With the release of Alder Lake less than a week away and the "Lakes" thread having turned into a nightmare to navigate I thought it might be a good time to start a discussion thread solely for Alder Lake.
If your "serious work" is a CPU render farm. Sure. But not all "serious work" is.
It seems like Intel should consider an all E-Core server chip for those kinds of loads.
S is the high end full power CPUs, you are talking about U for mobile or T for low power desktop.Or just the S variants that will be used for laptops.
Yeah an i9 12900k 24t beating a 32t 5950x on MT and murdering it on ST while also costing less is ''not much better''
You should be fine, my understanding is Z-height drops from 7.3mm to 6.53mm, so for spring loaded retention mechanisms it means some drop in pressure. My immediate thought for an ad-hoc solution would be some sort of mounting washers to reduce the height difference.Hm, as someone that went with an ASUS board to use with an LGA115x/12xx cooler (EK Quantum Velocity block), I'm now curious if I should've been more patient and waited for the LGA1700 block. Well, I guess I'll find out the hard way!
When you are limited to 65 W of power for 6 performance cores on the lower binned chips (the 12600 has the higher binned chips), the base frequency cannot be very high. For example, the 6 core 11400F base frequency is 2.6 GHz.Hm, why only 2.5ghz base CPU clock for i5 12400F.
I will take the 5950x.
Let me try, (8/6)*37=45.3W (Prof. Koch my partial differential equations instructor would be proud!), but still a way to go from 3.2GHz to 3.7GHz, even with Intel 7. There is room for debate on this one.
As for your first example, 2.9GHz to 3.7GHz is also a stretch at 48W.
I stand by my adjective. 48W for 8 Gracemonts @3.7GHz is "great" from a power/performance point of view.
Let's see where I put it on one subjective scale of adjectives!
Sucks
Bad
Not too bad
Meh
Kind of meh
Okay
Decent
Pretty good
Good
Really Good
Great <========This is the one I got scolded for using and made planets collide!
Really great
Fantastic
Astounding
Unbelievable
and the top of the chart...
Hulk SMASH!!!
You think it will have avx512 enabled by default because it doesn't have any e-cores? Could be.Does the 12700k significantly throttle in Prime95 v30.3 due to Intel stock power limits?
Yeah, however trivial it may seem now, affordable solid state storage was literally the greatest thing that's happened to consumer computing in the past 2 decades - my 2 cents.Well it's the only thing you can do to improve performance without really doing anything, it's like putting an ssd into an old laptop, or to stay closer to how cache works like putting an optane disk in a system.
I just hope it's no too complicated and expensive to produce because if they raise their prices once again it's not going to look good at all.
Based on just these figures:
12900k gaming: Yay
12900k serious work: Nay
BTW, what about Board (and DDR5) cost?
They are just running the benchmarks with proper power settings, resulting in 20W per core in full load compared to the 19W of the 5900x and the 14w of the 5950xSome of those numbers seem off to me, I haven't gone through them all but a quick glance at Blender shows something is off. The 12900K shouldn't be slower than the 5900X and it most definitively shouldn't be only 21% faster than a 12600K in a fully MT workload like Blender.
Tomshardware has it in between a 5900X and 5950X in Blender with a ~38% advantage over a 12600K:
Techpowerup has it slightly ahead of a 5950X with a ~36% advantage over a 12600K:
My guess is that there is some kind of thermal throttling going on with the Guru3D setup in the heavily threaded benchmarks like Blender, and the poor 21% scaling over the 12600K (while other sites show 36-38% gains) should be a telltale sign that something isn't right.
All 8 performance cores were overclocked to 5.3 GHz, we leave the E cores at defaults settings. Intel Core i9-12900K at 5.3 GHz outperformed the AMD Ryzen 9 5950X by a small margin in multi-core tests. However it also needs to be stated that a Ryzen 9 5950X easily consumes almost 33% less power than the Alder Lake CPU.
I made a mistake in my initial opinion that 48W for 8 Gracemont cores @3.7GHz was "great" considering Skylake level IPC for Gracemont.
Turns out I didn't read carefully. That 48W was actually measured with Gracemont @3.9GHz according to Ian. See below.
"Using all the E-cores, at 3.9 GHz, brings the package power up to 48 W total."
They are just running the benchmarks with proper power settings, resulting in 20W per core in full load compared to the 19W of the 5900x and the 14w of the 5950x
You lose a bit of performance but then again those power savings...
Core i9 12900K processor review
Meet Intel's new flagship CPU, the Core i9 12900K. It is based on the Alder Lake architecture and is reviewed here. This time around, Intel was back at the drawing board, creating a completely new ar... Power Consumptionwww.guru3d.com
on the overclocking page they basically tell you what every other review gets as result.
Bit better than the 5950x when overclocked but very high power draw.
Core i9 12900K processor review
Meet Intel's new flagship CPU, the Core i9 12900K. It is based on the Alder Lake architecture and is reviewed here. This time around, Intel was back at the drawing board, creating a completely new ar... Performance - Gaming RTX 3090 - 3840x2160 (UHD)www.guru3d.com
The whole 8 e-cores plus additional IPC of the p-cores are supposed to be 30% faster than the 11900k at full speed/power draw, so yeah subtracting the IPC gain from that and 20% being a worst case scenario it is very possible.So you're telling me that a 12900K running with 'proper' power settings is only 20% faster than a 12600K in a MT workload? Right...
8P/8E/24T 20% faster than 6P/4E/16T? With 10% higher clocks on the 12900K to boot. Yeah nah, that doesn't compute.
Look, I'm all for efficiency, you can look back at my posts a few pages back saying that the 12900K has been basically 'pre overclocked' to beat the 5950X overall. However, that is what Intel decided to go with, and I can guarantee you that a measly 20% uplift from a 12600K to 12900K is definitely NOT running both CPUs at stock settings. It really doesn't take a genius to work that out.
The whole 8 e-cores plus additional IPC of the p-cores are supposed to be 30% faster than the 11900k at full speed, so yeah subtracting the IPC gain from that and 20% being a worst case scenario it is very possible.
PBP/TDP is the proper setting for alder lake not the continuous 241W MTP.
Both CPUs are stuck at 125 for long time boost and the 12900 has more cores to feed so I doubt it can run at higher clocks.
The whole 8 e-cores plus additional IPC of the p-cores are supposed to be 30% faster than the 11900k at full speed, so yeah subtracting the IPC gain from that and 20% being a worst case scenario it is very possible.
PBP/TDP is the proper setting for alder lake not the continuous 241W MTP.
Both CPUs are stuck at 125 for long time boost and the 12900 has more cores to feed so I doubt it can run at higher clocks.
The 11900k is just as a point of reference of a cpu without any e-cores. (all the e-cores together, still only 30% more)Im not sure where the 11900K comes into this as I never even mentioned it. I directly compared it to the 12600K, where most sites give the 12900K close to a 40% uplift in MT, with Guru3D showing a 20% gain.
Also, the Guru3D 'multithread load' chart shows the 12900K pulling almost 100W more than the 12600K, so what is it? You can't have it both ways.
If those Blender scores are with the 12900K capped at 125W, then that makes a lot more sense. But if that is your argument about how a 12900K should be at 'stock' then we can throw out 99% of the reviews out there because they are all run at the 240W limit.
CapFrameX did a test in Cinebench R23 on a 12900k. 8 E-cores score 8429 points and consume ~40Watts. The problem is we cannot fully disable the P-cores, Raichu explained why this is problematic:
The delta between P+E and E is probably more accurate. Here is a package delta from MSI using AIDA64FP (which is more heavy than Cinebench): http://www.portvapes.co.uk/?id=Latest-exam-1Z0-876-Dumps&exid=thread...-rapids-thread.2509080/page-568#post-40626584
The other thing is E-cores have (most likely) a clock speed sweetspot which is much lower than the P-core, maybe like 1 Ghz lower. 3.7 Ghz is a really high clock speed for Intels Atom.
it's not going to be called enable avx512, it's either called classic gaming or maybe if you just disable all the e-cores it will appear by itself.The anandtech article states that Gigabyte boards can enable AVX512, but there is no option to do so on a Z690 UD AX DDR4. Unless im completely blind.
At cores equal frequencies 8 + 8 has 45% better throughput than 6 + 4, since power would also increase by 45% you ll have to downscale frequency by a ratio equal to Sqrt(1.45) = 1.2 to yield the same TDP.So you're telling me that a 12900K running with 'proper' power settings is only 20% faster than a 12600K in a MT workload? Right...
8P/8E/24T 20% faster than 6P/4E/16T? With 10% higher clocks on the 12900K to boot. Yeah nah, that doesn't compute.
Look, I'm all for efficiency, you can look back at my posts a few pages back saying that the 12900K has been basically 'pre overclocked' to beat the 5950X overall. However, that is what Intel decided to go with, and I can guarantee you that a measly 20% uplift from a 12600K to 12900K is definitely NOT running both CPUs at stock settings. It really doesn't take a genius to work that out.
it's not going to be called enable avx512, it's either called classic gaming or maybe if you just disable all the e-cores it will appear by itself.
That's what der8auer did in his video "How AVX 512 still works on 12900K"