Question Intel Raptor Lake vs AMD Zen 4 vs Apple M2

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Carfax83

Diamond Member
Nov 1, 2010
6,841
1,536
136
These CPUs are all going to square off against each other at some point this year assuming nothing catastrophic occurs to delay any of the product launches. So going by what we know from official sources and informed rumor mongers (many of which were very accurate before Alder Lake and the M1 launched), which CPU do you think will win out in these categories?

1) Single threaded performance
2) Multithreaded performance
3) Gaming performance
4) Performance per watt
5) Overall performance (who wins the majority of applications)

While I've been keeping a close eye on rumors and leaks for Zen 4 and Raptor Lake, I have not admittedly been doing so for the M2; as I'm unrepentant Apple hater :innocent: At least I'm honest about it... That said, this is my ranking based on what I've seen and heard:

I think the single threaded crown will go to Raptor Lake, and I say this based on informed rumors that Raptor Lake will have up to 10% more IPC from microarchitectural updates, cache upgrades and higher clock speeds than Alder Lake. From what I've seen, gauging IPC performance isn't easy as it varies so much based on application, but I'd say Alder Lake already has at least a 15% across the board IPC advantage over Zen 3, so Raptor Lake could conceivably have 25% better IPC than Zen 3, which is similar to what Zen 4 will reportedly possess. But I doubt Zen 4 will match Raptor Lake in clock speeds and memory latency performance, which is why I'm predicting Raptor Lake will take the single threaded performance crown.

For multithreaded performance, Zen 4 should easily take it due to having more big cores than its Intel counterpart and similar IPC.

Gaming performance is more complicated because while some games are inherently more reliant on single core performance (strategy games for instance), more and more 3D engines are becoming increasingly parallel due to the adoption of Vulkan and DX12 in addition to modernized programming methods. Still, very few 3D engines can scale beyond 8 threads and 6 to 8 cores remains the sweet spot for gaming and will be for some time. So overall, I feel more comfortable going with Raptor Lake for the gaming crown. Also if rumors are correct, Raptor Lake will officially support DDR5-5600 off the bat while Zen 4 will reportedly use DDR5-5200. The raw memory speed won't likely be a significant factor, but Intel's memory controller will be right next to the CPU cores while Zen 4's will be in the I/O die which while still on the same package will definitely incur a significant latency penalty; which I'm sure will be offset by a massive L3 cache. :)

On performance per watt, one would think the M2 should take this category easily......but from the small amount of research that I've collected on it, it seems that there won't be much of a performance increase with the M2, if any. Some rumors are even suggesting there may be a bit of a regression in that aspect. Also since Zen 4 will be on TSMC's 5nm node, it will undoubtedly have excellent performance per watt and I believe it will also easily crush Apple's best in single core. So for performance per watt, I'm going to go with Zen 4.

When it comes to overall performance, I'm leaning towards Zen 4 but it will be close. Raptor Lake will supposedly double the amount of Gracemont efficiency cores which will certainly help in multithreaded performance per watt, but ultimately they won't be a match for Zen 4's 16 big cores with SMT. AMD will have the core count advantage and when that's combined with IPC parity with Raptor Lake, Zen 4 will win the majority of the benchmarks.
 

scineram

Senior member
Nov 1, 2020
361
283
106
Alder Lake 12900KS has a 5.5 Ghz ST boost and 5 Ghz all core boost already, how is 5.1 Ghz All cores on 13900K delusional?


View attachment 57211
All core will be possible if molten VRM is your thing. But 5.8 GHz is pure hopium.
 

TheELF

Diamond Member
Dec 22, 2012
3,967
720
126
All core will be possible if molten VRM is your thing. But 5.8 GHz is pure hopium.
A hundred Mhz more all core isn't going to melt anything that isn't melting now, if we assume the same amount of cores it should be pretty in line with what we have now.
The 5.8Ghz single core is hopium indeed though, the same 5.2 as the normal 12900k or 1-2 hundred Mhz more is the most that would make sense. The 5.5 of the KS is extremely high binned in the first place and not the standard of this gen.
 

DrMrLordX

Lifer
Apr 27, 2000
21,582
10,785
136
Nah.. 5.8ghz single core abd 5.1 all core boost..

That's not what he said.

raptor lake will have new intel technology for voltage it decreases power consumption by 30%

They can't just magically change the v/f curve limitations imposed by the process.

Having more at lower clock speeds will still be a net gain. The efficiency takes a nose dive with Intel pushing the clocks to absurd levels just to have a slightly bigger bar.

True. But you can't just take the existing CB23 score, bump up the score up by 7%, then double the contribution of the E-core cluster and call it a day. That's not how it's going to work. Not unless Intel ups the TDP and PL values.

I really don't see that happening.

???

Why?

You really think Intel is going to have twice as many Gracemont cores, all running @3.9GHz in a workload like CBR23?
 

yuri69

Senior member
Jul 16, 2013
373
573
136
To OP: Too many variables.

* Intel's current IPC advantage can be 0% (CB) and up to 20% or so (CPU hard limited games).
* Next gen Apple's manufacturing process is still unknown.
* Golden Cove seems to be wide and big so there are might be bottlenecks to be removed in Raptor.
* Raptor power efficiency seems better as the mobo specs revealed.
* Zen 4 might be frequency optimized unlike its predecessors as hinted at CES.
* AM5 might improve on power delivery and allow AMD to go into higher power levels.
* The leaked Zen 4 spec doesn't look like a departure from Zen 3.
 

Mopetar

Diamond Member
Jan 31, 2011
7,797
5,899
136
ah, in that case we might have a much closer race for performance/watt between Apple and AMD

Not likely. Apple will smoke both AMD and Intel in low power situations. I have an M1 Max MacBook Pro and even under heavy loads I'm still amazed at how quiet and cool it stays. I'm sure Apple could have pushed for another 10% performance at the expense of 50% more power, but that's not something they would do.

While Apple isn't doing anything magical that AMD or Intel couldn't also do as far as CPU design is concerned, the fact that they don't have to break down x86 instructions will always give them an edge. Otherwise they just have a really good design team that can keep up with some of the best in the business.
 

gruffi

Member
Nov 28, 2014
35
117
106
Nah.. 5.8ghz single core abd 5.1 all core boost.. raptor lake will have new intel technology for voltage it decreases power consumption by 30%
You are talking about DLVR. That will only help with partial load. It's a tech for low TDP SKUs where every watt counts. That's why it was announced for mobile only. Desktop Raptor Lake won't benefit from it. I think 5.8 / 5.1 GHz is too much wishful thinking for the P-core. But something like 5.5 / 4.8 + 3.8 (E) GHz could be doable at 250W. OTOH Zen 4 could achieve similar clock speed single core if 5 GHz all core is possible according to AMD.

Alder Lake 12900KS has a 5.5 Ghz ST boost and 5 Ghz all core boost already, how is 5.1 Ghz All cores on 13900K delusional?
That's special selected silicon with higher power consumption. That doesn't count. Especially because Raptor Lake has to share its TDP with 8 more E-cores.




At the moment we have the following scenario for the top SKUs:
single core: Zen 3 < Alder Lake +15%
multicore: Alder Lake < Zen 3 +10%
gaming: Zen 3 < Alder Lake +5-10%

My expectation is Zen 4 will increase single core IPC by ~20-25% and clock speed by ~10% (something like 4.9 -> 5.4 GHz). So, overall single core performance could increase by ~30-35%. Multicore performance could be increased even more due to the 5nm process and higher all core clock speed, ~40-50%. Assuming top Raphael SKU will still offer 16 cores.

I don't expect Raptor Lake to increase single P-core IPC by much, maybe 5-10%. With 5% higher clock speed (something like 5.2 -> 5.5 GHz) it could be 10-15% faster. Currently performance distribution between P-cores and E-cores seems to be something like 3:1 on average. Which means the P-cores are responsible for ~75% of the multi core performance, the E-cores for ~25%. So, by doubling the number of E-cores we could see 25% higher multicore performance. Including the core improvements something like 40-50% should also be possible, just like Zen 4. But that's theory. Raptor Lake is still 10nm. There won't be much efficiency improvements. That's why Raptor Lake's clock speed should be more limited by power consumption than Zen 4. So, I rather expect 20-30% more performance multicore for Raptor Lake.

To sum it up, single core performance of Zen 4 and Raptor Lake should be very close. Maybe with slight advantages for Zen 4 (except in some pointless synthetic stuff, of course). Multicore performance should see Zen 4 as the winner, probably even increasing the gap again. Gaming is hard to guess. V-Cache based Zen 4 actually should win this easily. But we won't see it this year anymore. Zen 4 without V-Cache? With similar improvements as Zen 3 it could have the upper hand over Raptor Lake. Otherwise it's a tight race again.
 

TheELF

Diamond Member
Dec 22, 2012
3,967
720
126
OTOH Zen 4 could achieve similar clock speed single core if 5 GHz all core is possible according to AMD.
You guys have to keep in mind that that demo was running a game, which uses much less power and allows much higher clocks, ZEN 4 will have to hit high clocks in at least cinebench if not even in a torture test with AVX if it's going to compare favorable against intel.
If 5Ghz all core is the best it can do under very light load as games are then anything else will be running much lower clocks...or the power draw will be much higher, if the arch even allows it.

But no matter what, this is just a very early ES so we have to wait and see how it will turn out.
 

Markfw

Moderator Emeritus, Elite Member
May 16, 2002
25,481
14,434
136
You guys have to keep in mind that that demo was running a game, which uses much less power and allows much higher clocks, ZEN 4 will have to hit high clocks in at least cinebench if not even in a torture test with AVX if it's going to compare favorable against intel.
If 5Ghz all core is the best it can do under very light load as games are then anything else will be running much lower clocks...or the power draw will be much higher, if the arch even allows it.

But no matter what, this is just a very early ES so we have to wait and see how it will turn out.
I don't know where you get that from. The only thing the 12900k beats AMD (the 5950x specifically) in , is gaming and a few single core benchmarks anyway, so AMD is already favorable. Zen 4 will be better in every respect, and so will raptor lake. Its WAY too early to be staging things like AMD is already not favorable. the 5950x is already the winner in MT scenarios.
 
  • Like
Reactions: Drazick

Carfax83

Diamond Member
Nov 1, 2010
6,841
1,536
136
I don't know where you get that from. The only thing the 12900k beats AMD (the 5950x specifically) in , is gaming and a few single core benchmarks anyway, so AMD is already favorable.

Um, not really. Golden Cove is very strong in certain multithreaded apps like code compiling and encoding and punches well above its weight compared to Zen 3. That's why I think Golden Cove has a higher IPC in the double-digit range than Zen 3 compared to what some of the other posters stated.

compiler.png


embed.php

embed.php
 

Markfw

Moderator Emeritus, Elite Member
May 16, 2002
25,481
14,434
136
Um, not really. Golden Cove is very strong in certain multithreaded apps like code compiling and encoding and punches well above its weight compared to Zen 3. That's why I think Golden Cove has a higher IPC in the double-digit range than Zen 3 compared to what some of the other posters stated.

compiler.png


embed.php

embed.php
I did not mean in all tests, but on average...... Geesh... Still 3 out of the number out there is pretty small, and and only 2 of those are by more than margin of error.

And of course, taking twice that wattage to do it doesn't count, huh...

Edit: and Anands review agrees with me, its in between the AMD 5900x and the 5950x

Quote:
"At a retail price of around $650, the Core i9-12900K ends up being competitive between the two Ryzen 9 processors, each with their good points. The only serious downside for Intel though is cost of switching to DDR5, and users learning Windows 11. That’s not necessarily on Intel, but it’s a few more hoops than we regularly jump through. "
 
Last edited:

Carfax83

Diamond Member
Nov 1, 2010
6,841
1,536
136
Raptor Lake is still 10nm. There won't be much efficiency improvements.

I don't know man, I think Intel 7 is more power efficient than what most people think.

Go on the Intel Reddit board and you will find plenty of reports of people undervolting their 12900Ks and shaving off 50 to 80w off the stock power usage and exceeding the 5950x power efficiency in Cinebench.

Alder Lake is just horribly overvolted because Intel wanted to make a great showing in multithreaded performance vs the 5950x and 5900x.
 
  • Like
Reactions: Henry swagger

Carfax83

Diamond Member
Nov 1, 2010
6,841
1,536
136
I did not mean in all tests, but on average...... Geesh... Still 3 out of the number out there is pretty small, and and only 2 of those are by more than margin of error.

But it's still impressive that Alder Lake can outperform Zen 3 easily in these multithreaded applications despite having significantly less big cores.

And of course, taking twice that wattage to do it doesn't count, huh...

Check again. The TPU review showed the 12700K beating the 5950x in MVS C++ compilation. Also, the Phoronix review had the 12600K ahead of the 5800x.

And like I mentioned above, these CPUs respond very well to undervolting. You can easily shave off 50w to 80w of the max TDP without affecting performance by undervolting because the CPUs are essentially overvolted from the factory.
 
  • Like
Reactions: Henry swagger

Ajay

Lifer
Jan 8, 2001
15,332
7,792
136
Alder Lake 12900KS has a 5.5 Ghz ST boost and 5 Ghz all core boost already, how is 5.1 Ghz All cores on 13900K delusional?


View attachment 57211
The 30% power reduction bit was delusional at w/higher clocks. I’ll eat crow if true - but it’s a very high bar.
 

DrMrLordX

Lifer
Apr 27, 2000
21,582
10,785
136
With unrestricted TDP, Power Efficiency be damned, sure why not?

Intel can push power consumption only so far. They're already pushing the limits on what you can cool with a huge dual-tower HSF. At least as far as Raptor Lake clocks are concerned (not sure about his Zen4/Raphael clocks), @gruffi seems to be in the right ballpark.

Yeah, 30% reduction with higher clocks and on the same node? And all it took was adding some voltage regulators on package? Highly questionable.

Raptor Lake will save some power in high utilization scenarios from being able to run Gracemont clusters on separate voltage rails. But those gains won't be enough to offset the addition of 8 more Gracemont entirely. We should still expect lower Gracemont clocks than Alder Lake in full-bore MT scenarios.
 
  • Like
Reactions: Tlh97 and Kaluan

Carfax83

Diamond Member
Nov 1, 2010
6,841
1,536
136
Again huge amounts of hopium around code compilation, which is super memory intensive. Of course DDR5 kicks ass, which Zen 3 doesn’t have. But Zen 4 will.

AFAIK, code compiling is mostly CPU, cache and memory latency intensive rather than bandwidth. Look at the LLVM compilation graph. Phoronix used DDR5-4400 spec RAM for Alder Lake in that review, but the 12900K is nearly 75% faster than the 12600K despite having the same memory bandwidth. The 12900K and the 12600K differ greatly in cache and core count however.

For Zen 3, DDR4-3600 was used and you see good scaling from the 5600x to the 5800x to the 5900x. The scaling dramatically diminishes with the 5950x likely due to the application not scaling well to 32 threads. It has nothing to do with memory bandwidth.
 

gruffi

Member
Nov 28, 2014
35
117
106
Go on the Intel Reddit board and you will find plenty of reports of people undervolting their 12900Ks and shaving off 50 to 80w off the stock power usage and exceeding the 5950x power efficiency in Cinebench.
Either you compare both at stock or both undervolted. Everything else is quite pointless. Undervolting also doesn't mean anything for production. It's still 10nm. Probably with some minor improvements. Don't expect any miracles for desktop Raptor Lake.

But it's still impressive that Alder Lake can outperform Zen 3 easily in these multithreaded applications despite having significantly less big cores.
Not impressive. When 5 GHz OC falls apart you know that the other results are likely with significantly higher clock speeds. So, power consumption also should be significantly higher than 5950X. That "Software Compilation" chart also doesn't reflect full utilization. This is only partial load. So, it's pointless to talk about the number of cores. Btw, software compiling also depends a lot on disk and memory performance. It's not just about CPU performance.

Sure, there might be cases where Alder Lake offers more performance multicore. Which is the result of more IPC and higher clock speeds (or even DDR5). But on average Zen 3 is still faster and more power efficient. The more cores are being utilized the better Zen 3 will perform compared to Alder Lake. And all this comes at a very high price for Intel. You have to look at a crucial aspect, especially for the professional market, area efficiency. Zen 3 is not less area efficient than Gracemont, probably even better. Despite Intel's 10nm is a little more dense than TSMC's 7nm. Leave alone Golden Cove. It's huge compared to Zen 3, ~7 vs ~4 mm², and quite area inefficient.
 

JoeRambo

Golden Member
Jun 13, 2013
1,814
2,105
136
Not impressive. When 5 GHz OC falls apart you know that the other results are likely with significantly higher clock speeds. So, power consumption also should be significantly higher than 5950X.

Good thing that we have Phoronix to check facts, in compilation results 12900K is tying 5950x and power is -10W to +30W

12900K "bad efficiency" is completely inflicted by Intel themselves, both retarded marketing decisions and unfortunate engineering "shortcuts" to make mobile cpu into desktop CPU.
 

gruffi

Member
Nov 28, 2014
35
117
106
Phoronix is a special case. Especially with those outdated compiler optimization settings. Although I'm a big fan of Linux and use it at office it's still not very meaningful for the client market.

I think this chart demonstrates Intel's biggest problem very clearly.

wprime.png

The big.LITTLE implementation is mediocre at best. You often have to disable the E-cores to get best results. And then it's sometimes not better than its predecessor. So, I wonder how more E-cores can really change anything. Intel needs a better and more area efficient P-core. But that likely won't happen before 2024 and the newly designed "Royal Core".
 

DrMrLordX

Lifer
Apr 27, 2000
21,582
10,785
136
Good thing that we have Phoronix to check facts, in compilation results 12900K is tying 5950x and power is -10W to +30W

There are no per-benchmark power measurements. You don't know how much power it's using in any of those individual tests. nm found the buttons to switch to power measurements.

Not sure why you're bragging on the 12900k's compiler performance. 12900k on the Linux kernel compilation bench takes 200W vs 5950X 142W and performance is a dead heat. GNU Debugger compilation was almost as bad.
 
Last edited:

TheELF

Diamond Member
Dec 22, 2012
3,967
720
126
Phoronix is a special case. Especially with those outdated compiler optimization settings. Although I'm a big fan of Linux and use it at office it's still not very meaningful for the client market.

I think this chart demonstrates Intel's biggest problem very clearly.

View attachment 57232

The big.LITTLE implementation is mediocre at best. You often have to disable the E-cores to get best results. And then it's sometimes not better than its predecessor. So, I wonder how more E-cores can really change anything. Intel needs a better and more area efficient P-core. But that likely won't happen before 2024 and the newly designed "Royal Core".
Well one of those two must clearly be wrong...

Intel-Core-i9-12900K-Benchmarks-wPrime-v2.10-1024M.jpg


Here is another with similar scores.