KaitouX

Only if you're overclocking or remove the power limits, otherwise the maximum boost is 180W, but it should be around slightly below that at 165W or so under normal full load like Blender and CPU encoding, the maximum would probably only be reached in some stress tests. You can also lower the power limits for a small performance hit if you're worried about the power consumption or undervolt both CPU and GPU and lower quite a bit the power consumption without any noticeable hit to performance.
 
I'm assuming it's the XFX XTR 650W, which is in the A tier low priority in the tier list.

Assuming Zen 4 follows the expected gains from the demos and they don't change pricing/core count, only the 16-core part would be able to beat Alder Lake by a significant margin, let alone Raptor Lake.
Being pretty optimistic Zen 4 gains would make a 5600X and 5800X successor be slightly behind the 12600K and 12700K, while the 5900X successor would be slightly ahead of the 12900K. That's in MT, in ST Intel would be faster by a decent margin.
 
Power consumption might not be that great of a deal with Raptor Lake, at least if Intel doesn't do what they've done to the 12900K, and increased the power consumption in almost 100W for less than 10% more performance, at 190W it's less than 5% difference. The 12600K and 12700K are already basically as efficient as the 5600X and 5800X, while the 12700 is similar to the 5900X, they don't need much to compete in efficiency, only the 5950X and it's successor are probably going to stay isolated as the most efficient consumer CPUs in MT tasks.

Maybe the one they shown is a 16-core with 170W PPT, but they have/plan another SKU on top of that with 230W PPT and higher clocks, like Intel have the 12900KS?
Or it's just for future 24 or 32 core parts with Zen 5 or later.

Assuming Zen 4 follows the expected gains from the demos and they don't change pricing/core count, only the 16-core part would be able to beat Alder Lake by a significant margin, let alone Raptor Lake.
Being pretty optimistic Zen 4 gains would make a 5600X and 5800X successor be slightly behind the 12600K and 12700K, while the 5900X successor would be slightly ahead of the 12900K. That's in MT, in ST Intel would be faster by a decent margin.
 
Power consumption might not be that great of a deal with Raptor Lake, at least if Intel doesn't do what they've done to the 12900K, and increased the power consumption in almost 100W for less than 10% more performance, at 190W it's less than 5% difference. The 12600K and 12700K are already basically as efficient as the 5600X and 5800X, while the 12700 is similar to the 5900X, they don't need much to compete in efficiency, only the 5950X and it's successor are probably going to stay isolated as the most efficient consumer CPUs in MT tasks.

I read the comment wrong and thought you were the other guy

Yes, that's my point.

The demo on stage was a mixed multi core workload, Zen 3 maximum boost is for single core. 5950X typical gaming is ~4.7GHz and all core boost for blender is ~4Ghz, so it's I think about 15% uplift for similar workload.
 
Unless of course the demo game is extremely single threaded but I wouldn't realistically assume that.
 
Well that sucks 😢
 
I'd really like to know a lot more about these, what does this actually mean?!?!?! PCIe switch chips and expander chips? More stuff piped through the chipset?
 
Why should direct PCIe lanes to the CPU be affected by chipset like this? Shouldn't we just get the choice to spend more on a better X670 board that has full PCIe 5.0 rather than having an extra X670E chipset? Or is X670E "in name only" and it's just a cert spec for boards that are full PCIe 5.0 and it's actually the same chipset?
 
🤷‍♂️
 
Personally I'm quite disappointed, I was expecting better improvements with the change to TSMC 5nm. Maybe I need to wait for actual reviews and things like game performance has increased much more than indicated here.
 
Feels like the most effort was put in to the IOD changes to me and not Zen 4 and the 5nm chiplets. Appears to be just a node shrink and a sprinkle of cache and called it a day.
 
Nice, since they were talking about socket power I was suspecting it would actually be PPT.

The optimistic rumors seem to mostly be around 35% ST performance over Zen 3, with MT being a bit over that due to higher all-core boost(possibly using the 170W "TDP" that AM5 is supposedly rated for?). But even if we consider a absolutely best case scenario for AMD where the 8 core CPU is 50% faster in MT tasks on average, it would still at best match a stock 8P+8E ADL CPU(12900K), which is slower than what a likely 8P+8E 13700K would be.
One issue that Zen 4 is probably going to have is pricing, there are rumors that Zen 4 will have yet another price increase over the previous gen, that together with it being DDR5 only and the likely more expensive boards, might make it way too expensive.
My personal guess is that Zen 4 will be around 30% over Zen3 in ST, ~35% in MT. While RPL is going to be around 7% faster in ST and ~20% faster in MT compared to ADL.
 
Also this is probably unlikely to happen, but if Intel makes their i5 and i3 non-K line up include E-cores, it would be amazing value, so I hope they do it. 

Being somewhat optimistic Zen 4 will be 30~35% faster in ST compared to Zen 3, Alder Lake is ~18% faster, so if Raptor lake is around 7% faster than Alder lake in ST, which would be a fair improvement for a "refresh", this would put Zen 4 ST average under 10% ahead of Raptor lake.
When comparing the 12600(6P-cores, with power limits removed to match the K variant boosting behavior) to the 12600K(6P+4E-cores), the 12600K is ~30% faster in MT. When comparing the i9 12900K stock against itself with the E-cores turned off, it seems to be about ~35% faster in MT tasks.
 
Overall I think the situation will be somewhat similar to the Zen+ vs Coffee Lake, where one side has a slight ST advantage, but the other has way more cores/threads.

For reference of the performance difference between the 5700X and 12700:

[65W, RM1] = Power limited to 65W, using stock cooler.
[RM1] = Using stock cooler.
[H170i] = Using a good cooler, in this case a 420mm AIO.
 
The 5700X is slightly slower than the 5800X.
https://www.techspot.com/review/2391-intel-core-i7-12700/

Not really, the non-K(and 12600K) CPUs, with exception of the i9, use about the same amount of power as AMD parts with similar performance. 12400=5600X, 12600K=5800X, 12700=5900X. Only the 5950X and 5700X are actually more efficient.

Not really, the non-K(and 12600K) CPUs, with exception of the i9, use about the same amount of power as AMD parts with similar performance. 12400=5600X, 12600K=5800X, 12700=5900X. Only the 5950X and 5700X are actually more efficient.

This shouldn't really be an issue if you select the right parts. You can for the most part greatly reduce the power consumption from stock settings by undervolting and power limiting the GPU, as example my 3060Ti was able to go from 200W at stock to 120W while keeping 85% or so of the stock performance, the lowest I could go was 100W due to BIOS limitations on the GPU, and at 100W the performance would drop quite a bit, but if you get one that can go lower, you can probably configure it to use 80W, my old 960 was able to go to 60W or so without issues.
 
Naturally that doesn't mean it should be done, this would be incredibly hard and expensive to do properly. Properly cooling a desktop GPU in a notebook would be a nightmare.

Pcpartpicker shows this site as the cheapest option:
https://www.paradigit.ie/Intel-Core-i5-12600KF/21155147/product?
The site seems to be Irish even though the company is from Netherlands, so I think there would be no taxes, only the included VAT. Not sure on how good the shop is though.

You probably should, the auto voltage usually is focused on making sure the CPU is stable disregarding power consumption, so it often ends higher than it could be.
Also if you don't want a space heater (while in full load) you shouldn't OC, you're likely increasing the power consumption in 50%+ for a small performance improvement.

The 12700F is $30 cheaper if you don't need the iGPU.

Also Zen 4 is DDR5 only, which costs a lot more than DDR4 now, and even if it will be cheaper by the time Zen 4 comes out, it probably won't be anywhere near DDR4 prices, particularly for kits that aren't trash.

I mean, they could just make the parts with higher core counts all-core clock a bit lower. The 12900K when limited to 180W is only 5% or so slower than when at 250W and 165W is less than 10% slower, so if Intel is willing to have a 241W all core boost CPU they could make the P-core count higher without issues from the power side, they just need to sacrifice the clock a bit.

I mean, they could just make the parts with higher core counts all-core clock a bit lower. The 12900K when limited to 180W is only 5% or so slower than when at 250W and 165W is less than 10% slower, so if Intel is willing to have a 241W all core boost CPU they could make the P-core count higher without issues from the power side, they just need to sacrifice the clock a bit.

I guess this will have the 8P + 24E cores as the highest SKU?
12900K = 8P + 8E
13900K? = 8P + 16E
14900K? = 8P + 24E?
15900K? = 8P + 32E (This one was confirmed to be the target by Intel)
 
TBH I'm not sold on the core count increase being so focused on the E-cores, I personally would rather have something like 12P+16E, ideally having the option would be the best, so people that benefit from the core count can go that way(8+24), while others can go with a more balanced set up, but that probably wouldn't work/make sense for Intel.

You didn't really mention what you do that would benefit from a CPU upgrade, but for most people the Intel option will be better, as it is similar in MT workloads while being faster in ST ones.
If you're willing to sacrifice OC on the Intel side, you can drop the CPU to the 12700F and save 50 EUR there, depending on the features you want/need from the MB you can also drop the motherboard to a B660 if you go with the 12700F. You can also drop the X570 board to a B550 one depending on the features you want/need if you decide to go with the 5900X.
OC probably isn't really worth on either CPUs, on Intel it's a bit better, but you still would be increasing the power consumption in 50%+ for a ~5% performance increase.
The 12700F uses about the same amount of power as the 5900X, while the 12700K uses ~30W more at stock, if you want to tweak for better efficiency, the 5900X is the better option, as it loses less performance when power limited below 100W.

No it isn't. Just because Ryzen was better doesn't mean it still is, currently Intel have the core count advantage, which was the main reason why Ryzen used to be faster in MT applications. And on top of having the core count advantage, Intel also have way faster ST performance currently.
 
The 12700 performance is on the same level as the 5900X, it's way faster than the 5800X in pretty much everything.

It's an underclocked 5600G not 5600, it doesn't support PCIe 4.0 and have only 19MB of cache, which is almost half of what the 5600(X) has at 35MB. It's quite a bit slower in games compared to the 5600, in many cases it's about the same as the 3600.

There is a beta bios available that should support Ryzen 5000.
 
But like others said, if you don't need more performance, there's no need for an upgrade.

There is a beta bios available that should support Ryzen 5000.
 
But like others said, if you don't need more performance, there's no need for an upgrade.