Flagship AMD Ryzen Threadripper 16-core Chip Appears on GeekBench

[btarunr]

Apparently, AMD's nomenclature of its flagship Ryzen Threadripper won't look like years-of-birth of today's gamers after all. The flagship 16-core part will bear the model name Ryzen Threadripper 1950X (and not the previously-reported 1998X). This chip was put through GeekBench 4.1.0, on an ASRock X399 Professional Gaming, paired with 16 GB of DDR4-2133 MHz memory. Whether it's dual-channel or quad-channel, is not known at this point. What is known, however, is that 2133 MHz isn't the best memory frequency for Ryzen; and paired with quad-channel DDR4-3200, one could expect the best possible performance. The 1950X was clocked at 3.40 GHz for this test, which probably is its final nominal clock speed, after all.

The Ryzen Threadripper 1950X sample scored 4,167 single-thread performance, and 24,539 points in multi-threaded performance. To put these numbers into perspective, an Intel Xeon E5-2697A v4 16-core/32-thread processor based on the "Broadwell" architecture scores 30,450 points in multi-threaded performance, even if single-thread performance is as low as 3,651 points. Perhaps the memory setup or SMT isn't optimally set for the Threadripper chip. Among the other Threadripper SKUs AMD plans to launch on July 27 are the 12-core/24-thread Ryzen Threadripper 1920X and 1920 (non-X).



View at TechPowerUp Main Site

 

[[XC] Oj101]

No idea about the accuracy of these, but there are accurate Cinebench results out.

 

[silentbogo]

The Ryzen Threadripper 1950X sample scored 4,167 single-thread performance, and 24,539 points in multi-threaded performance. To put these numbers into perspective, an Intel Xeon E5-2697A v4 16-core/32-thread processor based on the "Broadwell" architecture scores 30,450 points in multi-threaded performance, even if single-thread performance is as low as 3,651 points. Perhaps the memory setup or SMT isn't optimally set for the Threadripper chip.

No, there is only one reason: Geekbench.
It's results may be inconsistent even between consecutive runs of the same bench on the same system.

Guess which one was made with real-time priority, and which one was made while playing Spider Solitaire at Normal priority?

That's 5% error margin per physical core. In 4.0 it was a lot more (my scores could vary from 4300 to 4900 under the same conditions)

 

[Deleted member 172152]

No, there is only one reason: Geekbench.
It's results may be inconsistent even between consecutive runs of the same bench on the same system.

Guess which one was made with real-time priority, and which one was made while playing Spider Solitaire at Normal priority?
View attachment 89055 View attachment 89056 View attachment 89057
That's 5% error margin per physical core. In 4.0 it was a lot more (my scores could vary from 4300 to 4900 under the same conditions)

And how can threadripper have much worse multicore and much better single core than that xeon? Also, 3.4ghz seems a bit low. If legit it probably was to test the mobo since asrock didn't want their name in one of these articles.

I'm keeping an open mind untill the threadripper cpu's go on sale.

 

[nemesis.ie]

The Xeon is also nearly $3000 or something, I suspect TR will be a "tad" cheaper ...

 

[Boosnie]

Very Low Attendibility Test.

Also, on multicore operation, memory bandwidth, speed and availability is crucial [see]

Also, free shipping.

 

[Boosnie]

And how can threadripper have much worse multicore and much better single core than that xeon? Also, 3.4ghz seems a bit low. If legit it probably was to test the mobo since asrock didn't want their name in one of these articles.

I'm keeping an open mind untill the threadripper cpu's go on sale.

Xeon was tested with 130GB ram @ +260Mhz.

 

[jboydgolfer]

No idea about the accuracy of these, , but there are accurate Cinebench results out.

you HAVE to laugh at the wording of that statement

 

[silentbogo]

And how can threadripper have much worse multicore and much better single core than that xeon?

That's that error margin I mentioned. With 16 cores and unreliable testing you can get anywhere between 30 and 40% discrepancy! Crazy stuff.
The first time I tried Geekbench, was the older 3.xx version on my X5650 and it had a much worse MT score with 4GHz overclock, than with 3.2GHz, though the single-threaded score did increase (also unproportionally to OC%).

 

[Dave65]

People will freak out over this, time to back away, slowly and take a breath and wait for more info

 

[uuuaaaaaa]

These ThreadRipper results do not seem legit:

3600MHz R7 1800X
Single-Core Score: 5219
Multi-Core Score : 28881

http://browser.primatelabs.com/v4/cpu/2689532

 

[_JP_]

That isn't a very even comparison, the Win10 version isn't even the same

 

[[XC] Oj101]

you HAVE to laugh at the wording of that statement

I'm missing something, please fill me in?

 

[Deleted member 172152]

These results do not seem legit:

3600MHz R7 1800X
Single-Core Score: 5219
Multi-Core Score : 28881

http://browser.primatelabs.com/v4/cpu/2689532

Maybe it's actually a 1700x or something.

 

[uuuaaaaaa]

Maybe it's actually a 1700x or something.

I meant the ThreadRipper results not being legit x)

 

[EarthDog]

People will freak out over this, time to back away, slowly and take a breath and wait for more info

lol, no doubt... same thing if the results smoked the intel, lolol!

 

[Vlada011]

This is not real result, because Ryzen 7 is 50% stronger than i7-5820K and this is result of
i7-5820K 4.2GHz and 4.0GHz Cache Frequency

That mean Threadripper 16 core is weaker than Ryzen 7 1800X.
That's not possible.

 

[Gasaraki]

LOL.

*goes make some popcorn*

 

[Patriot]

Arguing over if this was a real run on a terrible benchmark is such a waste of time. geekbench's variability and core scaling make it truly useless for ...anything. Run linux-bench on both setups and give us something to compare.

 

[Vlada011]

Wait, AMD just launch CPU with 8 core kill all processors with 6 cores in any application and now you want to say that CPU with x2 cores is somehow same power as i7-5820K in Geekbench.
That's not software from 2009. That's newest software with support for Intels and AMD Ryzen and valid for testing PC processor or cell phone processor and how CPU with 16 core competitior to i9-7900X could give same score as i7-5820K. I expect to Intel i9-7900X give 80% better result than i7-5820K in any situation.
4.5GHz clock, 4 cores more and newer architecture if AMD is not 50% stronger than i7-5820K I don't know why they expect.
I compare with 4.0.3 and last 4.1.0 and results of my CPU are better in last Geekbench and he is closer to AMD in last than previous version of software.
i7-5960X have over 30.000 multi score, even 35.000

We could say nothing is valid, only games are important, CINEBENCH, Geekbench, SiSoftware, Fritzchess, ... never mind, but that's not true... They measure CPU power and Geekbench know to recognize strong processor.
Intel Coffee Lake when show up will give single at least 6000 and multi at least 30.000 with his 6 cores and 95W. i9-7900X 45-46.000 multi my rough estimate.

i7-5820K/i7-5930K are killers on 4.5GHz imagine CPU with 10 cores on Intel Turbo Boost 3.0 and new architecture with better memory controller.

 

[uuuaaaaaa]

Wait, AMD just launch CPU with 8 core kill all processors with 6 cores in any application and now you want to say that CPU with x2 cores is somehow same power as i7-5820K in Geekbench.
That's not software from 2009. That's newest software with support for Intels and AMD Ryzen and valid for testing PC processor or cell phone processor and how CPU with 16 core competitior to i9-7900X could give same score as i7-5820K. I expect to Intel i9-7900X give 80% better result than i7-5820K in any situation.
4.5GHz clock, 4 cores more and newer architecture if AMD is not 50% stronger than i7-5820K I don't know why they expect.
I compare with 4.0.3 and last 4.1.0 and results of my CPU are better in last Geekbench and he is closer to AMD in last than previous version of software.
i7-5960X have over 30.000 multi score, even 35.000

We could say nothing is valid, only games are important, CINEBENCH, Geekbench, SiSoftware, Fritzchess, ... never mind, but that's not true... They measure CPU power and Geekbench know to recognize strong processor.
Intel Coffee Lake when show up will give single at least 6000 and multi at least 30.000 with his 6 cores and 95W. i9-7900X 45-46.000 multi my rough estimate.

i7-5820K/i7-5930K are killers on 4.5GHz imagine CPU with 10 cores on Intel Turbo Boost 3.0 and new architecture with better memory controller.

This geek bench score is not legit, I mean, a stock R7 1800x beats it...

3600MHz R7 1800X
Single-Core Score: 5219
Multi-Core Score : 28881

http://browser.primatelabs.com/v4/cpu/2689532

 

[Vlada011]

That's what I thought immediately.
If you ask me only Intel i9 18C/36T worth 1999$ should beat Threadripper.
But Intel was not fair with price of i9 10 core and 28 PCI-E lanes of i7.
Intel i7 8 core Turbo Boost 3.0 4.5GHz should be best for gamers, but people dont like paste and 28 PCI-E lanes.
I'm really dissapointed with paste. Because I compare i7-3770K and i7-5820K and i7-5820K is easier to cooldown with great coolers.

Price difference between 16 cores and 10 cores same as 10 vs 8 core
Only because Intel want their 1000$ no matter on what.
He is higher class and 44 PCI-E lanes, only to pull 1000$ instead of half of that.
But if AMD beat them for 850$ price of i9 will go down.

 

[Vlada011]

But look Cinebench results... almost as Xeon E5 2699...
That Xeon cost almost 4000$. I hope memory result will be similar at least to X99 platform.

How long we could play games and change Radeons with such processor??? 5 years?
He need to work on 4.2GHz at least. That's necessary.