Speculation: Ryzen 4000 series/Zen 3

[DrMrLordX]

I personally don't believe Warhol is real.

I hope not. It just lets them push Zen4 desktop (Raphael?) out even further.

Milan ES was spotted in June through publication of reports. There was talk of Milan ES being in the wild for many months prior to that initial public disclosure by Red Hat devs. I don't for a second buy that AMD wouldn't begin testing its flagship line over a year before it releases.

The Big Cloud Boys should already have early silicon, and probably got it months ago.

@Hans de Vries

The last time AMD had to change desktop platform, they had early Bristol Ridge with a limited number of A320 boards in a bundle that was only available in a few countries. People spotted it in something like Oct. 2016 or whenever it was. Granted that didn't seem to help much since the actual AM4 launch in March 2017 was a mess due to massive x370 board shortages that lasted for months.

With Matisse, there were plenty of boards but there were CPU shortages, and they weren't even changing platform.

 

[A///]

Push further? 12-18 month cadence. AMD has yet to falter on that. We're just about 14 months in since the last major CPU release.

The Big Cloud Boys should already have early silicon, and probably got it months ago.

June or July for Red Hat leaks. I was hearing about ES in the wild with the major cloud companies and partners as early as February.

 

[DrMrLordX]

Push further? 12-18 month cadence. AMD has yet to falter on that. We're just about 14 months in since the last major CPU release.

Sure, just keep telling yourself that. Meanwhile, the dubious leaked roadmap that had Warhol on it had Raphael solidly in 2022, though as we all know, roadmaps are marketing tools first and foremost, so everything is fuzzy on timing. In any case late H1 2022 for Raphael is a pretty distinct possibility if Warhol actually hits the market. No Warhol means AMD has to pull in Raphael earlier.

 

[A///]

Sure, just keep telling yourself that. Meanwhile, the dubious leaked roadmap that had Warhol on it had Raphael solidly in 2022, though as we all know, roadmaps are marketing tools first and foremost, so everything is fuzzy on timing. In any case late H1 2022 for Raphael is a pretty distinct possibility if Warhol actually hits the market. No Warhol means AMD has to pull in Raphael earlier.

Just keep telling myself that? It's AMD's own statement. Zen to Zen+ was 13 months, Zen+ to Zen 2 was 14 months. If AMD shipped products on 12/31/2020, they'd still have over a week before hitting 18 months. It would be within their proclaimed 12-18 month cadence. You keep doubting every comment on the matter, be it from AMD or someone who quotes a C-level employee of AMD, yet offer zero argument as to why they are wrong and you feel your opinion, or lack thereof, is correct.

The very last date AMD can deliver under that cadence windows is January 7th 2020, marking 18 months to the day they released Zen 2.

AMD can deliver Zen 4 on July 7 2022 and that day marks 18 months from the release of Zen 3. If AMD is releasing Warhol AKA Zen3+ in '22 then that means no one will see Zen 4 as the major upgrade until January 7th 2024. Again, going by 18 month intervals. By then you're looking at past 5nm.

If AMD can't push out an update with a performance increase in 18 months, it doesn't bode well for their engineering prowess, and that they were a three hit wonder.

 

[A///]

For what it's worth, some random on reddit posted this a while back but it was removed by admins on one of the subs.

I know a few editors at goss sites and they all got hit with this email about a month ago and again a week or two ago. They chose not to publish it because it sounded and, really it does, sounds incredibly stupid. To me it seems like pure coincidence the dates follow a pattern. I don't believe it, but I do believe reviews for the processor will go up following AMD's event on the 8th of October, and it'll become available for sale after in the weeks following with reviews of the RDNA2 cards coming in during the 3rd week and up for same day sale maybe on the 28th.

There was that baseless rumor of AMD considering doing a choose your own bundle when it came to ordering processors and video cards through them.


<------ Rumor crap taken from reddit ---- >

Ryzen 1800X was released on March 2nd 2017
Ryzen 2700X was releasing on April 19th 2018

1y,1m,2w,3d between those

The entire launch of 3000 was released on July 7th 2019

1y,2m,2w,4d

Moi predicts Ryzen 4000 will release Octobere 26th.

1y,3m,2w,5d

 

[DrMrLordX]

Zen+ to Zen 2 was 14 months.

15 months. April 2018 -> July 2019.

Zen2 to Zen3 will be 15 months, 16 months if you count general availability. Assuming the less-than-12c SKUs hit in November, which they might not. And you're still arguing cadence (which is now stretched as far as 18 months, which is frankly ridiculous), when you are missing the central point I'm trying to make.

Warhol lets AMD stall. No Warhol means Zen4 has to be pulled in to refresh AMD's product lineup. Simple as that. Doesn't matter if they stay within some 18 month time limit or whatever. No matter how good is Zen3, it will get stale sooner or later.

 

[Nereus77]

I haven't seen any slides from AMD about any Zen 3+. Only Zen 3 -> Zen 4 -> Zen 5. 'Warhol' or whatever is just the rumor mill entertaining itself.

 

[moinmoin]

Warhol lets AMD stall. No Warhol means Zen4 has to be pulled in to refresh AMD's product lineup.

The consumer only Zen+ let AMD stall as well. And I'm pretty sure that wasn't originally planned, we know AMD used the additional time won for Zen 2 to backport features originally planned for Zen 3. These things happen.

 

[A///]

15 months. April 2018 -> July 2019.

You may want to actually look at the dates before posting absolutes.

Zen+ 2700X Hard Launch: April 19th 2018
Zen 2 series Hard Launch: July 7th 2019

14 months and 2 weeks.

 

[A///]

Zen2 to Zen3 will be 15 months, 16 months if you count general availability. Assuming the less-than-12c SKUs hit in November, which they might not. And you're still arguing cadence (which is now stretched as far as 18 months, which is frankly ridiculous), when you are missing the central point I'm trying to make.

If you want to go by this asinine route, then Intel's 10th gen HEDT never launched because their numbers are still slim nearly a year into their launch. They're almost unheard of unless you want a crappy low end processor. X299 was a massive failure by Intel and AIB partners are still pissed off at Intel for screwing them over, because now they're stuck with unsold inventory.

Warhol lets AMD stall. No Warhol means Zen4 has to be pulled in to refresh AMD's product lineup. Simple as that. Doesn't matter if they stay within some 18 month time limit or whatever. No matter how good is Zen3, it will get stale sooner or later.

There is no confirmation Warhol is real. Do you understand this? There were two leaks; one in May and one in August. Both are part of the same image. Both are leaked by the same person. The same individual has posted shoddy leaks including stupid YoY IPC increases in chart form for Intel. And if you want to be clear on the topic, then I suggest going back where someone stated that even Mebiuw, the individual who posted both images, felt that the timelines were not current. In other words, he or she felt that these were old plans. Even if you go by a 15 month timeline, you're still launching Zen 4 in 2023.

If Zen 3 launched on October 1st, and you have your "Warhol" launching in 15 months as a refresh, that is a January '22 launch. 15 months from that is well into '23. Or if you want to go the XT route:

Zen 3 launched October 1st 2020. "Warhol" refresh launches exactly one year from then just like XT did in July for Zen 2. That would be October 1st 2021. Followed by a 15 month time span for Zen 4. For Zen 4, you'll be launching on January 1st 2023.

Warhol won't offer up a large improvement in clocks and IPC. In fact, there may be a slight regression. It's also a trash base for DDR5 and other hurdles. It is the Zen+ approach to Zen4, sans the bug fixing.

 

[A///]

I haven't seen any slides from AMD about any Zen 3+. Only Zen 3 -> Zen 4 -> Zen 5. 'Warhol' or whatever is just the rumor mill entertaining itself.

The only reason I'm cautious about it is because it sounds so dumb and it's the same person/source for both "leaks" who didn't and still cannot date it correctly. They're not very confident of its current state if it were real. They're merely a middle man in the rumor mill.

FWIW, Same dude also leaked that AMD snapped up the majority of Hauwei's cancellations. This same board discredited that idea and that AMD was and is going to be hamstrung on wafers "forever." Are we really going to pick and choose a guy's spotty record? He's fairly accurate, and I Wouldn't lump him with crap sites like respected and loved, or some of the YouTubers, but he's been wrong before.

 

[eek2121]

Logistics wise they need a Zen3 chiplet + IO die combination which can go either into an AM5 or an AM4 package.

In this way they can deliver the right CPU quantities during the AM4 to AM5 transition, instead of having to guestimate the production allocations 3 months earlier.

If you are producing millions and milions of CPUs you better make sure that supply meets demand.

I wouldn't be surpriced if Vermeer is the AM4 version and Warhol is the AM5 version, with the latter just waiting for the AM5 motherboards.

This theory is likely the closest to correct.

5nm is already in a good state with better yields than 7nm had during this time in it’s lifespan. AMD already has working 5nm silicon in the labs. Zen 4 based chips will be on time. If Warhol is real it is something akin to a new stepping or new socket. It will likely launch well before Zen 4 is due to launch.

 

[eek2121]

If you want to go by this asinine route, then Intel's 10th gen HEDT never launched because their numbers are still slim nearly a year into their launch. They're almost unheard of unless you want a crappy low end processor. X299 was a massive failure by Intel and AIB partners are still pissed off at Intel for screwing them over, because now they're stuck with unsold inventory.



There is no confirmation Warhol is real. Do you understand this? There were two leaks; one in May and one in August. Both are part of the same image. Both are leaked by the same person. The same individual has posted shoddy leaks including stupid YoY IPC increases in chart form for Intel. And if you want to be clear on the topic, then I suggest going back where someone stated that even Mebiuw, the individual who posted both images, felt that the timelines were not current. In other words, he or she felt that these were old plans. Even if you go by a 15 month timeline, you're still launching Zen 4 in 2023.

If Zen 3 launched on October 1st, and you have your "Warhol" launching in 15 months as a refresh, that is a January '22 launch. 15 months from that is well into '23. Or if you want to go the XT route:

Zen 3 launched October 1st 2020. "Warhol" refresh launches exactly one year from then just like XT did in July for Zen 2. That would be October 1st 2021. Followed by a 15 month time span for Zen 4. For Zen 4, you'll be launching on January 1st 2023.

Warhol won't offer up a large improvement in clocks and IPC. In fact, there may be a slight regression. It's also a trash base for DDR5 and other hurdles. It is the Zen+ approach to Zen4, sans the bug fixing.

By your logic Zen 3 shouldn’t launch until next year. Were you the source of that rumor?

XT was a pipe cleaner for minor process changes for Zen 3. It was also a bit of a celebratory release. AMD got such negative press for doing that, they won’t be doing it again.

 

[A///]

By your logic Zen 3 shouldn’t launch until next year. Were you the source of that rumor?

Err no. You may want to read what I wrote again and edit your post. Zen 3 is due next month or launching in November. I expect a 12-18 month cadence per AMD's official words through their C-level employees who've stated as such.

A refresh for Zen3+ that brings little in terms of performance gain and merely allows them to launch DDR5... You gonna launch a flagship series of boards with that or wait until Zen 4? Is this an AM4 socket meant for DDR5 adoption? Or an AM5 socket meant for early DDR5 adopters? Will there be a secondary flagship board series for actual Zen4? Will the AIBs update two different board systems? Will the pandemic still affect AMD even though they purchased a majority of Huawei's wafers? Apple moving to 5nm should clear up a lot of wafer spaces for AMD. Will AMD still have trouble then? You could attack the Warhol theory a hundred ways until it doesn't make sense financially. AMD has multiple teams working on different generations. Are we lead to believe they have to split the AM4/AM5 DDR5 launch down the middle? Why would AMD waste resources on this move? They'll only serve to piss people off.

In the end, AMD spends more, consumers are pissed and find themselves needing to purchase a new interim board before AM5 boards come out. The current AM4 pinout has reserves, but I've been told it wouldn't be able to support DDR5.

AM4 Pinout Diagram.pdf

A. B. 1. 4. 5. 6. 7. 8. 9. 10. 11. VSS. DP0_TXN[2]. RSVD. VSS. DP2_TXP[1]. VSS. DP2_AUXP. DP0_TXN[1]. F. G. H. I. J. K. L. M. N. O. VSS. P_GFX_TXP[0]. P_GFX_TXN[0 ]. VSS. P_GFX_TXP[3]. P_GFX_TXN[3 ]. VSS. P_GFX_TXP[6]. P_GFX_TXN[6 ]. VSS. P_GFX_TXP[9]. P_GFX_TXN[9 ]. VSS. 13. TEST11. 14...

www.docdroid.net

AM4+ or early AM5 launch. You piss customers off either way. And why just DDR5? Why not offer PCIe5 too on Warhol. Might as well get that out of the way, too. Make a AM4+ board for the crossover. Great. Probably good for that processor only, maybe low end AM5 processors assuming same pinout. Short lifespan on the AM4+ board. Pissed off customers. Bump it AM5, early teething problems galore. Pissed off customers.

Warhol is a red herring. If you do a crossover refresh, you don't adopt one new feature. You testbed it with multiple features from AM5 and let the customers dogfood that. By the time AM5 rolls around in that proposed scenario, all the features are taken care of. But you have a half-baked, short life span AM4+ socket.

 

[soresu]

Err no. You may want to read what I wrote again and edit your post. Zen 3 is due next month or launching in November. I expect a 12-18 month cadence per AMD's official words through their C-level employees who've stated as such.

A refresh for Zen3+ that brings little in terms of performance gain and merely allows them to launch DDR5... You gonna launch a flagship series of boards with that or wait until Zen 4? Is this an AM4 socket meant for DDR5 adoption? Or an AM5 socket meant for early DDR5 adopters? Will there be a secondary flagship board series for actual Zen4? Will the AIBs update two different board systems? Will the pandemic still affect AMD even though they purchased a majority of Huawei's wafers? Apple moving to 5nm should clear up a lot of wafer spaces for AMD. Will AMD still have trouble then? You could attack the Warhol theory a hundred ways until it doesn't make sense financially. AMD has multiple teams working on different generations. Are we lead to believe they have to split the AM4/AM5 DDR5 launch down the middle? Why would AMD waste resources on this move? They'll only serve to piss people off.

In the end, AMD spends more, consumers are pissed and find themselves needing to purchase a new interim board before AM5 boards come out. The current AM4 pinout has reserves, but I've been told it wouldn't be able to support DDR5.

AM4 Pinout Diagram.pdf

A. B. 1. 4. 5. 6. 7. 8. 9. 10. 11. VSS. DP0_TXN[2]. RSVD. VSS. DP2_TXP[1]. VSS. DP2_AUXP. DP0_TXN[1]. F. G. H. I. J. K. L. M. N. O. VSS. P_GFX_TXP[0]. P_GFX_TXN[0 ]. VSS. P_GFX_TXP[3]. P_GFX_TXN[3 ]. VSS. P_GFX_TXP[6]. P_GFX_TXN[6 ]. VSS. P_GFX_TXP[9]. P_GFX_TXN[9 ]. VSS. 13. TEST11. 14...

www.docdroid.net

AM4+ or early AM5 launch. You piss customers off either way. And why just DDR5? Why not offer PCIe5 too on Warhol. Might as well get that out of the way, too. Make a AM4+ board for the crossover. Great. Probably good for that processor only, maybe low end AM5 processors assuming same pinout. Short lifespan on the AM4+ board. Pissed off customers. Bump it AM5, early teething problems galore. Pissed off customers.

Warhol is a red herring. If you do a crossover refresh, you don't adopt one new feature. You testbed it with multiple features from AM5 and let the customers dogfood that. By the time AM5 rolls around in that proposed scenario, all the features are taken care of. But you have a half-baked, short life span AM4+ socket.

Warhol could be an IOD move to 7nm, and/or possibly higher end features like 2.5G ethernet and more PCIE4 lanes baked into the IOD so it doesn't need a toasty chipset with its own tiny fan.

 

[A///]

Warhol could be an IOD move to 7nm, and/or possibly higher end features like 2.5G ethernet and more PCIE4 lanes baked into the IOD so it doesn't need a toasty chipset with its own tiny fan.

It's a vaiable idea if you go down your suggested route. Otherwise a shrink has no real benefit to the IO die. I can see more lanes being offered, more SATA for rusters or older SSDs, USB4, etc. 2.5Gbe solutions at the moment are a small chip. I don't think I've come across a SoC or IOD with one integrated into it. AMD were or are looking into layered silicon spanning multiple facets. Even the Intel solution is a small chip on a card or mobo.

 

[DrMrLordX]

There is no confirmation Warhol is real. Do you understand this?

Yup. It fits a pattern but I do understand that it is just a leak. Regardless, if they DON'T release a product like Warhol, they may be riding Vermeer for a very long time. As to whether or not they should launch AM5 with Zen3 or Zen4 . . . OEMs would be more than happy to launch AM5 with Zen3, and then refresh product months later with Zen4. Look at Intel: they launched Z490 with Comet Lake, and they'll probably refresh the Z490 lineup for Rocket Lake, assuming something else doesn't go wrong.

And yeah when I see 14 months 2 weeks, to me, that is 15 months. So sue me.

(as far as Intel's 10th-gen HEDT goes . . . it may as well have never launched! Good luck finding one!)

The consumer only Zen+ let AMD stall as well. And I'm pretty sure that wasn't originally planned, we know AMD used the additional time won for Zen 2 to backport features originally planned for Zen 3. These things happen.

Maybe? Pinnacle Ridge did win AMD some performance improvements. It bought them a small bump in IPC and a small bump in clockspeed, and presumably had bugfixes. If a hypothetical product like Warhol can move things forward the same way Pinnacle Ridge did, then I might actually embrace the product and buy one. If it's another Matisse XT then forgetaboutit.

 

[soresu]

Regardless, if they DON'T release a product like Warhol, they may be riding Vermeer for a very long time.

As of October 8th Matisse will be 15 months old.

Going by that I'd say Vermeer could go through to mid Q1 2020 at least before you can really say it's getting long in the tooth.

 

[french toast]

Zen 3 being on 5nm is improbable to say the least.
Warhol actually makes sense for a higher binned 7nm+ Zen 3 CCD + new generation IOD chiplet on new cheap process.

As others have stated AM5 + 5nm + Zen 4 is a heavy bug ladden lift for AMD, it is also a fairly long lead time ( mid 2022) Having a new IOD made for AM5/DDR5/PCIE5 - mid 2021 whilst having slightly higher performing zen 3 would freshen up the product stack and allow them to iron out the bugs of AM5 before the Zen 4 hammer drops mid 2022.

Be interesting to me to see if AMD does some of tick tock strategy for both the CCD and IOD, alternating between refresh and new architecture between the two + process optimisations, but never all three at the same time =constant improvement, more manageable transitions.

Ryzen 5000 = Zen 3 ( new architecture) 7nm+ (mature node/refresh) IOD 12nm ( first generation AM4 refreshed, mature cheap node)

Ryzen 6000 = Zen 3 + ( Same architecture revised/binned) 7nm+ ( same process more mature) IOD 2 ( New IOD, new features-AM5, new cheap node 6nm?).

Ryzen 7000 = Zen 4 ( Heavy improvement to Zen 3) N5P (New Node) IOD 2= minor revision/ more mature process.

And so on, big redesigns on some chips whilst a small steady revision on other chips, alternating on process between new and optimisation, constant progression, intelligent manageable sized improvements, metronomic road map delivery, build up of trust, market share increases, stock increases.

Think that is Lisa Su's long term plan when she and others kick started this process many years ago, big bets and fortune telling now starting to bare fruit.

 

[A///]

Zen2 to Zen3 will be 15 months, 16 months if you count general availability. Assuming the less-than-12c SKUs hit in November, which they might not. And you're still arguing cadence (which is now stretched as far as 18 months, which is frankly ridiculous), when you are missing the central point I'm trying to make.

You're confusing OEMs such as Dell and HP with AIBs. I don't see the point in dropping $500-800 every 12 months for a 10% increase in performance, when I can wait 14-17 months and get maybe 17-20% increase in overall performance, or even more if major ground was broken.

Cadence was always set between 12 and 18 months. Zen to Zen+ was over 12 months. You could hardly call Zen+ an amazing feat by any stretch of the imagination. They added some performance, but also fixed what they had messed up and left on the table with Zen. AMD's last release saw the release of their entire retail box lineup. This pissed reviewers off, because not only did they have a bunch of work to do, they also released RDNA on the same day. If anything, I expect Ryzen 7 and 9 to launch first. These are expensive, but compared to Intel they're very affordable. If that 10 core rumor by Yuri is remotely accurate, we may see a 4800X, 4900X and a 4950X launch. 10, 12 and 16 core. Though I question why they would launch the 10 core since CML will have been out since April, and RKL at 8 cores for 5 Ghz or whatever's been stated won't be here until sometime in 1H21.

Would you prefer them to be on a strict 12 month cadence with much less YoY growth and charging the same price as they do now, with each generation increasing in price? Or would you prefer them to take their time and launch a great product every 14-18 months? You want them to be like Intel when it comes to releases.

Yup. It fits a pattern but I do understand that it is just a leak. Regardless, if they DON'T release a product like Warhol, they may be riding Vermeer for a very long time. As to whether or not they should launch AM5 with Zen3 or Zen4 . . . OEMs would be more than happy to launch AM5 with Zen3, and then refresh product months later with Zen4. Look at Intel: they launched Z490 with Comet Lake, and they'll probably refresh the Z490 lineup for Rocket Lake, assuming something else doesn't go wrong.

Again, what? What are you basing this info on? That specific leak? It says Raphael is due for 2022. Ok, let's go with that.

Say Warhol is real. It opens up the complex issue in my prior post where the entire idea becomes a money pit for everyone involved. It isn't a bad money pit, but it is a money pit. Now, say Warhol and that leak is fake. Let's say Raphael is due in 2022. It doesn't state a month. Is this what you're worried about? If AMD launches Zen 3 on October 8th, same day, at 14-16 months, they're still in that sweet January-March 2022 window. Right in line with their cadence.

The XT chips didn't garner favorability with people. And if they do launch an interim product on AM4+/AM5 to get people on it, you do realize that not only do people have to pony up extra costs not involved with the development of prior generation B550 and X570 motherboards, but they aren't good for more than a single or maybe two generations. Assuming Zen 4 is operable with a Zen 3+ motherboard. Early DDR5 will cost a fortune. Buggy dogfooded platform, high development costs that won't pay off in a single generation of product, and expensive hardware for the end user which won't see major benefit until DDR5 is mature.

And yeah when I see 14 months 2 weeks, to me, that is 15 months. So sue me.

You're only 50% through a month. Neither a majority nor a minority.

(as far as Intel's 10th-gen HEDT goes . . . it may as well have never launched! Good luck finding one!)

I know people... professionals, who've been waiting for their orders to come through. BH was selling retail boxes a few weeks ago but it might have been a right moment in time type of thing.

 

[JujuFish]

You're confusing OEMs such as Dell and HP with AIBs

AIB just means add-in board. You're literally talking about a graphics card. Nvidia's FE cards are AIBs. You're thinking AIB partners.

 

[A///]

AIB just means add-in board. You're literally talking about a graphics card. Nvidia's FE cards are AIBs. You're thinking AIB partners.

AIB Partners

The AIB term isn't even accurate today in the way that it's used. It's become a catch-all for mobo makers and GPU cards, when the original definition for AIB didn't apply to these.

Edit: NVidia themselves used add in card as opposed to add in board. Boards in the old days were large and cumbersome. Add in cards is more accurate nowadays. If you used a mezzb today you could call it a true AIB.

 

[lightmanek]

Sorry for joke post, but I need to lower tensions here

You are all missing the glaring hint AMD is sending us, Zen 3 unveil is on 8th of October, clearly it will be manufactured on Samsung 8nm process same as RTX 30xx series.
Radeon RX6000 unveil is on 28th ... yes, you guessed it right! GloFo 28nm chip is coming!

:Hides behind sofa:

 

[Thunder 57]

Sorry for joke post, but I need to lower tensions here

You are all missing the glaring hint AMD is sending us, Zen 3 unveil is on 8th of October, clearly it will be manufactured on Samsung 8nm process same as RTX 30xx series.
Radeon RX6000 unveil is on 28th ... yes, you guessed it right! GloFo 28nm chip is coming!

:Hides behind sofa:

Gotta do something with GloFo to satisfy the WSA

 

[NostaSeronx]

As of October 8th Matisse will be 15 months old.

Actually, it is 25 months old from the RTG sampling date.

The Radeon Technology Group (RTG) has received its first Zen 2 sample!

The RTG has just received its first Zen 2 sample (to optimize for) and it's really impressive. 8C/16T 4.0 GHz/4.5 GHz DDR4-3600 CAS 15 Radeon RX Vega 64 LE __________________________ The good: It's already nibbling at the Core i7-8700K. The bad: It crashes a lot. The ugly: It crashes all...

hardforum.com

Spec-wise it is identical to the 7/7/2019-AMD Ryzen™ 7 3800X.

Ax Stepping => Late 2018
Bx Stepping => Mid 2019

Summit Ridge for example was a mid-2016 sampling processor for Q4 HVM. In which it did Doom at E3 2016.

Zen ->
Leapfrog design teams -> Zen2(one team) + Zen3(another team)
Based on some factors the Zen3 team was pushed to Zen5. Which lead to another team to do Zen3, which also happened to be the Zen4 team. Through those factors Zen4's node is also Zen3's node.

If one would look up the people doing Zen3/Zen4. They would find out they have only done single node architectures in the format of next-gen microarchitecture(AVX) -> refined microarchitecture(AVX2).

14nm -> 7nm -> 5nm(To TSMC's Bugatti race track) is equivalent to 45nm -> 32nm -> 28nm(GloFo's bumpy road), keeping cadence relative to the foundry is very important.
32nm=130CPP -> 28nm=114CP
vs
7nm=57CPP -> 5nm=57CPP&51CPP.

Cadence Collaborates with TSMC to Advance 5nm and 7nm+ Mobile and HPC Design Innovation

Cadence announced its continued collaboration with TSMC to further 5nm and 7nm+ FinFET design innovation for mobile and high-performance computing (HPC) platforms.

www.cadence.com

7nm+ and 5nm seem to have same design flows and design rules. Most cases if 57nm is used, it is potentially 1:1 with 7nm+.

Setting up for some weird play from TSMC:
7nm family equals: N7, N6(w/ N7 RTO), N7+, N5(w/ N7+ RTO)
5nm family equals: N5(w/ 51CPP), N4(NTO & RTO)