Discussion Speculation: Zen 4 (EPYC 4 "Genoa", Ryzen 7000, etc.)

[Vattila]

Except for the details about the improvements in the microarchitecture, we now know pretty well what to expect with Zen 3.

The leaked presentation by AMD Senior Manager Martin Hilgeman shows that EPYC 3 "Milan" will, as promised and expected, reuse the current platform (SP3), and the system architecture and packaging looks to be the same, with the same 9-die chiplet design and the same maximum core and thread-count (no SMT-4, contrary to rumour). The biggest change revealed so far is the enlargement of the compute complex from 4 cores to 8 cores, all sharing a larger L3 cache ("32+ MB", likely to double to 64 MB, I think).

Hilgeman's slides did also show that EPYC 4 "Genoa" is in the definition phase (or was at the time of the presentation in September, at least), and will come with a new platform (SP5), with new memory support (likely DDR5).

What else do you think we will see with Zen 4? PCI-Express 5 support? Increased core-count? 4-way SMT? New packaging (interposer, 2.5D, 3D)? Integrated memory on package (HBM)?

Vote in the poll and share your thoughts!

 

[A///]

If anything, gaming is by far the biggest part of DIY. I can't emphasize that enough. And for good reason, pretty much everyone else is better off buying a prebuilt. Yes DIY is small in the scheme of things. You do have the e-penis crowd which can be quite profitable.

Do you have data to back this claim up or is it confirmation bias based on what you see and mass blanket evaluate the sector?

 

[A///]

It's still early yet to say final perf on Zen2, given how many BIOS updates and OS scheduler tunings you tend to get for signficant uArch changes.

Likewise I'd say the unified L3 change for Zen3 should be a doozy for those successive tuning updates on its own, even without whatever else comes with it.

That's one of the reasons I held off of buying Zen2. I wasn't in the mood to play with BIOS updates. I know other people who don't mind them. You say doozy here but doozy usually implies a negative thing. I know that's not what you meant, but it's a little weird. Lol

AMD has gone on record to saying they're working with MS. I have no idea what that implies unless they meant the scheduler and the custom processor for the surface. Or they're working with MS on those in addition to making Windows 10 better adapted for Zen or chiplet structure in its most basic form. Say what you will about W10 updates, but the fact they can push major improvements "quickly" is a god send.

 

[soresu]

The amount of bad uarchs in history outnumber good ones. Keller is an excellent engineer, but placing too much emphasis on his work when there were more hands on people involved with Zen, than him, is a disservice to all the others.

I noted this on another thread recently, it sounds like they pulled together a great team at AMD that continued on executing well/great after Keller left in 2016.

There was some comment/rumour last year I think that some of them had been redirected to course correct the RTG work (so likely RDNA), though who knows if it was baseline uArch for RDNA2+ or just sanding down rough edges for RDNA1.

 

[soresu]

I know that's not what you meant, but it's a little weird. Lol

I use it because of Groundhog Day and Steve Tobolowsky, whatever the context

 

[soresu]

If anything, gaming is by far the biggest part of DIY. I can't emphasize that enough. And for good reason, pretty much everyone else is better off buying a prebuilt.

I was at Uni and interacted with a group of Music Tech students not so long ago, they all DIY'd their own build for maximum silence.

I try to do the same myself, just because I like to listen to music and watch films with my PC, and the noise can be bothersome in quieter parts.

 

[soresu]

AMD has gone on record to saying they're working with MS. I have no idea what that implies unless they meant the scheduler and the custom processor for the surface. Or they're working with MS on those in addition to making Windows 10 better adapted for Zen or chiplet structure in its most basic form. Say what you will about W10 updates, but the fact they can push major improvements "quickly" is a god send.

Phoronix recently tested Threadripper 39xx on Windows and Linux, and it seems they didn't have the same problems as the 29xx generation on Windows, so definitely something going on there in terms of collaboration.

Previously certain workloads on the Threadripper 29xx showed a huge delta with Linux in front, now it seems Windows is back up on top with 39xx.

As for the "major improvements quickly" part, it's both a blessing and a curse - it means as you say big hardware related improvements don't have to wait too long, but it also brings that terrible unstable nature of new updates that W10 has been plagued by.

 

[jpiniero]

Do you have data to back this claim up or is it confirmation bias based on what you see and mass blanket evaluate the sector?

It's pretty much been that way since DIY popularity exploded in the late 90's. There's a reason discrete GPU gaming has always been front and center in reviews. Really until relatively recently OEMs had mostly ignored discrete GPU gaming because the market was too small so you didn't have much choice in the matter.

 

[A///]

It's pretty much been that way since DIY popularity exploded in the late 90's. There's a reason discrete GPU gaming has always been front and center in reviews. Really until relatively recently OEMs had mostly ignored discrete GPU gaming because the market was too small so you didn't have much choice in the matter.

You're conflating better performing parts with lesser performing parts. Akin to saying the only reason 9900Ks are sold are for gaming, while ignoring any actual benefit high IPC and high clock speeds offer up. DIY didn't just explode in popularity in the late 90s. You're timeline is off. It was earlier than that.

 

[A///]

Phoronix recently tested Threadripper 39xx on Windows and Linux, and it seems they didn't have the same problems as the 29xx generation on Windows, so definitely something going on there in terms of collaboration.

Previously certain workloads on the Threadripper 29xx showed a huge delta with Linux in front, now it seems Windows is back up on top with 39xx.

As for the "major improvements quickly" part, it's both a blessing and a curse - it means as you say big hardware related improvements don't have to wait too long, but it also brings that terrible unstable nature of new updates that W10 has been plagued by.

I don't keep up with TR but I believe some changes in TR3 made that possible. Current performance of TR3 on Windows and Linux is probably that and due to an issue with Linux which Phoronix covered in a separate article. These were published in August, I think. No ETA on Linux patches for improvement.

 

[jpiniero]

DIY didn't just explode in popularity in the late 90s. You're timeline is off. It was earlier than that.

I guess you could call it mid-late 90s if you wanted. Earlier than that the market was more about whiteboxes but that was more of a commercial operation than anything else.

 

[A///]

I guess you could call it mid-late 90s if you wanted. Earlier than that the market was more about whiteboxes but that was more of a commercial operation than anything else.

Show a source for your claims.

 

[soresu]

I guess you could call it mid-late 90s if you wanted. Earlier than that the market was more about whiteboxes but that was more of a commercial operation than anything else.

Ah yes, the era of a small boulder weight CRT monitor resting on top of your desktop case.

Suffice to say picture quality loss aside, the difference in weight and volume of a CRT is not something I will miss, certainly not my back in any case.

 

[A///]

Ah yes, the era of a small boulder weight CRT monitor resting on top of your desktop case.

Suffice to say picture quality loss aside, the difference in weight and volume of a CRT is not something I will miss, certainly not my back in any case.

I have some good memories of those old CRTs including a heavy Sony my eldest brother was planning on giving me after he switched to those early LCDs but that thing broke internally. I got the LCD instead and used it for many years. I still have it. It still works. It's beat to hell and back including the base being sawed off in certain places to fit an old desk I had about 13 years ago as a second monitor for chats while online gaming.

 

[jpiniero]

Show a source for your claims.

The website you are on now? AT is one of the earliest hardware review sites out there, and you could say they had an impact of making DIY more popular.

 

[A///]

The website you are on now? AT is one of the earliest hardware review sites out there, and you could say they had an impact of making DIY more popular.

Anandtech is a tech news website. You can be a DIY builder but not be an enthusiast. You can be a gamer and seek a middle option rather than the high end options. The take rate of high MSRP parts is lower than that of middle-ground parts.

AT isn't an enthusiast website. Their focus has always been general tech news and what's new in the hardware world. This isn't an extreme gaming site, an extreme overclocking site, or one specific to anything in particular.

To be so blase about labels, it would be akin to saying HuffPost is for dilettantes and dilettantes only, and all who seek HuffPost are most definitely dilettantes.


Here's an example of a source.

"In Europe, for example, the majority of processors being sold are 3600X and 9700Ks. This is a result of the higher end AMD Ryzens being more expensive and getting less sales put on, and the higher end 9900K, KF and KS extremely hard to get and months in preorder."

"Can you prove what you say is true?"

"Yes, here is a link to Mindfactory's sales charts. They don't cover all of Europe, but they cover a decent amount of countries and have always provides monthly and yearly sales charts, as well as data down to specific motherboards. As you can see in the below chart, the 3600X does have a much higher take rate than the 9700K due to supply, but also because the 3600X offers value for money. However, the majority of motherboards paired with this processor are on the B450 chipset, specifically MSI's offerings.

https://imgur.com/a/UK8dTad

Below is an RMA data interpretation for the Zen generation. These were published in March 2018. While this doesn't reflect Zen+ and Zen2 motherboard data, it gives you an idea of just how much more popular a middle ground option is for the vast majority of consumers.

RMA quotes AM4

Sheet1 MSI,Asus,Asrock,Gigabyte Board,Sales,RMA %,RMA count,Board,Sales,RMA %,RMA count,Board,Sales,RMA %,RMA count,Board,Sales,RMA %,RMA count B350 PC MATE,6520,3,195.6,Strix B350-F,3040,2,60.8,AB350M Pro4,2580,8,206.4,GA-AB350M-DS3H,140,0,0 B350M PRO-VDH,6020,3,180.6,X370-Pro,4890,5,244.5,Fat...

docs.google.com

I can't offer you global sales data, but Mindfactory is one of the largest sellers in Europe and they're as well known as Amazon."


That is how you provide source and not hearsay.

 

[UsandThem]

AT isn't an enthusiast website. Their focus has always been general tech news and what's new in the hardware world. This isn't an extreme gaming site, an extreme overclocking site, or one specific to anything in particular.

Maybe not so much anymore, but Anandtech most definitely an enthusiast site back in the day. Their target audience was those who wanted a deep dive into hardware technology. Times surely have changed, and most tech sites now offer general tech news to cater to a general audience. How sites like Tom's and Anandtech now are isn't how they always were.

Just because they weren't a hardcore gaming site, or extreme overclocking site, doesn't mean that people who wanted a deeper understanding of how hardware worked and performed weren't "enthusiasts".

 

[Thunder 57]

Maybe not so much anymore, but Anandtech most definitely an enthusiast site back in the day. Their target audience was those who wanted a deep dive into hardware technology. Times surely have changed, and most tech sites now offer general tech news to cater to a general audience. How sites like Tom's and Anandtech now are isn't how they always were.

Just because they weren't a hardcore gaming site, or extreme overclocking site, doesn't mean that people who wanted a deeper understanding of how hardware worked and performed weren't "enthusiasts".

I appreciate that Anandtech goes much deeper into the technology rather than just reusing press material and doing benchmarks. Sometimes the reviews take longer, but they are more comprehensive then just about anywhere. They probably were even better in the past, but I still think they do a good job most of the time explaining the how and why of hardware.

 

[A///]

Maybe not so much anymore, but Anandtech most definitely an enthusiast site back in the day. Their target audience was those who wanted a deep dive into hardware technology. Times surely have changed, and most tech sites now offer general tech news to cater to a general audience. How sites like Tom's and Anandtech now are isn't how they always were.

Just because they weren't a hardcore gaming site, or extreme overclocking site, doesn't mean that people who wanted a deeper understanding of how hardware worked and performed weren't "enthusiasts".

Guess we see things differently. I'm not saying deep dives don't count towards the enthusiast, but how many deep dives are published yearly versus regular reviews a few pages long with some benchmarks of a hardware device? Some of the companies I've worked for have offered deep dives into background ops, but those don't necessarily attract upper echelon enthusiasts. The yearly conferences they hold and sponsor do, however, satisfy those wants or needs.

My issue with the poster was his repeated claims which were corrected, given into, and then another claim was made without much basis. It would have been simple for him or her to link to last month's Steam hardware survey. While Steam is just a small segment in terms of online gaming and as an e-tailer themselves, they offer some interesting data for October 2019.

Steam Hardware & Software Survey

Right off the bat we get these results.

75% of users utilize NVidia graphics cards.
Of all graphics cards, 73.17% offer DX12 on W10
80.54% of users have Intel processors
Of all the processors reported, 51.21% use 4 cores (labeled as CPUs), less than 24% use other, which is for more than 4 cores.

You can see the breakdown data:

Steam Hardware & Software Survey

GFX card breakdown:

Steam Hardware & Software Survey

One could argue that a 1060 counts as enthusiast. I really don't. That's entry level decent graphics. You could argue on price, because of NVidia's sly shift towards a higher price that the 1060 is an enthusiast card now. I could also argue the opposite on the CPU front, that the 10980XE that was just released doesn't qualify for HEDT based on price because Intel slashed 50% over a generational shift which is really a 2 year old processor with boosted clocks, and that it should only be compared with the 3950X because of the pricing bracket.

The above is an argument people have been using, not necessarily on AT, over the past week. The reality is if Intel slashed another $800 off, are we supposed to compare it to a sub $200 processor? Or compare it for what it is? You wouldn't compare the 2080ti to the rumored 5500 XT, now, would you? No.

 

[Markfw]

Guess we see things differently. I'm not saying deep dives don't count towards the enthusiast, but how many deep dives are published yearly versus regular reviews a few pages long with some benchmarks of a hardware device? Some of the companies I've worked for have offered deep dives into background ops, but those don't necessarily attract upper echelon enthusiasts. The yearly conferences they hold and sponsor do, however, satisfy those wants or needs.

My issue with the poster was his repeated claims which were corrected, given into, and then another claim was made without much basis. It would have been simple for him or her to link to last month's Steam hardware survey. While Steam is just a small segment in terms of online gaming and as an e-tailer themselves, they offer some interesting data for October 2019.

Steam Hardware & Software Survey

Right off the bat we get these results.

75% of users utilize NVidia graphics cards.
Of all graphics cards, 73.17% offer DX12 on W10
80.54% of users have Intel processors
Of all the processors reported, 51.21% use 4 cores (labeled as CPUs), less than 24% use other, which is for more than 4 cores.

You can see the breakdown data:

Steam Hardware & Software Survey

GFX card breakdown:

Steam Hardware & Software Survey

One could argue that a 1060 counts as enthusiast. I really don't. That's entry level decent graphics. You could argue on price, because of NVidia's sly shift towards a higher price that the 1060 is an enthusiast card now. I could also argue the opposite on the CPU front, that the 10980XE that was just released doesn't qualify for HEDT based on price because Intel slashed 50% over a generational shift which is really a 2 year old processor with boosted clocks, and that it should only be compared with the 3950X because of the pricing bracket.

The above is an argument people have been using, not necessarily on AT, over the past week. The reality is if Intel slashed another $800 off, are we supposed to compare it to a sub $200 processor? Or compare it for what it is? You wouldn't compare the 2080ti to the rumored 5500 XT, now, would you? No.

Steam numbers are crap. And have been proven so. They show only online gamers numbers. What about the rest of the world ?

 

[A///]

Steam numbers are crap. And have been proven so. They show only online gamers numbers. What about the rest of the world ?

By whom? How was that statement established? Are you referring to the Chinese gaming cafe numbers corruption that was addressed and fixed a year ago to prevent data corruption (https://www.extremetech.com/gaming/...steam-hardware-survey-went-nuts-last-summer)? Games sold on Steam which normally have DRM (call home ability) utilize Steam's DRM. Even if you play in offline mode on Steam, Epic, UPlay, et al. it does call in at one point or another as these games do have offline mode built into their DRM function.

Is your post referencing older games that don't utilize a call home function or weren't purchased on a digital storefront? I play a lot of Rollercoaster Tycoon 3 Platinum, maybe 20-40 hours a month. This game doesn't require an internet connection. Does that make me a hardcore, enthusiast gamer?

What about the rest of the world? Is Steam ignoring the rest of the world? Who's to say another platform isn't offering a real and accurate representation of their player base if your point stands? Or are you referring to pirated games which have had their DRM stripped? At which point, why care about those numbers?

 

[soresu]

Or are you referring to pirated games which have had their DRM stripped? At which point, why care about those numbers?

As someone who once bought games and stripped DRM to play them without the annoying need for putting in a CD/DVD, I can attest not all those that strip DRM are pirates.

DRM is an annoyance that seldom dissuades a true pirate from getting around it, it only serves to put people off, especially if it sometimes causes performance degradation too - this is why GOG has a market that no one thinks twice about the same way that they did for Epic Store.

 

[A///]

As someone who once bought games and stripped DRM to play them without the annoying need for putting in a CD/DVD, I can attest not all those that strip DRM are pirates.

DRM is an annoyance that seldom dissuades a true pirate from getting around it, it only serves to put people off, especially if it sometimes causes performance degradation too - this is why GOG has a market that no one thinks twice about the same way that they did for Epic Store.

If so, then I'd be a massive pirate because I stripped DRM from DVDs and BluRays I bought and made copies for my own use. Piracy involves distribution to unauthorized parties. An EU court made claim that even self copies could not be tolerated, I'm unaware of any landmark US case set fort in front of the US Supreme Court that touched on that subject. The cases the MPAA had were against software companies that made products such as DVD Shrink et al. that offered DRM bypass, that prevented users from making copies of movies they bought. Or software that virtualized a regional code instead of reading it off the drive firmware.

When people talk about game pirates, they mean torrent files or downloading off of a DDL or direct download link, that offers software plus a patcher. That is firmly illegal and I can't see why you'd want to count in those people. The Valve survey makes no distinction between MP online games and single player games that require a phone home function to verify that it's a legitimate copy.

According to a 2019 report, Steam's got 1B users, of which 90M are active. I think that's a very good baseline to work off of in terms of hardware being deployed.

Steam Now Has One Billion Accounts (And 90 Million Active Users)

Valve's digital PC games storefront Steam got its billionth registered user on Sunday, though it has 90 million monthly active users.

variety.com

On the other hand, Epic has 85M users according to the article below, and that 40% of their users don't use Steam.

The creators of 'Fortnite' now have more than 85 million people using their store to buy games on PC

In less than four months, Epic Games has turned 'Fortnite's' success into a power play in the PC gaming market.

www.businessinsider.com

40 percent of Epic Games Store users say they don't have Steam

Epic said at GDC that 85 million people now use its store.

www.pcgamer.com

Epic doesn't use a hardware survey. They have surveys which you fill in manually, but these are often unrelated (I play Fortnite). Steam's is anonymous and auto-sent.

 

[soresu]

If so, then I'd be a massive pirate because I stripped DRM from DVDs and BluRays I bought and made copies for my own use.

This is also common - those anti piracy adverts are the most ridiculous thing I have ever heard of in media.

It's practically a material visual definition of the term "preaching to the choir" - telling the people that have bought it not to pirate, when anyone worth their salt copying it could strip the stupid annoyance and just enjoy the film.

If anything it is so annoying that it makes me want to pirate films just to stick it to them.

 

[DrMrLordX]

By whom?

Pretty much the entire VC&G subforum. Ask there if you want clarification.

 

[A///]

Pretty much the entire VC&G subforum. Ask there if you want clarification.

That is purely anecdotal experience without hard numbers.

Querying against the site shows threads from 2011 where people criticized it based on data prior to 2011. I'm sorry, if the site has a bias against it that's fine, but suggesting it has junk data because Anand's users have some decade old grudge against it isn't a valid reason to exclude it. If it isn't allowed, then I'd like some hard data showing individual units of video cards being sold. It would only be fair, wouldn't it?

Provide a better source if you have one. Perhaps a breakdown of sales data from 2010 to 2019. Major GPU makers AMD/ATI and NVidia, the model names of cards, the amount sold each quarter or year, and some data that show capture data of equipment still in use.