Question Which is better? More E-cores or more P-cores ?

[FlameTail]

Apple is betting on it's Big cores. The M1 Pro doubles the number of firestorm cores but halves the number of icestorm cores compared to the M1. A future 'M1 Quadra' is rumoured to have 32 Big cores and just 8 small cores. So, yes, Apple is focusing on it's big cores.

Intel on the other hand seems to be betting on it's E-cores( small cores ). Alder Lake has upto 8P+8E. Raptor Lake is rumoured to bump that upto 8P+16E. Arrow Lake will supposedly have 8P+32E ! So intel will be putting in more small cores, which is a complete opposite to Apple's approach.

Which approach do you think is better ? Why ?

Who will win ( performance, efficiency, cost wise ) in the end ?

 

[gtrxpro]

Hi! @FlameTail

My opinion, apple is doing somewhat better, because : bigger the processor more easy to make
easy to make = low price
low price = more audience
more audience = intel's goin' down





but... it is more like a win - win situation here, because some people like the budget and some people don't care about it and they like nano tech and stuff, which makes it equal at the end




by the way (sorry to burst your bubble but Amd is doing a whole lot better than the both of them)

 

[FlameTail]

Hi! @FlameTail

My opinion, apple is doing somewhat better, because : bigger the processor more easy to make
easy to make = low price
low price = more audience
more audience = intel's goin' down





but... it is more like a win - win situation here, because some people like the budget and some people don't care about it and they like nano tech and stuff, which makes it equal at the end




by the way (sorry to burst your bubble but Amd is doing a whole lot better than the both of them)

Could you elaborate on your last statement about AMD ?

 

[Leeea]

P-cores

We have not quite left the single threaded world behind just yet.

All AMD does is P-Cores, and it works just fine for them.

by the way (sorry to burst your bubble but Amd is doing a whole lot better than the both of them)

That is incorrect.

AMD could double its current market share, and Intel would still be the market leader.
Apples profits are 18x that of AMDs.

Yes, AMD makes a very compelling high performance part. But market inertia is on Intel's side.

 

[dullard]

All AMD does is P-Cores, and it works just fine for them.

Even AMD might be going to Big Little designs: https://wccftech.com/amd-big-little...st-patent-coming-to-next-gen-ryzen-cpus-apus/

As we go to more and more cores, each big core has less and less and less power to work with. Thus, ultimately you either need to castrate big cores, use only little cores, or go hybrid.

 

[FlameTail]

Both top dogs from Intel and AMD, the i9 12900K and 9 5950X have 16 cores. But half of the 12900K's cores are small cores.

DESPITE THAT, the 12900K beats the 5950X in benchmarks.

Sure, you can argue that the 5950X is last gen and the real rival to the 12900K is the 7950X ( or whatever it is called ).

But then, intel is also at a node disadvantage. That makes me wonder- if both of them were on the same node... the 12900K would absolutely roast the 5950X.

The small cores are no joke.

Very interesting.

 

[Qwertilot]

As per the initial question, it's just not that comparable - Intel's big cores are very, very big and power hungry.

Somewhat undesirably so to be honest. Apple (and AMD) have got their ones rather more under control, so it makes sense to load up on them instead.

 

[insertcarehere]

Even AMD might be going to Big Little designs: https://wccftech.com/amd-big-little...st-patent-coming-to-next-gen-ryzen-cpus-apus/

As we go to more and more cores, each big core has less and less and less power to work with. Thus, ultimately you either need to castrate big cores, use only little cores, or go hybrid.

Certainly, they are already paving the way with Zen 4 by splitting it into Zen 4 (vanilla) & Zen 4c (maximum density, reduced cache & clocks).

As to the question, there's no black and white answer as it all depends on how the big cores stack up against the small cores in terms of performance/efficiency/area... Etc.

For Apple, their big cores are plenty efficient enough (the same cores go in phones) that the incremental efficiencies associated with offloading significant compute to the small cores isn't worth the hassle. (Suspect the frequency-scaling mechanism within the SoCs isn't as sophisticated as what's being employed by AMD/Intel, which doesn't help). Quite the opposite with Intel.

 

[moinmoin]

Certainly, they are already paving the way with Zen 4 by splitting it into Zen 4 (vanilla) & Zen 4c (maximum density, reduced cache & clocks).

AMD's approach seems to be to take the same core and same feature set and optimize the hell out of it. The learning through that can be fed back to development of future cores as the lineage is still the same. I'm honestly not sure how much of that is true for Arm's and Apple's cores, and Intel's P and E cores are completely different.

 

[Timmah!]

All i know i am not buying next CPU from Intel, that has less P-cores than my current 7940x. So if they intend for the forseeable future in coming generations (Raptor, Meteor Lake) just increase the number of E-cores, and stick to the 8 P-cores, my next CPU will be AMD (24/32C depending on price). Presuming they will keep their current performance parity with Intel per core.

 

[gtrxpro]

P-cores

We have not quite left the single threaded world behind just yet.

All AMD does is P-Cores, and it works just fine for them.


That is incorrect.

AMD could double its current market share, and Intel would still be the market leader.
Apples profits are 18x that of AMDs.

Yes, AMD makes a very compelling high performance part. But market inertia is on Intel's side.

Im not talking about AMDs profit, but im talking about their processors

 

[gtrxpro]

All i know i am not buying next CPU from Intel, that has less P-cores than my current 7940x. So if they intend for the forseeable future in coming generations (Raptor, Meteor Lake) just increase the number of E-cores, and stick to the 8 P-cores, my next CPU will be AMD (24/32C depending on price). Presuming they will keep their current performance parity with Intel per core.

yea, good choice intel's ones are cheaper, less performance, if you are:

an overclocker? go with AMD,
are a gamer?? AMD!
Wanna multitask? : AMD threadripper

k, my point is most gamers and tech guys go with AMD because they wanna "SHOWOFF"
It is costly but.. Owning the world best known processor is better than saving a few bucks
right?

****** My personal choice??? AMD

 

[gtrxpro]

Could you elaborate on your last statement about AMD ?

Ohh, just saying AMD Is the BEST processor 2021

 

[lyonwonder]

IMO, Intel should go with 12 P-cores maximum in 14th Gen Meteor Lake to further differentiate i9 from i7 since the current i9s are basically just higher binned i7s that still have 8 P-cores.

 

[IntelUser2000]

All i know i am not buying next CPU from Intel, that has less P-cores than my current 7940x. So if they intend for the forseeable future in coming generations (Raptor, Meteor Lake) just increase the number of E-cores, and stick to the 8 P-cores, my next CPU will be AMD (24/32C depending on price). Presuming they will keep their current performance parity with Intel per core.

You are going to change your mind if they prove otherwise. It's about a generation away before each individual E cores are going to outperform every pre-Golden Cove cores.

Architecture-wise Gracemont E cores are pretty much at the level of the cores in the 7940X already and while the interconnects and caches aren't up to the level we're talking future CPUs here.

The idea of the hybrid is to be able to maximize low thread count performance, and I believe even with Golden Cove Intel still has way more work to do to make it actually a competitive core perf/watt perf/area wise. So when they actually get there and have a hybrid chip that uses it then we'll talk.

 

[FlameTail]

I think the best option is to have few Little Cores ( for efficiency ), many big cores (for strong single core performance ) and add hyperthreading to the big cores ( for strong multi threading performance ).

And BOOM

WE HAVE AN AWESOME CPU

 

[gtrxpro]

well as for newbies (who don't multitask) lower the better,


anyone here a BIG fan of overclocking?

 

[IntelUser2000]

I think the best option is to have few Little Cores ( for efficiency ), many big cores (for strong single core performance ) and add hyperthreading to the big cores ( for strong multi threading performance ).

See, that makes sense for mobile but for desktop it doesn't. In desktop, the platform power absolutely dominates power in idle, so you won't care about even saving 1W on the CPU level.

So adding E cores have to be about adding performance on the desktop, cause no other option makes sense.

They do have an option for you. I think it's called the 12700K.

 

[TheELF]

AMD's approach seems to be to take the same core and same feature set and optimize the hell out of it. The learning through that can be fed back to development of future cores as the lineage is still the same. I'm honestly not sure how much of that is true for Arm's and Apple's cores, and Intel's P and E cores are completely different.

TSMC 7nm HD and HP Cells, 2nd Gen 7nm, And The Snapdragon 855 DTCO

Update and analysis of TSMC 7-nanometer node low-power and high-performance cells, 2nd generation 7nm, and the design technology co-optimization (DTCO) effort that went into the Snapdragon 855 SoC.

fuse.wikichip.org

TSMC is the one that provides a high density and a high performance variant of their nodes.
AMD doesn't need to do anything here expect for figuring out how many cores of high density (smaller less performance) and high performance they want on their next design.

 

[Timmah!]

You are going to change your mind if they prove otherwise. It's about a generation away before each individual E cores are going to outperform every pre-Golden Cove cores.

Architecture-wise Gracemont E cores are pretty much at the level of the cores in the 7940X already and while the interconnects and caches aren't up to the level we're talking future CPUs here.

The idea of the hybrid is to be able to maximize low thread count performance, and I believe even with Golden Cove Intel still has way more work to do to make it actually a competitive core perf/watt perf/area wise. So when they actually get there and have a hybrid chip that uses it then we'll talk.

Well maybe, remains to be seen. Regarding current Golden Cove having more work to done regarding their watt performance, thats why, if i were to buy new CPU right now, i would lean toward AMD.

Merry Christmas to everyone!

 

[BorisTheBlade82]

IMHO...
If one would take a look of the general distribution of how many threads the entirety of common workloads (games, productivity, content creation, business applications, scientific applications, server applications, etc.) uses it could look a bit like in the chart below (no real numbers, just for illustrative purposes).



There are a lot of workloads that are rather lightly threaded - either because the algorithm in itself cannot be parallelized, or due to dependencies because of user interaction like in games, or due to the developer being lazy (although this might be the minority after almost two decades of multi-cores being common).

But the point is this:
Most workloads fit into one of the following categories. Either they benefit from few very powerful cores. Or they scale rather indefinitely with the amount of cores.

So as a CPU manufacturer that needs to pay attention to PPW as well as PPA, there is a break-even-point where adding 1 further big core is less beneficial in the grand scheme than adding 2 or even more small cores for the same area and/or consumption.

As Intel is making a living with that, I am pretty sure that they heavily investigated in that matter in order to reach design decisions. And as of now it looks like they came to the conclusion that adding more than 8 big cores is not worth it in comparison to adding another 4 small cores for each big core they do not include.

Increasing the amount of small cores while staying on the same number of big cores for Raptor Lake supports that line of thinking. And if I had to guess I would say that when AMD will start producing hybrid SKUs that they might also stay with 8-12 big cores.

The big cores will get bigger for single-thread-tasks - but they might also get fewer, while the small cores will increase in numbers.

 

[Zucker2k]

The small cores are no joke.

This.

Gracemont: Revenge of the Atom Cores

This article can be considered a part 2 to our Golden Cove article because today we are looking at the other core in Alder Lake, Gracemont. Which is in my opinion more interesting than Golden Cove …

chipsandcheese.com

 

[Hulk]

Having experimented and worked with my 12700k for a month+ now I think 8+16 would be perfect for me. The 8 P's handle any foreground user latency task well for me and having 16 E's for background tasks like rendering, processing RAW files, etc... would be more than enough. I'm actually hoping for a 13700k that is 8+12 because that would do it as well.

 

[Leeea]

anyone here a BIG fan of overclocking?

Not worth it.

Ohh, just saying AMD Is the BEST processor 2021

Depends on the metric.

Performance per watt -> Apple wins, but $$$$
Absolute best performance -> Intel wins, but $$$ + water + good AC
Budget -> Intel wins, those last gen i3 10k units can be picked up real cheap
Performance per $ -> AMD won in 2021, knocking it out of the park with the APU line
Good enough performance at a good enough price*? -> AMD won

*may change in 2022, but as of 2021 Alderlake mainboards are still to expensive.

 

[The Hardcard]

This.
View attachment 55014

Gracemont: Revenge of the Atom Cores

This article can be considered a part 2 to our Golden Cove article because today we are looking at the other core in Alder Lake, Gracemont. Which is in my opinion more interesting than Golden Cove …

chipsandcheese.com

Exactly. Gracemont efficiency is area efficiency not power efficiency. Virtually no one needs more than eight strong threads (are there even people who need more than eight strong threads at a time).

Nearly always, if not always, using more than eight threads means multi-threaded. With the E cores getting half the performance in one quarter of the space - put otherwise double the performance in the same space as a Golden Cove core - I’m curious if anyone can define a scenario where a ninth or higher number P core is better than adding a four pack of E cores based on pure performance alone. I’ll be waiting to see if someone can pull that off.