Intel Core Ultra 7 265KF and Core Ultra 5 245K benchmarks leaked
- Thread starter marquimvfs
- Start date
- Rumor Analysis Platform
drotahorror
Member
nvm clicked wrong one
Last edited:
Insane Metal
Member
8 cores and somewhat low clock rate
OT, kinda feels the same as Zen 5... if you compare the 265K to the 14700K and the 245K to the 14400K... now let's see the 285K which would be comparable to the 14900K.
Last edited:
marquimvfs
Member
It sure is a great upgrade. It is only a sidegrade compared to Core i7-14700KF, that does 3.004 points single and 19.583 multi. But there are others aspects to be considered, like power consumption, heat, efficiency and if it isn't affected by any silicon problems like the previous gen.
Last edited:
kiphalfton
Member
Is "Core Ultra" replacing "Core i" naming convention?
Seems needlessly confusing.
Seems needlessly confusing.
ShaiKhulud1989
Gold Member
I've tried to figure out for one solid minute what chips are we even talking about here.
Great naming scheme. Terrific even.
Great naming scheme. Terrific even.
Skifi28
Member
It almost seems like they want to distance themselves from the previous line. I wonder why
.
Silver Wattle
Gold Member
Hopefully they fix the stability/oxidation issues, Intel desperately needs a win.
I personally would not risk this gen from them though as they have behaved abhorrently.
I personally would not risk this gen from them though as they have behaved abhorrently.
Leonidas
AMD's Dogma: ARyzen (No Intel inside)
Leonidas
AMD's Dogma: ARyzen (No Intel inside)
The i was replaced with Ultra, and now the model number is 3 numbers instead of 5 numbers that's pretty much it...
Ultra 9 285K replaces i9 14900K
Ultra 7 275K replaces i7 14700K
Ultra 5 245K replaces i5 14600K
marquimvfs
Member
Possible, even probable. It looks like they're trying to improve the single core performance at every cost, what will for sure helps in gaming. The question that's in the air is what is the efficiency of those mew chips and the price.
Also, why is Geekbench useless? I was always under impression that it simulates everyday tasks, not some obscure tools.
Last edited:
Gameplay Gods Bless
Member
Multithreaded is impressive considering they aren't using hyper threading. Single threaded is disappointing as are the clocks TSMC N3 is enabling.
Gameplay Gods Bless
Member
I've heard Geekbench multithreaded results are very unreliable but ST is solid
Celcius
°Temp. member
Yeah I usually use cinebench when I run benchmarks, I wish they had used it here instead of GeekBench
Gameplay Gods Bless
Member
Clocks are going backwards due to TSMC. IPC ain't going up enough to provide a big bump. Should be better in games than Zen 5 but Zen 5 X3D is coming.
Leonidas
AMD's Dogma: ARyzen (No Intel inside)
With a purported 100w drop for the i9 and rumors of AMD increasing 9700X TDP, Arrow Lake could even beat Zen5 in efficiency.
I'll still get a clock bump, as I'm on a mildly overclocked 13600K. And if I jump up to the i7, I'll be getting a decent clock bump + the improved IPC + 6 more cores... and the E-cores are clocked half a Ghz higher than what I'm at today.
Useless speculating Zen5 3D, we don't know when it launches, we don't know if it will be out when Arrow Lake arrives when I make my purchase, and it could be another dissapointing 5% increase over the last gen 3D.
And top gaming performance never mattered to me. I'll take 90% the top gaming performance if I can get 30% better MT and pay less.
Last edited:
marquimvfs
Member
Well, we will see. Even with a 100w tdp it coud be above Zen. If they manage to be below Zen 5, it'll be a massive improvement for them. Also why Geekbench is useless? I was always under impression that it simulates everyday tasks, not some obscure tools.
Leonidas
AMD's Dogma: ARyzen (No Intel inside)
Its a 2 node improvement from Raptor Lake so I wouldn't be too surprised.
After seeing Zen5's 15% IPC lead to 3% gaming performance, I am not really concerned with any of these early benchmarks. I'll care when we see some gaming numbers...
marquimvfs
Member
Right, then it's a personal matter? Because I thought that you had something to back up that claim.
Leonidas
AMD's Dogma: ARyzen (No Intel inside)
Geekbench is useless for gaming. This is a gaming forum.
marquimvfs
Member
Yeah. Just like you CPU being efficient at doing nothing is useless for this forum.
Leonidas
AMD's Dogma: ARyzen (No Intel inside)
Web browsing and posting right now, my 13600K is using less power than your CPU at this very moment
Very useful when it comes to posting here, and browisng other sites, and reading e-books, something I do a lot of...
Last edited:
marquimvfs
Member
Is it? What CPU im using?Web browsing and posting right now, my 13600K is using less power than your CPU at this moment
Leonidas
AMD's Dogma: ARyzen (No Intel inside)
Don't care
marquimvfs
Member
By pure efficiency matter, the phone I'm posting from is better than your CPU.
Leonidas
AMD's Dogma: ARyzen (No Intel inside)
your speed efficiency of typing is worse.
marquimvfs
Member
Cannot disagree. Sausage fingers sucks.
Leonidas
AMD's Dogma: ARyzen (No Intel inside)
and i can easily switch between tabs on a much larger screen, whilst reading ebook on the side as well, all while my cores are ilding in a low power state.
marquimvfs
Member
PC's are practical at multitask compared to smartphones. News at 11. But, like you said, that's a gaming forum, it doesn't matter.
Leonidas
AMD's Dogma: ARyzen (No Intel inside)
lots of users browse from their desktops, so it does matter if one CPU is more efficient in a low power state compared to another, if people really care about efficiency...
TheBreezyBB
Member
Web browsing and posting right now, my 13600K is using less power than your CPU at this very moment
You win.
Reizo Ryuu
Gold Member
seems like the ultra 7 scores twice as much as my i9 109k, so by the time I upgrade years from now, performance/power usage ratio should be even better.
thebigmanjosh
Member
NPU is the biggest addition to Intels processors this cycle, but outside don't expect more than the standard power bump. Big gains will come with RibbonFET and backside power.
Intel making it to 2026 is an entirely different question
Intel making it to 2026 is an entirely different question
Dr.D00p
Member
Still in a CPU performance plateau...prioritising power, heat and efficiency are going to mean quite a few boring, small incremental upgrades over the next few years.
Ever increasing GPU performance Is going to leave bottlenecks all over the place. Only playing at 4K or higher, is going to push them.
Visual fidelity will greatly increase but framerates will remain stuck at current levels for the foreseeable future.
Ever increasing GPU performance Is going to leave bottlenecks all over the place. Only playing at 4K or higher, is going to push them.
Visual fidelity will greatly increase but framerates will remain stuck at current levels for the foreseeable future.
marquimvfs
Member
Well, NPUs are used to accelerate Neural Network processing. It have several use cases, and one of them is the Generative IA and voice recognition (the gimmick you're talking about). I think that, once it becomes more popular, it could be broadly used in gaming, but now, it isn't. It could be used in upscaling, accelerate physics calculations and so on.
Last edited:
winjer
Member
Though not confirmed, Arrow Lake on the desktop is using Intel 20A. And on the Laptop, it's using TSMC N3.
The reason for these lower clocks might be to avoid the degradation they got with 13 and 14th gen.
winjer
Member
It's mostly for things like Copilot. Tough there is also Microsoft Auto Super Resolution.
For games, we'll be using the Tensor cores in the GPUs, for things like DLSS, XeSS, PSSR, etc.
Dr.D00p
Member
There's no doubt,IMO, that the Arrow Lake we are getting isn't going to be the Arrow Lake we could have had without all the recent trouble.
And yet despite it all my natural inclination is to still build a top end Intel PC over a top end AMD one, mainly because I think Intel is still inherently more stable when it comes to RAM speeds.
Multi threaded performance only matters when Intel is losing in gaming.
winjer
Member
One of the Agesas released last year, improved memory compatibility for AM5 a lot. To the point that overclockers can now use 8000MT/s.
Though at higher speeds, Intel still has a slight edge.
But for the common, 6000MT/s, AMD and Intel are almost on par. As long as both are using QVL dimms, with EXPO/XMP.
On the other hand, the X3D parts have so much cache that memory speeds and tweaking become much less important.
So for people that don't care about spending hours tweaking memory, using an X3D part is the easiest and fastest way of getting top performance.
Von Hugh
Member
blastprocessor
The Amiga Brotherhood
This will use far higher wattage than Ryzen 9000 it seems.
Last edited:
GreatnessRD
Member
Fuck all this. We wanna see it used with CAMM2.