Intel to Go Ahead with "Meteor Lake" 6P+16E Processor on the Desktop Platform?

[Crackong]

If you need commercially equal cores, buy an EPYC or Xeon.

Exactly.
I do need equal cores for virtualization
So I had TR PRO as bare metal

If you think that all P-cores work equally well, you lack experience in dealing with many-cores.

If you think they cannot make all P-cores working equally well, what makes you think they can handle mixed Hybrid cores work well ?

 

[hs4]

Although we see the opinion that "E-cores are unnecessary for games," the difference in frame rates between 12600 (6P+0E) and 12600K (6P+8E), for example, does not seem to indicate that E-cores are having a negative impact. Also, when games were being streamed or recorded, or when games were being played while connected to Discord, the thread director stabilized affinity, so that the 12600K sometimes performed as well as or better than the 5900X in both game and recording frame rates.

Most of the complaints about the E-core are conjecture. Incidentally, I have also encountered hybrid-specific problems, but it is questionable whether those who complain about E-cores are really putting in the kind of load that runs into E-core-specific problems.

 

[john_]

Intel to Go Ahead with "Meteor Lake" 6P+16E Processor on the Desktop Platform?

I guess Intel needs something to beat the 7800X3D in gaming for mid range CPUs.

 

[Dirt Chip]

You know, less is more...

 

[Crackong]

Although we see the opinion that "E-cores are unnecessary for games," the difference in frame rates between 12600 (6P+0E) and 12600K (6P+8E), for example, does not seem to indicate that E-cores are having a negative impact. Also, when games were being streamed or recorded, or when games were being played while connected to Discord, the thread director stabilized affinity, so that the 12600K sometimes performed as well as or better than the 5900X in both game and recording frame rates.

Most of the complaints about the E-core are conjecture. Incidentally, I have also encountered hybrid-specific problems, but it is questionable whether those who complain about E-cores are really putting in the kind of load that runs into E-core-specific problems.

I am sry but tests have shown having 2 more P cores is always better than 8 e cores in gaming

And disabling the e-cores are more likely to benefit you in gaming overall

So my points still stands.
As a gaming CPU,
Intel would be better off giving us 10/12 P-cores instead of 8+16

 

[hs4]

Exactly.
I do need equal cores for virtualization
So I had TR PRO as bare metal



If you think they cannot make all P-cores working equally well, what makes you think they can handle mixed Hybrid cores work well ?

Since the days when all the cores were equal, the program is built on the assumption that each logical core does not function equally (mainly for HT/SMT). This has been done by everyone before, and even the first review at the launch of Alder lake showed that the actual results were in line with logical performance for Blender and video encoding. This means that those applications were previously built on the assumption of unequal logical cores.

 

[Shtb]

So much theoretical talk about E-core and nothing about practical applications... well, let's get back to reality.

 

[hs4]

I am sry but tests have shown having 2 more P cores is always better than 8 e cores in gaming

And disabling the e-cores are more likely to benefit you in gaming overall

So my points still stands.
As a gaming CPU,
Intel would be better off giving us 10/12 P-cores instead of 8+16

It is known that the ring bus clock is lowered by the E-core in ADL, but this problem is solved in RPL Has the same test been done in RPL?

 

[john_]

Although we see the opinion that "E-cores are unnecessary for games," the difference in frame rates between 12600 (6P+0E) and 12600K (6P+8E), for example, does not seem to indicate that E-cores are having a negative impact. Also, when games were being streamed or recorded, or when games were being played while connected to Discord, the thread director stabilized affinity, so that the 12600K sometimes performed as well as or better than the 5900X in both game and recording frame rates.

Most of the complaints about the E-core are conjecture. Incidentally, I have also encountered hybrid-specific problems, but it is questionable whether those who complain about E-cores are really putting in the kind of load that runs into E-core-specific problems.

No indication of negative impact, doesn't mean they are necessary.

Having to stream your games, or if you are not a streamer, maybe run continuous virus checks or other tasks in the background just to justify the presence of E cores, is not exactly a good argument.

As for complains, well, when the new Xeons get out and you start seeing left and write what 16 P cores can do, you might understand why we SHOULD HAVE a problem seeing DESKTOP CPUs copying SMARTPHONE SOCs in having more than one type of cores. When in the future you will be reading in the specs 3 types of cores, like for example
"16 Core CPU: 2 Performance cores, 2 (mid range) M cores (maybe even older arch) and 12 E cores"
at the price of $300, you will understand. Intel waits for AMD to get in the game of E cores, with their Zen4c and then the consumers lose.
On the other hand, shareholders win, so I shouldn't complain.

 

[Karti]

[...]
Yeah awesome, the return of the quadcore in 2023. Intel selling you regression at premium, where can I sign up.
[...]

Yep, because if you think you need more than 2-4 strong cores for gaming and e-cores are worthless is just proving how wrong you are

and if i3 would use p+e cores, it would no longer be quad core... depending on how many e-cores you would give it, it could be 6 or 8 core unit
open your eyelids i guess

 

[Crackong]

This has been done by everyone before, and even the first review at the launch of Alder lake showed that the actual results were in line with logical performance for Blender and video encoding. This means that those applications were previously built on the assumption of unequal logical cores.

That's maybe your use case.
As I said,
I value these CPUs based on my own use cases.
And my use case works best on many equal cores.

E-cores are totally useless in my use cases.

 

[oxrufiioxo]

I'm a little surprised there is so much push back about E cores sure if intel released a 10/12 Pcore cpu I'd be on Alderlake/raptorlake but the majority of people I personally know that have 12/13th gen intel are very happy with their systems same with people I've personally done AM5 systems for... Don't get me wrong I'm not thrilled about meteor lake being 6P cores but I'd rather see benchmarks before deciding if it's not for me. My biggest worry if it does suck we will just get amd pricing 6 cores at 300+ and 8 cores at 400+ again which is worse than intel using E cores imo.

 

[JustBenching]

So much theoretical talk about E-core and nothing about practical applications... well, let's get back to reality.

View attachment 285652

You realize that post is from more than a year ago right? That was a couple of months after the launch of the actual 12th gen CPUs with ecores.

 

[Crackong]

Has the same test been done in RPL?

Idk, much appreciated if you could find some for us.

 

[hs4]

I provide the results of my meta-analysis of the AnandTech and Phoronix reviews of the actual application impact.
At the time of the 13900K review, the processing speed of each application, at least in Windows 11, was as expected from typical benchmarks.

 

[Dirt Chip]

If 'Meteor' will fail to deliver (preformance) zen5 might have a "zen" moment.
I give it low chance though.

 

[lemonadesoda]

I am OK with 4x e-cores on my (future) 8 core CPU. So long as;

• 8 real cores, doing work,
• With the 4x e-cores tied and bound to the web-browser, incl. all other background apps like messengers, skype and other micro-apps hogging resources etc.

Leave my 8x real cores alone. And the OS scheduler had better not put any real work onto those e-cores. And had better keep micro-hogs off my p-cores!

But then, the scheduler needs rethinking. I don't want it doing dynamic load balancing - it needs to tie workloads to certain core-types.

Is that a waste of silicon? No more so than having silicon dedicated to sound, or to network, or to GPU, and not using it when you are not using it.

 

[Dimitriman]

My guess is that Intel wants to transition to E-cores only that are fast enough in large numbers, hence the gradual reduction in P cores...

 

[lemonadesoda]

I really do understand the use of e-cores in laptops to extend battery life.

And possibly in corporate office desktops in their thousands to reduce power consumption.

But for workstation users - I see not benefit going beyond 4 e-cores to cover standby and background tasks with scheduler providing binding. All silicon and second processor sockets to p-cores! ie. keep e-cores off xeon thank you!

=====
6P + 16E is just nonsense and just marketing and PR
moar cores
lower power usage per core
show benchmarks? scratch that

Anyone that "needs" 6P cores would get better performance with 8P+8E than 6P+16E, but one says "16 cores" the other says "22 cores" and on retail shelves that might make a difference to Joey and Granddad.

 

[hs4]

In last summer's Hot chips presentation, there was a slide that read that the limit of power supply via the base die was around 150-180W. the reason for the low number of P cores is probably due to the smaller capacity for power supply.

- IPC +10%.
- Same max clock
- 30% reduction in power consumption per core

With this assumption, 6P+16E can be estimated to have the same MT performance as 13900K at 180W. Frankly I still doubt the existence of MTL-S, but if 6P+16E exists, it will be similar to the relationship between Rocket lake and Comet lake.

 

[GerKNG]

It's amazing how every "ecores bad" comment is here from AMD users who will buy an 6+8 zen 5 and love it

The day AMD does the same will make me buy the faster CPU from either intel or AMD with the least amount of cinebench accelerators and i'll disable them.

 

[Denver]

One more step and they can go back to plans to sell quad-cores, now with E-cores to disguise

 

[Vario]

I'm a little surprised there is so much push back about E cores sure if intel released a 10/12 Pcore cpu I'd be on Alderlake/raptorlake but the majority of people I personally know that have 12/13th gen intel are very happy with their systems same with people I've personally done AM5 systems for... Don't get me wrong I'm not thrilled about meteor lake being 6P cores but I'd rather see benchmarks before deciding if it's not for me. My biggest worry if it does suck we will just get amd pricing 6 cores at 300+ and 8 cores at 400+ again which is worse than intel using E cores imo.

For starters, I will say the present AMD offering is really fantastic. However, the people complaining about E cores are AMD enthusiasts looking to gripe about the other "team". You can tell based on their "system specs". I have the 12900KS and Win 10 LTSC and even in that suboptimal OS for E cores I have never had a single problem, my machine just straight up chews and spits out any application I throw at it.

Your pragmatic approach of "Wait and see benchmarks" is the sensible approach.

 

[dyonoctis]

That's maybe your use case.
As I said,
I value these CPUs based on my own use cases.
And my use case works best on many equal cores.

E-cores are totally useless in my use cases.

Or you could buy AMD. Isn't that's why competition is a good thing ? anyone who doesn't like the e-cores can just get a 7950x/7900x and call it a day.

 

[hs4]

The day AMD does the same will make me buy the faster CPU from either intel or AMD with the least amount of cinebench accelerators and i'll disable them.

Zen4D/5D is said to be simply a reduced cache (in Zen4, L2+L3 occupies about 60% of the CCD area; in Cezanne/Barcelo, L3 is halved to get iGPU area, which is similar). In other words, if there is a Ryzen hybrid in the future, it will be close to the 7900X3D / 7950X3D. We will see tomorrow how much of an impact that will have. And realistically, the 7950X3D will probably get higher FPS with non-vcache ccd disabled (this has also been confirmed with vanilla 7950X)

Anyone that "needs" 6P cores would get better performance with 8P+8E than 6P+16E, but one says "16 cores" the other says "22 cores" and on retail shelves that might make a difference to Joey and Granddad.

The MT performance of 8P+8E is equivalent to 12P+0E, and 6P+16E is equivalent to 14P+0E in MT perf. If someone is using it mainly for rendering and encoding, 6P+16E will work better. And for most people, there will not be much difference.