Core i9 12900K beats Ryzen 9 5950X in Ashes of Singularity by 40%.

You all be acting crazy brainwashed with brand loyalty to a faceless institution/company.

We should expect every new CPU/GPU to give better performance especially with PC gaming as the life blood for people who have a computer at home.
I want better performance numbers every new release.AMD/Intel/Nvida giving us crap forever with small increases and charging a crap ton of cash.
I am also a sucker who buys the new GPU/CPU all the time,so I have no reason to complain,just wishing like a dumbass for better,when I know it will never arrive to where I am fully feed to quash the need to buy the next one.

 

You sound like me years back. I use the term "MILKING WALL THEORY." If they keep pushing technology faster and faster they will likely back fire and hit a wall or limit. Computer tech now isn't booming like it did in the year 2000. We humans are pushing the limits with what we have now and what we can use. We will always hit a wall but you gotta have a new wall to start far away again and go again and again and see how long that path will travel until the next wall shows up.

 

Yup, that is pretty important factor here - DDR5 chipsets from AMD have not been released, and, Windows 11 is still a distance off from being used in any form of testing.

 

Intel brag beat AMD over ONE freaking game!? STFU Intel.

 

If you mean that ddr5 was used on this particular bench, I'm just telling you that there is no way ddr5 that is currently on has 100ns latency, I guess the E cores have been disabled and the other cores have been pushed with some kind of exotic water cooling.

 

Same bullshit like with Rocket Lake https://www.notebookcheck.net/The-I...s-of-the-Singularity-benchmarks.509533.0.html
Wait for the reviews

 

Maybe it's time for the X86 instruction set to be re-evaluated.
The M1 chip shows that a refresh in the architecture can get more speed using lower clocks and less power.
If the X86 platform can be evolved over time to maintain compatibility, maybe we can get a little more boost.
In the end though, integration of the CPU into the GPU will probably make more sense.

 

Define 'speed'.

Ask a bunch of marketing executives to define speed and you'll get a variety of answers, which is part of the problem.

A classic: "This generation has 2X power than the last generation".

And the nasty truth that marketing executives rationalize over is they meant 'their' generation, and not the competition.

 

"speed" is usually marketing execs best case scenario. Very possible that the 12900k is 38% faster than 5950x in AOTS, but it may be behind in other benches. Only reviews will reveal full picture.

Remember bulldozer (FX-8350)? It was thrashed pretty bad by Sandy Bridge in majority of benches, but in 7-zip it beat it.

 

Yup, could very well be stock CPU.
Not stock cooling though :

https://www.intel.com/content/www/us/en/products/docs/processors/core/cryo-cooling.html

"Intel CryoCooling Tech" sort of says it all really.

I suppose it'll eventually come out in the very tiny small print.....

 

Yep. AMD coming out with all these Ryzen CPUs kickstarted everything. I still don't understand why some people wanted AMD to dominate and kill Intel off. The best possible scenario that could have happened is happening.

We're getting great CPUs from both AMD and Intel and I hope they keep pushing themselves.

Remember people, buy what's best for your budget and usage scenario. Ignore the brand. They're both only here to make money, not to make anyone happy on some personal level.

 

There was a recent patch that enabled

In a way, you are very right.
As Jim Keller recently said, "X86 is probably the worst CPU arquitecture in the planet"

But the problem is that X86 is a very old ISA, with a ton of old instructions, that are still being used by companies and users all around the world.
This is an advantage, because it means that many are locked to the X86 ISA.
But it also means that Intel and AMD have to maintain support for these instructions.

We must also consider that X86 started as a CISC ISA, at a time when every one was still debating what was best. CISC vs RISC.
Eventually, it was proven that RISC was the best way. So with the P6 arch, Intel changed X86 to RISC.
But this was done by creating a big decode unit, that converted X86 instructions into uOps.

Because of these two factors, todays X86 CPUs, have a bigger decode unit than ARM.
This means more area, more transistors, and more power being spent by a X86 CPU.

But keep in mind that ARM has been adding more instructions to it's feature set.
These ad benefits, like performance improvements with certain types of calculations.
But it also means that ARM is becoming more "bloated"

 

@metagamer Except that there is no competition in pricing. Last year it was all about Zen 3, with Intel absent in Q4, leaving AMD to up their prices to the tune of record sales. This year we have a reversal, with AMD leaving Q4 all to Intel and their Alder Lake, so that Intel is under no pressure to do anything about their price. Sure, AMD will probably drop the MSRP on Zen 3, but who in their right mind will buy into a dead-end platform at its literal death bed? We're getting fleeced by both of them. The only way we win is if they both undercut each other while outperforming previous generations. Doesn't seem like that's going to happen.

 

Maybe there is (simply speaking) more profit in not releasing something until the competition beats you...which kinda makes you think why they have not done anything until 'now'.

 

Supposed result on R23

 

How exactly does this prove this benchmark wasn't done on stock cooling?

 

As someone who keeps background tasks to a bare minimum, even I acknowledge this technology path doesn't suck. So far, it seems to me Intel is charging proportionately less for the little cores than they used to charge for HT. The little cores are better than HT. Not everything needs advanced instructions. Most tasks don't scale up that well, if at all. The big complex cores are not efficient when they're being pushed hard yet not fully utilized. I would soooo much rather spend less for more overall performance and lower heat output.I just don't understand this mindset of people who just want more big cores. If your computer earns you money, buy a workstation or server CPU, where there either won't be little cores or relatively few.
I say it over and over again: CPUs are generally better for multitasking, not parallelization. You don't need homogeneity in your architecture if there isn't homogeneity in your workload.

Might not be crap for AMD. Maybe the game just peaks in performance after a certain core count.
Remember too that for games, each individual thread must be synced with a master thread. AMD's core cluster approach is pretty slow with inter-core communication, which can really drag down performance every time threads need to sync.

 

A 40% uplift over AMD seems to good to be true or some sort of freak result. Not to mention that there are more interesting games to bench than Ashes, like Cyberpunk.

Still, competition is always good, even if they don´t result in lower prices, like mackintosh as posted before me...

 

That explains alot .

However all should be happy because competition is real this time and we can hope for gains on PERFORMANCE/$ ratios. Which is imo most important thing , driving technology .
Remmber QUAD CORE STAGNATION ?

 

40% faster, 200% power draw...