Feasible 16 core AMD Threadripper 3000 series CPU crushes the Ryzen 9 3900X and Threadripper 2950X in multi-core workloads

[Redaktion]

Well, well, well. We might have just got our first sight of an AMD Threadripper 3000 series processor, courtesy of a posting on UserBenchmark. The processor, which we expect to see before the end of the year, is compatible with the TR4 socket and existing X399-based motherboards. It leaves the Threadripper 2950X and Ryzen 9 3900X in its wake in multi-core tasks too.

https://www.notebookcheck.net/Feasible-16-core-AMD-Threadripper-3000-series-CPU-crushes-the-Ryzen-9-3900X-and-Threadripper-2950X-in-multi-core-workloads.428020.0.html

[Konstantinos]

NO IT IS NOT FASTER!

ou do realise that the alleged 16 core threadripper has 33% more cores than the Ryzen 9 3900X, right??

So do the math: CPU multi-core score for 3900X * 1.33 = 2200 * 1.33 = 2926

3000 series Threadripper 16-core multi-core score = 2811

WHO IS FASTER AGAIN?

[Alex Alderson]

NO IT IS NOT FASTER!

ou do realise that the alleged 16 core threadripper has 33% more cores than the Ryzen 9 3900X, right??

So do the math: CPU multi-core score for 3900X * 1.33 = 2200 * 1.33 = 2926

3000 series Threadripper 16-core multi-core score = 2811

WHO IS FASTER AGAIN?

Hey ALL CAPS,

Well, it is still the 16 core CPU...

Unless you know some way of adding more cores to the Ryzen 9 3990X?

Faster per core does not make it faster overall.

[Booyah]

"NO IT IS NOT FASTER!

ou do realise that the alleged 16 core threadripper has 33% more cores than the Ryzen 9 3900X, right??

So do the math: CPU multi-core score for 3900X * 1.33 = 2200 * 1.33 = 2926

3000 series Threadripper 16-core multi-core score = 2811

WHO IS FASTER AGAIN?"

The 16 core one, the one that, as you've just shown, doesn't need its actual performance multiplied by 1.33 in an imaginary scenario to beat the other CPU. You don't seem to understanding your own argument that that 16 core argument has 16 cores, and the 3900X is not a 16 core machine, therefore it is not hitting 2926 in the real world, but only as a hypothetical and non existent 16 core model. In the real world, things don't scale perfectly linearly and neither existing CPU is hitting those numbers. An individual core may be slower, but as a whole CPU, as your math and even your chosen criteria for "multi-core speed" has just shown, the 16 core one is faster.

[Konstantinos]

@Booyah and @Alex Alderson

You are missing the point. It is like comparing a i9-9900K 8-core with a Ryzen 3 6-core and reaching the breaking news conclusion: THE i9-9900K IS CRUSHING the Ryzen 3!! It is a complete fault comparison.

[Konstantinos]

Or to show how ridiculous is the comparison I can also reach the conclusion:
Threadripper previous 2 nd generation (2990WX) is crushing next 3rd generation Threadripper 3000 16-core!

What is the point of this comparison? That is why I calculated what is the equivalent performance for the same number of cores...

[Alex Alderson]

Or to show how ridiculous is the comparison I can also reach the conclusion:
Threadripper previous 2 nd generation (2990WX) is crushing next 3rd generation Threadripper 3000 16-core!

What is the point of this comparison? That is why I calculated what is the equivalent performance for the same number of cores...

So, your suggestion is to just compare CPUs with the same core count? Or am I missing the point again?

As Booyah explained, your "equivalent performance for the same number cores" calculation is fantastical, leading you to draw a misinformed conclusion. There is no way to tell whether the Ryzen 9 3900X would be more powerful if it had the same volume of cores as the CPU in this Sharkstooth system; no 16-core Ryzen 9 3900X exists, so it is a pointless comparison.

[Konstantinos]

@Alex  Alderson
Yes, it makes sense to compare two CPUs only if they are in the same class, either technically (same no of cores) or financially (same price point) and same product lines (eg mobile, mainstreem desktop, HEDT). So is Threadripper 3rd Gen 16-core and Ryzen 3900X in the same technical class? No. Are they at the same price point? No. They are not even at the same product category.

You are also confused what to expect from Threadripper 3rd Gen 16-core:
"Better still, the Sharkstooth has 35% fast multi-core speeds than the Ryzen 9 3900X, but almost identical single and quad-core speeds.

If this is an indicator of what is to come with the Threadripper 3000 series, then we could be in for something special."
No we are not in for something special. The Threadripper 3rd Gen 16-core will be almost identical in performance to the already known Ryzen 3950X 16-core, just in a different socket (TR4) since they belong to the exact same architecture (Zen 2).

You copied the original article from "Tom's Hardware" were they knew that they are talking about the same architecture but different core-count: "As expected, the Threadripper sample beats the Ryzen 9 3900X by up to 35% in multi-core workloads."

And added some fanfare of a "Special" product that "Crushes" the 3900X like we are in for a surprise of some sort!

[Alex Alderson]

@Alex  Alderson
Yes, it makes sense to compare two CPUs only if they are in the same class, either technically (same no of cores) or financially (same price point) and same product lines (eg mobile, mainstreem desktop, HEDT). So is Threadripper 3rd Gen 16-core and Ryzen 3900X in the same technical class? No. Are they at the same price point? No. They are not even at the same product category.

You are also confused what to expect from Threadripper 3rd Gen 16-core:
"Better still, the Sharkstooth has 35% fast multi-core speeds than the Ryzen 9 3900X, but almost identical single and quad-core speeds.

If this is an indicator of what is to come with the Threadripper 3000 series, then we could be in for something special."
No we are not in for something special. The Threadripper 3rd Gen 16-core will be almost identical in performance to the already known Ryzen 3950X 16-core, just in a different socket (TR4) since they belong to the exact same architecture (Zen 2).

You copied the original article from "Tom's Hardware" were they knew that they are talking about the same architecture but different core-count: "As expected, the Threadripper sample beats the Ryzen 9 3900X by up to 35% in multi-core workloads."

And added some fanfare of a "Special" product that "Crushes" the 3900X like we are in for a surprise of some sort!

Again, you've drawn conclusions and then worked backwards to reinforce them. Oh, and thanks for the lovely jibes about my ability to read and interpret the UserBenchmark results that I linked in the article. Tom's Hardware must have copied UserBenchmark too, then.  8-|

Best of luck making slanderous remarks against other Editors elsewhere though, because I shall not entertain them.

But please, feel free to waste your time in concocting some more drivel, you have given me a good laugh.

[Konstantinos]

@ Alex Alderson

What I have given you up to now are pure logic arguments. But you are reacting like you are beyond any criticism. And very edgy too.

Not professional at all for an "Editor" with capital "E" as you have written it in the middle of your sentence :-)

[Alex Alderson]

@ Alex Alderson

What I have given you up to now are pure logic arguments. But you are reacting like you are beyond any criticism. And very edgy too.

Not professional at all for an "Editor" with capital "E" as you have written it in the middle of your sentence :-)

Pure logic? No. Delusional? Yes. Using a fictional CPU to argue something is not "pure logic".

No-one, including me, is beyond criticism or reproach. However, everyone is entitled to call out slanderous remarks when they are false and self-serving.

As for your comment about capitalising a proper noun in the middle of a sentence, I despair.

[Konstantinos]

You are writing like a teenager having a tantrum and not as a professional journalist. And NO, "editor" is not a proper noun, it is just your megalomania in combination with lack of skills to perform this job that makes it capital.

If you were more modest you could actually be benefited from my constructive criticism but you are blinded by your inflated ego.

[Alex Alderson]

You are writing like a teenager having a tantrum and not as a professional journalist. And NO, "editor" is not a proper noun, it is just your megalomania in combination with lack of skills to perform this job that makes it capital.

If you were more modest you could actually be benefited from my constructive criticism but you are blinded by your inflated ego.

The arrogance and ego amazes me. Please, debase yourself further.