AMD Readying 16-core "Zen 4" CCDs Exclusively for the Client Segment with an Answer to Intel E-cores?

[Wirko]

The picture you show, where the whole L3 (maybe? including L3-Control, L3-Tags and L2-ShadowTags) are only in the V-Cache-Layer. I find that problematic.
And the area the Zen4c cores take in your picture, again seem not necessary to me in the layout you chose,
because in the Zen3 layout below full equally big cores are placeable easily where the L3 sits normally.
So for me it makes nearly no sense area wise or efficiency wise.

I find it problematic for another reason.
With V-Cache as we know it now, AMD can choose between two variants: a "normal" one-layer package and a "super-cached" two-layer package, considerably more expensive and with a thermal tradeoff. This affords them quite a lot of flexibility. I'm sure each of them will find its place in servers and HPC clusters.
Now with this new proposed configuration, only the latter can exist. That's unless AMD also puts a part of L3 cache, certainly slower, on the I/O die.

 

[SaLaDiN666]

I think we are in for some pretty awesome competition with Raptor Lake and Zen4 this year. Competition I don't think we have seen since Core 2 and Phenom days to be honest.

Phenom I was an ultra trash product and never competed with Core2, it couldn't.

Phenom II couldn't compete either due to cheap Q6600 and was slower as well.

 

[ghazi]

The main problem I have with this... leak? theory? idea? is the area consideration. It doesn't make sense to have "E-cores" using the same uarch as "P-cores"; while as another poster pointed out it does greatly simplify things from a software perspective, it also largely defeats the point. If I'm not mistaken Intel's entire cluster of eight "E-cores" uses less area than two "P-cores". This is key, because the extra MT performance is coming almost free of charge. If you have to use up your die area for more big cores, why would you downclock them? Now, that maybe makes sense on a power constrained part, but if you have the full cores in silicon, why not unleash their full potential on top SKUs with higher TDPs? You pay all that money for the silicon just to gimp it? Come on. This only makes sense if, say, you allege there's a 16-core CCD and that AMD will do this to make it fit in a 28W laptop.

Now I did see a poster above talking about a slimmed down "Zen 4C". Some quick optimizations could keep the uarchs fundamentally similar while drastically cutting down on the die area through elimination of some of the logic that's relatively expensive for relatively little return (like AVX512 on ADL) and does not have substantial impacts on feature support (unlike AVX512 on ADL). That I can believe and I would of course be surprised if such a thing was not in the works. The big.LITTLE thing actually does work out fairly well for consumers, since very few people need the full ST performance on more than 8 cores, while per-core performance is mostly irrelevant to the highly parallel workloads and more smaller cores crank out more throughput at the same cost to power/area, and the execution turned out to be less rocky than I and others had anticipated.

 

[Melvis]

The 'big-little scheme' seems to be working.

https://www.newegg.com/intel-core-i7-12700f-core-i7-12th-gen/p/N82E16819118359
Intel Core i7-12700F $329.99

AMD Ryzen 7 5800X - Ryzen 7 5000 Series Vermeer (Zen 3) 8-Core 3.8 GHz Socket AM4 105W None Integrated Graphics Desktop Processor - 100-100000063WOF - Newegg.com

Buy AMD Ryzen 7 5800X - Ryzen 7 5000 Series Vermeer (Zen 3) 8-Core 3.8 GHz Socket AM4 105W None Integrated Graphics Desktop Processor - 100-100000063WOF with fast shipping and top-rated customer service. Once you know, you Newegg!

www.newegg.com

AMD Ryzen 7 5800X $378.98

Yeah again your just linking from USA which means nothing to the real world when it comes to actual prices, as I will give you a eg

https://www.pccasegear.com/products/52253/amd-ryzen-7-5800x-processor

https://www.pccasegear.com/products/56718/intel-core-i7-12700f-processor

and if your lucky you can find the 5800x on AUS Amazon for as low as $450 (last week actually)

and what is this! your linking Hardware unboxed? are you feeling ok? you hate those guys!

Phenom I was an ultra trash product and never competed with Core2, it couldn't.

Phenom II couldn't compete either due to cheap Q6600 and was slower as well.

Yeah wrong sorry dude, Phenom II destroyed Core 2 in both price and performance back then it was only when Core i whatever came out that AMD had no real competition for intels high end and never recovered till Zen. But Phenom II came out late and it came against the end of the Core 2 lineup and the begining of the Core Series so it never really looked great, but was a good CPU if you couldnt afford Core i 920 etc

 

[RedBear]

It kind of sounds like wishful thinking from Intel fans who are (rightly or wrongly) scared by Amd's hybrid approach.

Hypothetically speaking, as long as they can keep the first place in benchmarks and reviews they could simply make few parts and ask for an arm and a leg for them. I'm looking forward to how they're going to price the i7 12800KS in order to test this hypothesis.

Fixed that for you, ironic as shit the real world where both do the same shit for benches, completely different ways, but the tudes not warranted, you could be wrong.

You seem to be a bit confused, the thread is about what appears to be a completely hypothetical hybrid AMD CPU, somehow conjectured by someone for some reason, that doesn't exist, hasn't been announced and hasn't been leaked in any way whatsoever. Intel's fans can't be scared by AMD's CPUs that don't exist and at any rate my comment referred to this fantasy CPU in the first place, people aren't speculating about fantasy Intel CPUs with YUUUGE dies. Of course Intel does offer senseless halo products that are essentially made to pad the benchmarks, price aside even the basic i9 12900K counts as that if you consider the power draw, but that's beyond the point of this discussion.

 

[Why_Me]

Yeah again your just linking from USA which means nothing to the real world when it comes to actual prices, as I will give you a eg

https://www.pccasegear.com/products/52253/amd-ryzen-7-5800x-processor

https://www.pccasegear.com/products/56718/intel-core-i7-12700f-processor

and if your lucky you can find the 5800x on AUS Amazon for as low as $450 (last week actually)

and what is this! your linking Hardware unboxed? are you feeling ok? you hate those guys!


Yeah wrong sorry dude, Phenom II destroyed Core 2 in both price and performance back then it was only when Core i whatever came out that AMD had no real competition for intels high end and never recovered till Zen. But Phenom II came out late and it came against the end of the Core 2 lineup and the begining of the Core Series so it never really looked great, but was a good CPU if you couldnt afford Core i 920 etc

I admit got a bit of pleasure knowing that it must have been killing that shill to post that review. With that said I wonder why Oz is getting hosed on the prices of LGA 1700 cpu's. It seems to also be happening in a few other countries as well.

 

[Melvis]

I admit got a bit of pleasure knowing that it must have been killing that shill to post that review. With that said I wonder why Oz is getting hosed on the prices of LGA 1700 cpu's. It seems to also be happening in a few other countries as well.

Haha no no they just post what is true and fair and if there is a mistake they will openly admit it on youtube and also add other links to other reviews to back up there claims and or mistakes. When it comes to Aus and prices its always been this way for the past 20yrs at least, intel here in Aus no matter what has always been the more expensive buy, very rare you would find it cheaper and in NZ its even worse! Those guys get utterly ripped off big time. Its better to buy second hand honestly. This also comes down to games as well, we get ripped off on games and pay double the price compared to others, eg back 4 blood when it came out was $90 for the base version, and $150 for the Ultimate version, just to much!

 

[TheoneandonlyMrK]

It kind of sounds like wishful thinking from AMD fans who are (rightly or wrongly) scared by Intel's hybrid approach.

Hypothetically speaking, as long as they can keep the first place in benchmarks and reviews they could simply make few parts and ask for an arm and a leg for them. I'm looking forward to how they're going to price the 5800X3D in order to test this hypothesis.

Confused, nah you brought up the 5800X3D and I'm confused.

@Why_Me your calling HWunboxed a shill , now that's f£#@ing ironic, you may not be,I don't know but your actions are considerably more schill like them HWUB IMHO.

 

[gruffi]

It kind of sounds like wishful thinking from AMD fans who are (rightly or wrongly) scared by Intel's hybrid approach.

Scared of Intel's hybrid approach? You are kidding, right? Intel's hybrid approach is an utter joke right now. Their P-cores are more power efficient than their E-cores. Their E-cores prevent using AVX512 on the P-cores. The E-cores cause more problems than they solve.

The only reason why Intel needed Gracemont is because their P-core is quite area inefficient. AMD don't have such problems. They have a much more area efficient P-core. As they said, "AMD is the only team that can deliver a performance core and an efficient core in the same core". And I think that's a much better approach for the desktop, laptop and server markets. Developing a performance core and deriving a power efficient core from it.

 

[RedBear]

Confused, nah you brought up the 5800X3D and I'm confused.

V-cache isn't likely to be cheap, so the 5800X3D might kind of work as a proof of concept for increasingly expensive CPUs with gains limited to specific workloads.

Scared of Intel's hybrid approach? You are kidding, right? Intel's hybrid approach is an utter joke right now. Their P-cores are more power efficient than their E-cores. Their E-cores prevent using AVX512 on the P-cores. The E-cores cause more problems than they solve.

The only reason why Intel needed Gracemont is because their P-core is quite area inefficient. AMD don't have such problems. They have a much more area efficient P-core. As they said, "AMD is the only team that can deliver a performance core and an efficient core in the same core". And I think that's a much better approach for the desktop, laptop and server markets. Developing a performance core and deriving a power efficient core from it.

As they said using the same cores in that way will result in a large and expensive chip, it might be the better approach in terms of performance, but it's not necessarily the better approach in terms of price-to-performance ratio.

 

[gruffi]

As they said using the same cores in that way will result in a large and expensive chip

Why do you think so? AMD's Zen 3 core is almost as area efficient as Intel's Gracemont on comparable nodes. And Zen 4(c) will be much better than Zen 3, of course. As long as the deign is scalable I don't see any problems for AMD with large and expensive chips. At least not compared to Intel.

 

[TheoneandonlyMrK]

V-cache isn't likely to be cheap, so the 5800X3D might kind of work as a proof of concept for increasingly expensive CPUs with gains limited to specific workloads.

As they said using the same cores in that way will result in a large and expensive chip, it might be the better approach in terms of performance, but it's not necessarily the better approach in terms of price-to-performance ratio.

I'm not going to bother debating with you, you flip flop like mad and your convinced you have it sussed, I await your much better chip design though, I just hope it's cheap.

I'm also not sure what your on, you seem to think this a fan theory.

I'm seeing you spouting fan theories about Intel's hybrid approach being better than This theory of AMDS.

I'll await release reviews before I'm that outspoken, but I am already of the opinion Intel's way has merits but also has a slight inadequacy.

And was done out of necessity.

To beat Amd in single core Intel had to build a massive core ,are they winning by much, no.

And to have any chance of competing on core count and multicore loads they couldn't fit enough big cores in.

And then there's the power use on those big guns.

 

[RedBear]

I'm not going to bother debating with you, you flip flop like mad and your convinced you have it sussed, I await your much better chip design though, I just hope it's cheap.

Yet, you are. Funny, isn't it? For the record, I don't engage in speculating chip designs and I don't see how I would have flip flopped at all, my point was and is that this doesn't look like a practical design, at least from a cost perspective, something that the author of the article acknowledges, but you fail to consider.

I'm also not sure what your on, you seem to think this a fan theory.

In the article I see only your name being credited for the tip, there's no other source, as far as I'm concerned it is a fan theory until and unless it's proved to be otherwise. At the very least it's explicitly mentioned in the article as a theory, not a leak or a somewhat reliable rumour coming from an industrial source (anonymous or otherwise), even if it's not a "fan theory" tout-a-court it would still appear to be mere speculation.

IAnd was done out of necessity..

Of course it was done out of necessity, Intel is coming from how many 14nm iterations? But it's said that necessity is the mother of invention.

Why do you think so? AMD's Zen 3 core is almost as area efficient as Intel's Gracemont on comparable nodes. And Zen 4(c) will be much better than Zen 3, of course. As long as the deign is scalable I don't see any problems for AMD with large and expensive chips. At least not compared to Intel.

I'm not sure about that, do you have exact measures of each core? The difference in die area between the C0 (215.25 mm²) and H0 (162.75 mm²) Alder Lake dies is relatively small, so the eight Gracemont cores seem to occupy a fairly small space in comparison to the space occupied by each core on a Zen 3 CCD (205.7mm²). At any rate, even admitting that it wouldn't be a problem for AMD, it would still be a problem for the end users (us), AMD doesn't have an unlimited amount of wafers at its disposal, a significantly larger CCD would translate in definitely less CPUs and possibly much higher prices (both from the higher manufacturing cost and the lower availability).

 

[TheoneandonlyMrK]

Yet, you are. Funny, isn't it? For the record, I don't engage in speculating chip designs and I don't see how I would have flip flopped at all, my point was and is that this doesn't look like a practical design, at least from a cost perspective, something that the author of the article acknowledges, but you fail to consider.

In the article I see only your name being credited for the tip, there's no other source, as far as I'm concerned it is a fan theory until and unless it's proved to be otherwise. At the very least it's explicitly mentioned in the article as a theory, not a leak or a somewhat reliable rumour coming from an industrial source (anonymous or otherwise), even if it's not a "fan theory" tout-a-court it would still appear to be mere speculation.

Of course it was done out of necessity, Intel is coming from how many 14nm iterations? But it's said that necessity is the mother of invention.

I'm not sure about that, do you have exact measures of each core? The difference in die area between the C0 (215.25 mm²) and H0 (162.75 mm²) Alder Lake dies is relatively small, so the eight Gracemont cores seem to occupy a fairly small space in comparison to the space occupied by each core on a Zen 3 CCD (205.7mm²). At any rate, even admitting that it wouldn't be a problem for AMD, it would still be a problem for the end users (us), AMD doesn't have an unlimited amount of wafers at its disposal, a significantly larger CCD would translate in definitely less CPUs and possibly much higher prices (both from the higher manufacturing cost and the lower availability).

I provided a news tip is all, you can argue with yourself, trust I'll disagree in all likelihood, seams to be my path

Alleged AMD Ryzen 7000 'Zen 4' 16 Core & 8 Core Desktop CPUs Spotted

AMD's next-generation Ryzen 7000 'Zen 4' 16 & 8 core desktop CPUs have allegedly been spotted over at the MilkyWay@Home database.

wccftech.com

AMD Ryzen 7000 "Zen4 Raphael" 8-core and 16-core engineering samples have been spotted - VideoCardz.com

AMD Ryzen 7000 gets an early showcase on the BOINC platform Despite not launching till later this year, the first engineering samples of the Zen4 desktop CPUs are already out in the wild. Two models from AMD Family 25 (Zen4 Raphael) have been spotted over at the MilkyWay@Home project website...

videocardz.com

 

[gruffi]

I'm not sure about that, do you have exact measures of each core?

At least some numbers you can find on the net:

Zen 3
Alder Lake

1x Zen 3 + L2: ~3.2 mm²
4x Gracemont + L2: ~8.8 mm² / ~2.2 mm² per core
1x Golden Cove + L2: ~7 mm² / as Locuza mentions "with the black bar" ~7.4 mm²


Now let's put these sizes into perspective according to the core performance. Unfortunately I didn't find single core benchmarks of each core capped at a certain power target. Which would be most useful to see what's possible with high scalable core designs. But ComputerBase made a detailed test where we can find some interesting numbers. Based on the average results of Cinebench R15/20/23 and POV-Ray.

8x Golden Cove, w/ SMT, 3.9 GHz: 81%
8x Golden Cove, w/o SMT, 3.9 GHz: 62%
8x Gracemont, 3.9 GHz: 43%

So, one Golden Cove core (including SMT) offers almost 90% higher performance than Gracemont at ~4 GHz.

Comparing to Zen 3 is a little more difficult.

1x Zen, 3.6 GHz: 90%
1x Golden Cove, 3.6 GHz: 100%

Unfortunately this is without the impact of SMT. But due to the fact that the 12900K is also ~10% faster in MT than the 5800K @ 3,6 GHz in this review let's assume a comparable SMT speedup for Zen 3.

Now put all the numbers together, baseline is Gracemont:

1x Gracemont (w/o SMT): 100% performance, 2.2 mm² => ~45.5% perf/mm²
1x Golden Cove (w/ SMT): 188% performance, 7 mm² (7.4 mm²) => ~26.9% (25.4%) perf/mm²
1x Zen 3 (w/ SMT): 169% performance, 3.2 mm² => ~52.8% perf/mm²


So, I was not quite correct. According to these numbers Zen 3 isn't "almost as area efficient as Intel's Gracemont". Actually Zen 3 is more area efficient than Intel's Gracemont and about twice as area efficient as Golden Cove. Of course, this is a rough comparison. So, take these numbers with a grain of salt. But it clearly shows why AMD don't need separate cores. Current Zen already is very power and area efficient. And Zen 4 will improve that by a significant margin.

 

[Wirko]

At least some numbers you can find on the net:

Zen 3
Alder Lake

1x Zen 3 + L2: ~3.2 mm²
4x Gracemont + L2: ~8.8 mm² / ~2.2 mm² per core
1x Golden Cove + L2: ~7 mm² / as Locuza mentions "with the black bar" ~7.4 mm²

Gracemonts are probably bigger - there are unrecognized structures above and below each core cluster, and to the right of their L3 slice (violet coloured). It looks like at least some of that belongs with the cores but who knows.

 

[lexluthermiester]

I wonder if that shared L2 cache will be a problem?...