the 3xxx series of chips.

Has anybody seen any reviews that have tested it?

I decide to play around with it myself to see what it does.

Here's the normal stock setting behavior of my 3600



The ECO mode results



Guess I'll have to do some more playing around to see how it effects other use scenarios.

 

[Panino Manino]

Intel money is at play again:

https://www.reddit.com/r/Amd/comments/chal0r

 

[Thunder 57]

Intel money is at play again:

https://www.reddit.com/r/Amd/comments/chal0r

Wow. This is worthy of a new thread IMO as many won't even see it buried in this one. So what they did was change the weighting from Single Core/ Quad Core/ Multi core from 30/60/10 to 40/58/2. I'm sorry but the timing seems awfully suspicious.

 

[moinmoin]

Intel money is at play again:

https://www.reddit.com/r/Amd/comments/chal0r

The effective CPU speed index measures performance with the following weights: 40% single-core, 58% quad-core and 2% multi-core.
Source

Before Ryzen was released the ranking was based on: 30% Single core performance 60% Quad core performance 10% multi core performance
Internet Archive

ROFLcopter

 

[PotatoWithEarsOnSide]

There's a low clock ceiling of about 4.4GHz-ish which some individual cores seem to be able to breach. In my opinion, Zen2 is mostly a 4.3GHz product, if AMD had gone conservative with their boost algorithms. Zen+ would probably had exhibited the same problems we're witnessing here but AMD wisely held back at 4.35GHz. Look at the voltages they're pumping into these 7nm chips; 1.47v for the 3900x, for example. I don't know any process that'll scale up high (4.6GHz+) on those voltages on air. In light of the above, is it any wonder that the lower core products have no issues clocking up to their advertised boost clocks? I think not. The CCX design, and die layout all seem to be accessories to the heat issues as well. Hot-spotting, anyone?

What relevance is a clockspeed when it hits its performance levels irrespective of what number you get to jerk off to?
Too many idiots thinking that clockspeed is the ultimate arbiter of performance. Newsflash: it isn't. Bulldozer was hitting high clocks with crap performance, and Sunny Cove supposedly hits great performance with "crap" clocks.
It doesn't matter where your performance comes from; if the performance is there, the performance is there. Ryzen 3000 is there (or thereabouts when it comes to gaming).

 

[Thunder 57]

The effective CPU speed index measures performance with the following weights: 40% single-core, 58% quad-core and 2% multi-core.
Source

Before Ryzen was released the ranking was based on: 30% Single core performance 60% Quad core performance 10% multi core performance
Internet Archive

ROFLcopter

If anything multicore should be going up at the expense of single core. Maybe 20/60/20 or even 20/50/30 since there are so many 6+ core CPU's out these days.

 

[Markfw]

The effective CPU speed index measures performance with the following weights: 40% single-core, 58% quad-core and 2% multi-core.
Source

Before Ryzen was released the ranking was based on: 30% Single core performance 60% Quad core performance 10% multi core performance
Internet Archive

ROFLcopter

At least Ryzen is now much stronger on single core, but yes, sounds like Intel again and money can buy anybody (almost)

 

[ondma]

Yea, intel has bought them off for sure. That is why the first thing you see on their homepage is huge pictures of 2 ryzen processors.

 

[moinmoin]

At least Ryzen is now much stronger on single core, but yes, sounds like Intel again and money can buy anybody (almost)

It looks like only delaying the inevitable, Ryzen 4k may well be surpassing 14nm++ Intel in single, quad (thanks to further CCX optimizations) and multi core all at once, then the balance between all of them is moot.

The result of this balance tweak is pretty funny though even for Intel chips looking through the list sorted by "Avg. bench %" with fun stuff like i3 8350K and i5 8400 being on the same level.

 

[Markfw]

I tried it for grins.

UserBenchmarks: Game 138%, Desk 172%, Work 173%
CPU: AMD Ryzen 9 3900X - 95.8%
GPU: Nvidia GTX 1080-Ti - 158.8%
SSD: Corsair Force MP600 NVMe PCIe M.2 1TB - 418.8%
HDD: Seagate ST1500DL003-9VT16L 1.5TB - 66.7%
RAM: G.SKILL F4 DDR4 4266 C19 2x8GB - 116.8%
MBD: Asus PRIME X470-PRO

Looks pretty good to me.

 

[Thunder 57]

It looks like only delaying the inevitable, Ryzen 4k may well be surpassing 14nm++ Intel in single, quad (thanks to further CCX optimizations) and multi core all at once, then the balance between all of them is moot.

The result of this balance tweak is pretty funny though even for Intel chips looking through the list sorted by "Avg. bench %" with fun stuff like i3 8350K and i5 8400 being on the same level.

At which point they'll change it so multicore is 90% to make sure Zen pays an inter-CCX penalty .

 

[Zucker2k]

But I thought the consensus was that AMD has the ipc advantage now; apparently not.

 

[Space Tyrant]

But I thought the consensus was that AMD has the ipc advantage now; apparently not.

AMD may have the IPC advantage, but Intel has the clock speed advantage, and it is surely larger.

 

[Zucker2k]

What relevance is a clockspeed when it hits its performance levels irrespective of what number you get to jerk off to?
Too many idiots thinking that clockspeed is the ultimate arbiter of performance. Newsflash: it isn't. Bulldozer was hitting high clocks with crap performance, and Sunny Cove supposedly hits great performance with "crap" clocks.
It doesn't matter where your performance comes from; if the performance is there, the performance is there. Ryzen 3000 is there (or thereabouts when it comes to gaming).

Forget clocks, the performance came about by sorcery.

Don't change the topic. The question at stake is about unreachable boost clocks and potential risks to AMD's reputation and possible legal jeopardy over false advertising.

 

[Zucker2k]

AMD may have the IPC advantage, but Intel has the clock speed advantage, and it is surely larger.

But clock speeds don't matter /s

 

[Markfw]

Forget clocks, the performance came about by sorcery.

Don't change the topic. The question at stake is about unreachable boost clocks and potential risks to AMD's reputation and possible legal jeopardy over false advertising.

This is a benchmark thread. And the statement "the performance came about by sorcery. " is just trolling. And "legal jeopardy over false advertising" ? again, trolling. It has been shown they do it in some cases, just not as many as people like;

 

[Zucker2k]

This is a benchmark thread. And the statement "the performance came about by sorcery. " is just trolling. And "legal jeopardy over false advertising" ? again, trolling. It has been shown they do it in some cases, just not as many as people like;

How come you missed the part where the person I responded to said everybody who doesn't agree with him that clocks are not the ultimate arbiter of performance is an idiot? Here, I'll quote it:

Too many idiots thinking that clockspeed is the ultimate arbiter of performance

That whole post is neither here nor there because performance is always hinged on clockspeed. One user even calculated exactly what the contentious 100MHz amounts to - 2.2%. So yeah, that means something to some people, but they should all discard their opinions for one that says Zen 2 hits its "performance levels" (whatever that is) and anybody who thinks otherwise is an idiot.

Clock speed IS indeed the ultimate arbiter of performance. IPC is more or less fixed in silicon. The main thing that can change how that ipc interacts with code is clockspeed, all else being equal. That is why this is important; along with whatever the implications are for the manufacturer vis a vis the enthusiast community.

 

[Markfw]

How come you missed the part where the person I responded to said everybody who doesn't agree with him that clocks are not the ultimate arbiter of performance is an idiot? Here, I'll quote it:



That whole post is neither here nor there because performance is always hinged on clockspeed. One user even calculated exactly what the contentious 100MHz amounts to - 2.2%. So yeah, that means something to some people, but they should all discard their opinions for one that says Zen 2 hits its "performance levels" (whatever that is) and anybody who thinks otherwise is an idiot.

Clock speed IS indeed the ultimate arbiter of performance. IPC is more or less fixed in silicon. The main thing that can change how that ipc interacts with code is clockspeed, all else being equal. That is why this is important; along with whatever the implications are for the manufacturer vis a vis the enthusiast community.

You are not making any sense. In many cases, the Ryzen 3000 series beats even the 9900k at a lower clockspeed. So how can you possibly say that "Clock speed IS indeed the ultimate arbiter of performance " ?

 

[Vattila]

The question at stake is about unreachable boost clocks and potential risks to AMD's reputation and possible legal jeopardy over false advertising.

Look at it like this: With the Ryzen 3000 series, AMD now includes XFR in the stated Max Boost printed on the box.

https://forums.anandtech.com/threads/cpu-boost-frequency-and-marketing.2567796/post-39880295

 

[Zucker2k]

You are not making any sense. In many cases, the Ryzen 3000 series beats even the 9900k at a lower clockspeed. So how can you possibly say that "Clock speed IS indeed the ultimate arbiter of performance " ?

You're taking things out of context by comparing two separate architectures. Which is faster: A 3900x @4.5GHz or a 3900x @4.6GHz? Never mind the answer. I give up.

 

[PotatoWithEarsOnSide]

Which is faster?
A 3900X that hits 196fps at 4.5GHz or a 3900X that hits 196fps at 4.4GHz?
Makes no difference where the performance comes from. If X is the product of Y and Z, with X being constant, it doesn't matter what Y and Z are individually. They hit max performance out of the box, which ultimately means that Y and Z are irrelevant in the overall context of things.
We could be talking about Intel CPUs here for what it matters.
I'm tired of people hanging onto any little shred of information as a means to burn AMD when it is the bigger picture that counts. The bigger picture is that these CPUs are competitive, and in very many cases the performance leaders. How they got there is neither here nor there in the grand scheme of things. Sure, from a technical point of view it is interesting, but certainly not something with which to scold them.

 

[Shivansps]

The 3200G is more impressive by the day.