Core i9-10900K vs. Ryzen 9 3950X Cinebench R15 Comparison Leaked

[btarunr]

Ahead of its launch a leaked ASUS ROG marketing slide reveals Cinebench R15 performance comparisons between the new Intel Core i9-10900K and AMD's current MSDT flagship part, the Ryzen 9 3950X. The graphs also include Intel's previous gen flagship, the i9-9900K, which should provide a reasonable indication of where the new Core i7-10700K performance could land.

In the single-threaded Cinebench R15 test, the Core i9-10900K scores 222 points, while the 3950X scores 213, which is a 4.22% lead for the new Intel flagship over AMD's. The i9-9900K is 2.81% faster than the 3950X in the same test. The landscape changes completely with multi-thread. Armed with 16 cores and 32 threads, the 3950X tests 48.61% faster than the i9-10900K, and a whopping 94.14% faster than the i9-9900K, which means the 3950X should land around 90% (±5%) faster than the i7-10700K. Core i9-10900K vs. Ryzen 9 3900X should make for a fascinating contest.



View at TechPowerUp Main Site

 

[Vayra86]

4% in CS GO. That ain't much.

 

[Jism]

Intel still has that single core lead performance, but i'm sure that gap becomes smaller with the Ryzen 4 series. I mean a well tweaked AMD system with tight timings would proberly be very very close to it. There's so much untapped potential in AMD hardware really, just see the DDR4 timing guide here on Techpowerup.

 

[ZoneDymo]

Intel still has that single core lead performance, but i'm sure that gap becomes smaller with the Ryzen 4 series. I mean a well tweaked AMD system with tight timings would proberly be very very close to it. There's so much untapped potential in AMD hardware really, just see the DDR4 timing guide here on Techpowerup.

Well yeah but only because they do the one thing they can, keep pushing those clocks, no matter the power consumption and the heat (although for the end user especially heat can limit heavily what the processor can actually do)\

But like, who is even suprised by any of these results, cant we just produce these numbers before Intel even releases these?
We know its just the same cpu again but with a few more cores and higher clocks.

 

[Raendor]

Intel still has that single core lead performance, but i'm sure that gap becomes smaller with the Ryzen 4 series. I mean a well tweaked AMD system with tight timings would proberly be very very close to it. There's so much untapped potential in AMD hardware really, just see the DDR4 timing guide here on Techpowerup.

That's what is currently still stopping me from comfortably jumping on the Ryzen platform. I don't want to spend too much time on tweaking memory or many other settings that apparently is very useful for getting the best performance and sometimes even basic stability from Ryzen cpus. Intel seems like a much lower maintenance with easy xmp enabling and very basic approach to OC (if you're going the OC route).

 

[Zubasa]

That's what is currently still stopping me from comfortably jumping on the Ryzen platform. I don't want to spend too much time on tweaking memory or many other settings that apparently is very useful for getting the best performance and sometimes even basic stability from Ryzen cpus. Intel seems like a much lower maintenance with easy xmp enabling and very basic approach to OC (if you're going the OC route).

TBH, you are over-thinking things a bit too much. For Zen2 / 3rd gen Ryzen, the difference is really isn't that huge.
The reason why people often mentions about memory speed, is because Ryzen and Intel 14nm CPUs are within spitting distance of each other.
As a rough estimate, the 3950X's 213 points in single thread, at a 105% it is around 223 points. DDR4 3600 is generally achieveable XMP speed for Zen2.
As usual the bars on marketing graphs makes the difference looks bigger than it is.

https://www.techpowerup.com/review/amd-zen-2-memory-performance-scaling-benchmark/4.html

 

[ratirt]

Intel still has that single core lead performance, but i'm sure that gap becomes smaller with the Ryzen 4 series. I mean a well tweaked AMD system with tight timings would proberly be very very close to it. There's so much untapped potential in AMD hardware really, just see the DDR4 timing guide here on Techpowerup.

Considering the boost speed no shock there. the 3950x on the other hand can go faster if you tweak it a bit.

That's what is currently still stopping me from comfortably jumping on the Ryzen platform. I don't want to spend too much time on tweaking memory or many other settings that apparently is very useful for getting the best performance and sometimes even basic stability from Ryzen cpus. Intel seems like a much lower maintenance with easy xmp enabling and very basic approach to OC (if you're going the OC route).

I didn't have that impression jumping to TR's.
It sound weird because I can assure you, you would tweak your Intel counterpart if you get it. OC it and maybe tweak ram as well. undervolting not to fry the damn chip but still Ryzens require too much attention in your opinion. I didn't have that impression, actually it's been fun for me though and seeing the performance boost, made my day

 

[john_]

If AMD starts winning on R15, I wonder what benchmark they will start pushing. Maybe some single core benchmark from before 2010 that only runs on Intel processors.

 

[harm9963]

One run, neck and neck.

 

[bobsled]

What a joke. Multi core graph seems zero based yet single thread isn't... non zero graphs need to be removed. (in terms of representing data, worldwide)

 

[delshay]

3950x. ..Cinebench multi-thread score 3932. Err, the score almost matches the processor part number.

 

[DonKnotts]

That's what is currently still stopping me from comfortably jumping on the Ryzen platform. I don't want to spend too much time on tweaking memory or many other settings that apparently is very useful for getting the best performance and sometimes even basic stability from Ryzen cpus. Intel seems like a much lower maintenance with easy xmp enabling and very basic approach to OC (if you're going the OC route).

You're staying with Intel because of something you would not even have to do if you didn't want to, and if you did want to, you only had to do once in less than an hours time? I'm sorry, I just understand that logic at all.

 

[Mysteoa]

Doesn't the space between the 9900k and 10900k looks bigger than between the 3950x and 9900k on the single core bench?

 

[Assimilator]

Comparing a 10-core Intel CPU to a 16-core AMD one is somewhat unfair, methinks. Pity ASUS didn't bother to actually be useful and compare to the closer 3900X instead.

That's what is currently still stopping me from comfortably jumping on the Ryzen platform. I don't want to spend too much time on tweaking memory or many other settings that apparently is very useful for getting the best performance and sometimes even basic stability from Ryzen cpus. Intel seems like a much lower maintenance with easy xmp enabling and very basic approach to OC (if you're going the OC route).

I built a system in December 2019 using a Ryzen 3600X, 2x 16GB DDR4-3000, and an X370 motherboard. I didn't have to tweak anything. I have had zero blue screens or other stability issues with it. And I am extremely impressed with the performance. The only tweaking I did was to overclock the memory to DDR-3200 and that was utterly painless. You don't need to mess around with manual overclocking on Ryzen, just enable PBO and the chip does it for you.

You're missing out.

 

[SL2]

Those charts are wrong, both of them. Assuming the numbers are right.

Single core is obviously wrong, but so is the multi core. The 9900K bar should be closer to half the length of the 3950X bar.

I adjusted it ugly style, correct me if I'm wrong here.

 

[Chrispy_]

GOSH WOW, If you take a 7 year old single-threaded test that was optimised with Intel compilers and doesn't stress the limited cache sizes of Skylake++++++++++, 5.3GHz is more than 4.7GHz. Whaddayaknow!

Those charts are wrong, both of them. Assuming the numbers are right.

Single core is obviously wrong, but so is the multi core. The 9900K bar should be closer to half the length of the 3950X bar.

"When you can't win a fair fight, fight dirty."
- Intel Corporation

 

[Kapone33]

That's what is currently still stopping me from comfortably jumping on the Ryzen platform. I don't want to spend too much time on tweaking memory or many other settings that apparently is very useful for getting the best performance and sometimes even basic stability from Ryzen cpus. Intel seems like a much lower maintenance with easy xmp enabling and very basic approach to OC (if you're going the OC route).

Interesting how people look at Ryzen performance as necessary to tweak to work nicely but I can tell you as an AMD system builder that the only issue with Ryzen was some 1st gen memory issues. Today even the the Athlon 3000G fully supports DDR4 3600. Even though the Intel CPUs (10th gen of the same node) OC better there really is no need to OC Ryzen either as the that AMD has built these to factory OC anyway. If you want to talk about stability Zen is pretty stable and has a ton of next gen features that you just cannot get from Intel.

 

[Caring1]

"marketing slide "
Says it all really.

 

[GreiverBlade]

wait for 4XXX .... 3XXX is not the concurrent of the 10XXX series ... (yeah i know ... that's a lot of X ... not a X rated i hope ... )

 

[Psinet]

4% in CS GO. That ain't much.

wot

 

[m4dn355]

I would like to see R11.5 results with huge yellow and orange graphs spanning across A0 landscape mode.

 

[AnarchoPrimitiv]

That's what is currently still stopping me from comfortably jumping on the Ryzen platform. I don't want to spend too much time on tweaking memory or many other settings that apparently is very useful for getting the best performance and sometimes even basic stability from Ryzen cpus. Intel seems like a much lower maintenance with easy xmp enabling and very basic approach to OC (if you're going the OC route).

That's ridiculous, and in all honesty, I'd be willing to be that isn't your real reason, just your pretext. On my 2700x/X470 system, when I first built it, I got into all that tweaking, I was able to push my 2700x to 4.4Ghz all core at only 1.3125v, did the RAM timings calculator, etc etc etc and you know what, I didn't notice a single difference in my daily use, in the way my games played or how quickly my videos rendered, so I said screw it, just left the PBO on, and again, didn't notice any degradation in performance.

While tweaking can be fun, I truly believe that it's a placebo effect that can only be seen in a benchmark,, in other words, I'm sure all that tweaking only gains 1%-2% performance which isn't even perceived by a human being... What I'm saying is, staying away from a better platform because you're worried about having to put in some time to accomplish a 1% performance increase that isn't even noticeable is ridiculous.

 

[Vayra86]

wot

Its a wink to the last bastion Intel has for MSDT... high refresh rate gaming. Most notably, high refresh rate games that rely heavily on ST.

4%. Not much left of that bastion Its nearly margin of error territory and definitely not anything noteworthy any more.

 

[harm9963]

I have a 2700x and x470, as well , but 1.3125v at 4.4 all cores, have to see that please.

 

[jonup]

I have a 2700x and x470, as well , but 1.3125v at 4.4 all cores, have to see that please.

Picture of phase changing cooler incoming...