News Core i9-13900K Creeps Behind Ryzen 9 7950X In Blender Benchmarks

Toggle sidebar Toggle sidebar
Do these numbers reflect the security mitigations? I've read that the new Ryzen CPUs aren't as badly impacted by those as the Intel ones. All benchmark articles I've read lately don't even mention this.
 
Starry Dynamo said:
Do these numbers reflect the security mitigations? I've read that the new Ryzen CPUs aren't as badly impacted by those as the Intel ones. All benchmark articles I've read lately don't even mention this.
Click to expand...
These are not (more than likely) using final BIOS revisions as well, so don't trust these numbers blindly. Plus, not much other information was given on the platform either.

Other than those caveats, this is as close as I'm expecting the general 13K vs Ry7K to go down, so maybe the 13900K will reach partity with the 7950X using a tad more power across and gaming-wise, all of the Intel lineup will be a hair faster than Ry7K, except the 5800X3D and the memes will be glorious.

Regards.
 
  • Like
Reactions: KyaraM
hotaru251 said:
i'd actually be shocked if Intel DIDN'T beat amd in blender...given its got more cores.
Click to expand...
Look at it this way: they added 8 more E-cores while only increasing the power by 12W (241 to 253 was it?). Unless they go above that power (253W), the clocks across the board can't be higher than Alder Lake's, so it'll be interesting to see the clocks behaviour there. For an unlimited Raptor Lake, holy cow, I'm expecting some really high wattage and impresive clocks.

Regards.
 
No matter what the data says, somehow the articles always find a way to talk about Intel's chips winning.

When the quad-core i5's got smoked by Ryzen 5's, the articles were about how Intel's i9's beat Ryzen R9's. Now that the Ryzen chips hold the performance crown, most of the discussion is about how they're overpriced and how the i5's are faster. Just spin the articles both ways sometime.
 
  • Like
Reactions: cc2onouui and SunMaster
dalauder said:
No matter what the data says, somehow the articles always find a way to talk about Intel's chips winning.

When the quad-core i5's got smoked by Ryzen 5's, the articles were about how Intel's i9's beat Ryzen R9's. Now that the Ryzen chips hold the performance crown, most of the discussion is about how they're overpriced and how the i5's are faster. Just spin the articles both ways sometime.
Click to expand...
Heh, that is true. A better click-grabbing title would've been "AMD keeps the top while Intel can't catch it in Blender".

I'm sure a lot of... Hm... Very civilized and corteous Intel fans would have made their opinions clear on the matter xD

Regards.
 
  • Like
Reactions: RodroX and cc2onouui
BeedooX said:
Nice! I think it's going to be the 7950X running in ECO mode, and a top model multi-chip RDNA3 GPU.
Click to expand...
It will be interesting to see the eco mode 7950x vs the 13900.
7950x may win in productivity and lose in gaming as the previous gens did.
But not that interesting since I like to power plan limit my Intel cpu in windows when I value silence over unneeded performance already.
If you have an efficient volt/ frequency curve set up you can just limit the upper clocks. Persistent over reboots, 0 additional overhead and you can change max clocks as fast as you can choose a different power plan.

Edit: The admin command/terminal/powershell line to get Windows Power Plan to show the max clock option is:
powercfg -attributes SUB_PROCESSOR 75b0ae3f-bce0-45a7-8c89-c9611c25e100 -ATTRIB_HIDE
and if you have Alder Lake with e-cores active you will also need:
powercfg -attributes SUB_PROCESSOR 75b0ae3f-bce0-45a7-8c89-c9611c25e101 -ATTRIB_HIDE
For some reason Windows mixes up which one is the power efficient cores and you need both types just for Alder (and Raptor). Won't hurt if you open both options for other arches, but it will clutter your processor power management section if you add a useless option. These won't raise clocks above your bios limit, but they can reduce them and the corresponding power consumption. Also I don't get a perfect correlation between my p-core entry and what the result is, and it even varies between bios versions and overclocks, but it is proportional. I just load the cpu, check the clocks/power in hwinfo and adjust until I get what I want.

It's been my favorite power saving method as of late and it seems like it would be handy for Zen4 temp control, but I have no idea if it works with them. I like to set up a new balanced power plan and name it whatever max frequency I've chosen for p-cores.
 
Last edited:
Soon enough well need some small nuclear reactor to power theses rigs AMD, Intel, Nvida ... And yes i would wait a few weeks to see how it goes, there will be surely some bios quirks to correct ??
 
  • Like
Reactions: cc2onouui
Starry Dynamo said:
Do these numbers reflect the security mitigations? I've read that the new Ryzen CPUs aren't as badly impacted by those as the Intel ones. All benchmark articles I've read lately don't even mention this.
Click to expand...

I don't think we know yet if either of the CPUs are vulnerable to existing attacks and/or if they have been mitigated at the hardware level. At this point, if you are on the fence I would say wait until after the new year when user reviews and security reviews are completed.
 
  • Like
Reactions: KyaraM and purple_dragon
So these benches got me curious enough to run the Blender 3.3 bench on my system and my 12700k got about 333, stock. My AMD 6800 got 1486, totally stock and it seemed like exactly the same test so blender seems like a bad cpu bench if cpus are obsolete running it? By comparison their listed 6800 av got 1542(must be oc), Intel Arc A770 gets 1624, and a 3080 like mine gets 5000. Really can't call this a "productivity" case if a cpu the same price as a gpu gets 607.53/5053.93 or 12% the performance.

But it is a relative performance comparison if nothing else.
 
Starry Dynamo said:
Do these numbers reflect the security mitigations? I've read that the new Ryzen CPUs aren't as badly impacted by those as the Intel ones. All benchmark articles I've read lately don't even mention this.
Click to expand...
AFAIK AMD still hasn't done anything about SQUIP and the mitigation for that is disabling SMT. I don't think that was done for these benches.
Hopefully I'm wrong and they have released a less consequential mitigation.
 
  • Like
Reactions: KyaraM
dalauder said:
No matter what the data says, somehow the articles always find a way to talk about Intel's chips winning.

When the quad-core i5's got smoked by Ryzen 5's, the articles were about how Intel's i9's beat Ryzen R9's. Now that the Ryzen chips hold the performance crown, most of the discussion is about how they're overpriced and how the i5's are faster. Just spin the articles both ways sometime.
Click to expand...
What? The main point of the article is about how Intel closed the multi-core performance gap based on this benchmark; on an unoptimized system, and will probably shave off a couple more points when they do get peoper BIOS support. Nothing more, nothing less. And at considerably lower price point, too. Once again, when AMD does it, it's all fine and dandy, when Intel does, it's worth nothing. You do want AMD to be the only one on the market, do you? Even when you claim you don't. Be glad they are so close, it's beneficial for all when there is competition. As long as you don't glorify either, nobody will say anything.
 
BX4096 said:
Or temperatures. Some of these new products can be marketed as room heaters for the cold winter ahead.
Click to expand...
Perhaps it's counterintuitive, but chip temp has nothing to do with heating a room. Wattage is the main factor. I'll go back to my previous (ludicrous) example: a 5w raspberry pi running at 100C isn't going to change room temp. A 10kw+ cerberus wse2 running at 60C will heat your whole neighborhood.
 
  • Like
Reactions: JamesJones44, RodroX and -Fran-
jp7189 said:
Perhaps it's counterintuitive, but chip temp has nothing to do with heating a room. Wattage is the main factor. I'll go back to my previous (ludicrous) example: a 5w raspberry pi running at 100C isn't going to change room temp. A 10kw+ cerberus wse2 running at 60C will heat your whole neighborhood.
Click to expand...
Yeah it's hyperbole, but you are still going to feel 230-40W as an increase in room temperature. (Let alone the GPU and the rest of the system)
Not as much as a 2000W space heater but still.
 
KyaraM said:
What? The main point of the article is about how Intel closed the multi-core performance gap based on this benchmark; on an unoptimized system, and will probably shave off a couple more points when they do get peoper BIOS support. Nothing more, nothing less. And at considerably lower price point, too. Once again, when AMD does it, it's all fine and dandy, when Intel does, it's worth nothing. You do want AMD to be the only one on the market, do you? Even when you claim you don't. Be glad they are so close, it's beneficial for all when there is competition. As long as you don't glorify either, nobody will say anything.
Click to expand...
You say that like Intel is the one in a bad spot, but it's the other way around. I want parity and that includes articles with clickbait written both ways.
 
There has been too much comparing of one CPU to another...when one might be a year old.

So, I guess it's official - in CURRENT GENERATATION (and granting Intel their generation which has yet to be released), AMD is the NEW KING as far as the tests are able to show?
 
You must log in or register to reply here.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom