The 15th gen Intel CPUs are apparently using TSMC 3nm, apart from the i5 processors, which are rumoured to use Intel 20A.
Still a better node than Zen5, if so.
65W vs 105W, 7700X also 105W
Ok, it seems there isn't any confirmation either way about Arrow Lake, apart from the GPU tile.
Yes, like a half-node better. But I wouldn't be surprised to see a clock regression vs. Raptor Lake, given Intel's record when migrating to a new processes and Intel 7 being on its third iteration.
4nm is better than 7nm and 5nm, so yes, I'd expect an improvement in that area. Interested to see actual power numbers, but I would never buy an 8-Core CPU in 2024.
And its the first time Intel has had a node advantage against AMD in many years. And Intel didn't even need a better node in 2022 to beat Zen4 in gaming...
Does anyone want to tell him...?
My point is that sitting on a given node and refining it, is what allows them to hit crazy clock speeds in the first place. So it's both an advantage and a disadvantage for them, relative to AMD.
7800x3d was on a good discount for quite a while. That's what I went with.
intels in house 2nm? Thats not better then what TSMC offer now
AMD piss me off with these delaying shenanigans, I get that for first generation products, its understandable that there is no X3D at launch, but why there isn't one now...
AMD waiting to drop the X3D chip right after. They have finally learned after years of getting their asses kicked by Intel.
Pat Gelsinger has said they won't have process leadership until 18A.
Same here. But if people can get a 5800x3D for ~$200, that's a hell of a deal on a CPU that I think could get you into next generation.
They've been delayed until at least December and by then, the 3D chips will be out.
Leonidas
would have to agree with that.
Many insiders have speculated that max clocks will be a bit lower for Arrow Lake. But if the max power draw decreases by quite a bit (and I mean the 700-series CPU is working at less than 100W) then that would be a worthwhile tradeoff.
What's wrong with an 8-core CPU in 2024? Especially if you're only gaming?
Nothing is wrong with it. 20% of steam users have an 8-core CPU. While 14 core sits at 3% of users.
They also need to not crash. Who in their right mind would buy Intel until this crashing debacle has been resolved?
No, only the top end CPUs. It's complex, and dedicated threads for this here.
Yep, their 700 and 900 series i7 and i9 are degrading, crashing and there are issues even with server chips. It’s kind of crazy to even think about Intel’s high end till this is resolved.
You shouldn’t be upgrading every generation though. I’m on a 7600x and MAY go to the 9800x3d when it goes on sale. I’ve only played one game where my CPU was a problem (Spiderman with RT).
Seems there is a slight regreession in P-Core, but E-Cores are actually getting an increase in clocks, if this is true.
If you're only gaming it is okay, but as someone who has 14 cores in their rig today, I am not ever going backwards, I do more than just gaming on my PC.
The topic was whether finally being on a more advanced process was going to give Intel a huge advantage. I gave Meteor Lake as a counter to the claim that being on a newer process automatically gives Intel a competitive advantage in all aspects. Now you link to a rumour that seems to suggest 5.4 GHz all core clocks for the P Cores, which is below both the maximum for the 14900k and 13900k, as confirmed in TPU's testing. (In any case the 14900K was already at 5.7 GHz for all core clocks, so if 5.7 GHz is the max *boost* frequency for the Ultra 9 285K, we would expect its all core clock to be lower than that of the 14900k).
It should be obvious to anyone that using a much better node than last time is a huge advantage.
Compared to the 13900K, 5.4 would be the same as the 13900K spec, with the max turbo being 100 MHz lower.
That there is a slight P-Core regression? Yes, it seems likely, but the E-Core speeds improved if the leak is true, you seem to be ignoring that).
Because we're debating whether Intel has a much better chance this time due to finally being on a better process. Assuming no hyperthreading and a 15% P-Core performance improvement, then I would expect that to put us back where Raptor Lake was vs. Zen 4 in games (vs. Zen 5), but with more multithreaded performance. So nothing to suggest they are going to crush the eventual X3D part this time.
And that would lead me to get an i5 yet again, since for ~$350 I will get much more MT and probably better gaming performance than a $350 AMD CPU.
AMD will price their products in a way that reflects market demand. If they can continue to hold the $450 price because Intel's process advantage does not result in gaming performance leadership, they will do so.
