ppl stop buying Ryzens, waiting for next gen on 7nm.
Unfortunatelly, it won't be much better, up to 15% in best case

TristanXppl stop buying Ryzens, waiting for next gen on 7nm.
Unfortunatelly, it won't be much better, up to 15% in best case

Yeah, they stopped. Only twice more Ryzens are sold compared to Intel. :D
LOL "won't be much better, up to 15% in best case" while Intel getting CPUs gen by gen by a 5% jump.

TristanXppl stop buying Ryzens, waiting for next gen on 7nm.
Unfortunatelly, it won't be much better, up to 15% in best case

Hoping this isn't a case of Introlling.
A potential 15% increase is still MUCH better than Intel was dishing out with each chip refresh while requiring a new board for it too.
Sometimes even a 5% increase wasn't truly realized - Even a 7% increase beats what Intel had been doing.

from conference call: Navi in Q3, will be placed under VII from pricing standpoint, Epyc Rome in Q3 (shipping in Q2) and AMD even thinking on EPYC APU in future(?)
On Navi raytracing- Lisa didn't want to say- more info only near launch

41% Gross margins... while still undercutting their competition...

VERY Nice! :respect:

Jeez...just shows how bad RTG is doing right now. A new arch or finally fixing GCN is very much needed.

moproblems99Jeez...just shows how bad RTG is doing right now. A new arch or finally fixing GCN is very much needed.

I very much believe that the chief reason they remained profitable right now is the lack of soaking money into RTG. It's doing well enough considering it's shoestring budget.

TristanXUnfortunatelly, it won't be much better, up to 15% in best case

This we don't know better wait for actual reviews before making such claims , by the way 15% is outstanding comparing to what Intel has been offering for years now !

TristanXppl stop buying Ryzens, waiting for next gen on 7nm.
Unfortunatelly, it won't be much better, up to 15% in best case

I can see two sides of it. Zen+ was only a refresh and when you think of it that way, 15% is pretty good for a simple refresh these days! You don't see even entirely new generations do that anymore. So another 15% for the actual overhaul is still respectable in my eyes. Then you have to consider... if this winds up being true, it shows them consistently gaining ground that actually adds up to something in a reasonable time frame. It means that while they're still behind Intel in many ways, they're still accelerating much faster. Remember, Zen was... what? Spring 2017? So that would mean in just a couple years we saw a 30% performance increase on the Zen architecture family, which I guess doesn't sound like much to some, but isn't that kind of unheard of with the current state of things? Especially not without a socket swap. If you bought into Zen 1 in 2017, there is a decent chance that you can drop a chip that's at least 30% faster (more if you go up the line) into the same motherboard with nothing but a BIOS update. In 2019. Sounds like a nice upgrade to me. *shrugs*

Buuut it's all kind of moot right now. Gotta see them in the wild for a bit before anyone can really say how much of an upgrade they are. What is for certain is that from the original launch of Zen, AMD has gained an astonishing amount of ground. And it's starting to show. Even if Zen 2 only offers 15% over Zen+, that's still enough to keep them moving and surely enough to keep expanding profits and market share. This in turn leads to more R&D which leads to more significant performance increases, not to mention the need for more competition from Intel. To me, it would still be a good thing, for everyone.

AMD has made some big moves, to be sure. But I think where they have really been excelling is in the 'slow and steady' game. It's easy to look at what they've been up to and not think much of them. There are legitimate arguments to be had there. But that's also why they're succeeding. "One rung at a time" has been their MO. They put enough skin in the game to keep playing and snatching up some smaller victories, as well as a few key ones... which is really another way of saying they've focused on staying in the black as a way of advancing their position in the market. And it seems to be working. To play with the big boys when you're not nearly as big, you have to leverage effectively. As the saying goes "When you do something right, people won't know you've done anything at all." As long as you're making profits, you can always have growth. Whether it's a little or a lot, that's a good thing... but only when there is also sustainability.

You could think of it like this. Say they dumped more R&D in to their CPU's (or even their GPU's - like so many seem to expect,) reaching for that showstopper. The problem I think a lot of people don't see there is that R&D doesn't scale linearly with profitability in pretty much any area of tech. It's a balancing act. A major cost that needs to be offset as quickly as possible. The longer it takes to pick up, the more dire and limiting the situation becomes. Better to spend the money you know will make you money than go all in on what turns out to be a flop... or, just as bad, come out with the best thing ever and make no money on it because people aren't willing to pay enough to offset the huge cost difference. It is very hard to generate large performance increases without drastically increasing the price point. Every now and then R&D comes up with a silver bullet and a way to make it affordable to make turns up right in time, but that is the exception. Usually, to make a big move like that, you need to be able to afford being in the red for a good long time, with plenty of other revenue sources far ahead enough to make up a net profit. Very, very unlikely for any company to pull this off. Even if they CAN do it, it's almost never the best move in the long run. At best you take greater risks to ultimately arrive at the same point. On the surface it sounds like a no-brainer, but it's a total hail mary in practice. You see these flashes in the pan with new tech all of the time. Stuff takes off for a while, but ultimately dies because the progress made wasn't sustainable and production was never really affordable. So as the interest and excitement died off, so did the people behind the tech.

Honestly, a modestly respectable uptick with Zen 2 is exactly what I expect from them. They're simply tending to their little Zen garden. Most reasonable people, I think, don't expect any more or less. It's easy to look at the performance and say "whatever" but when you look at the numbers, you see it's actually working amazingly well for them. And people actually like their products. The niches for them continue to grow. Nobody ever thought this would happen in the beginning. From where they started till now, they've reached a point where they're becoming real contenders. A power move would be a total waste of everything they've built. Why fix what isn't broken? I'd love a supercharged Ryzen 3000, but more than that I would like to see AMD continue to pursue long-term stability and establish themselves in a way that continues to allow for a level of steady growth that the competition has been lacking in for too long now. Can't climb a mountain in an afternoon. Keep on chuggin along, setting and meeting expectations, while bringing a little something new to the table every time. Not every new launch can or should be a total revolution. It just needs to be another step forward, and one that leads to many more steps forward. That's how you do it. It's a little different from Intel's crab walk in that every launch is opening doors for them. A very good thing to see in any tech-driven industry. It's not always about who's got the best stuff. It's about who's consistently moving forward. To take and hold those positions you have to put in the time to grind-out a foundation... a way of developing and releasing tech that continues to raise the bar sufficiently and secure profits/market. One big move is not nearly enough to make that happen, and often has the opposite effect.

Basically, their next move is likely still going to be on the safe side... something they know they will profit and grow from. Zen 2 won't and doesn't need to be their Intel killer. That's never been how they've operated.

Their GPU's are a different story. But I think that too is a calculated choice. They're not taking any risks or stretching themselves thin. This may change if they ever get on top of the CPU game. Barring that I think they'll continue to float by and let Nvidia have their empire. Not much in it for them right now. I think not pursuing that is part of the reason they're more successful. They'd no doubt be hurting quite a lot right about now if they acted any differently. It's enough for them to not be a total failure, which for the most part, I think they've accomplished. GPU market changes have set them back a little, but it looks like they must've anticipated that. And what they have put forth there has been reasonably successful in my eyes - for what it is. Maybe not as much as people have hoped, but not less so than AMD expected.

Good, good.

kastriotGood, good.

For some reason I can't not read this post in an evil Emperor Palpatine voice...



Not saying anything... I just watched way too much star wars as a kid.

Proceed with topic.

B-RealYeah, they stopped. Only twice more Ryzens are sold compared to Intel. :D
LOL "won't be much better, up to 15% in best case" while Intel getting CPUs gen by gen by a 5% jump.

Twice more Ryzens are sold compared to Intel? Are you kidding? You think AMD sells 2x as many CPUs as Intel?

So basically: everything is (much) worse than before - apart from a slightly increase in gross margin.
Gross margin makes it to the article title (how much am I not surprised...?)

yakk41% Gross margins... while still undercutting their competition...

VERY Nice!

?
That's still very poor compared to Intel. Long way to go.
But sure: by asking lower prices, AMD is in fact lowering Intel's margins. Intel is few % down, AMD is few % up. But it's still 57% vs 41%.
As a result, Intel is making a lot of money and AMD is not making any.

And remember: Zen is all about gross margin. They should beat Intel here - especially since the Blues are also making other, less profitable products.
In other words: Zen CPUs are way too cheap.

B-RealYeah, they stopped. Only twice more Ryzens are sold compared to Intel. :D

In a particular store... Intel's revenue is still 10x larger and stable. AMD revenue is dropping quickly. 1st gen Ryzen lineup was more expensive than 2nd gen.

HwGeekfrom conference call: Navi in Q3, will be placed under VII from pricing standpoint

It's expected to be a slower GPU, so hard to imagine it costing more...

AMD even thinking on EPYC APU in future(?)

Workstation CPU with IGP - to compete with Xeon G lineup.

No, not to compete Xeon G lineup, remember it's EPYC, the question was more like if they will add GPU chiplete so the Compute power will be improved, so CPU and GPU compute units on single package-and you got a lot of space on EPYC :)..
This is very interesting IMO.

Mark Lipacis -- Jefferies -- Analyst :
... And I was wondering are your customers on the server side asking you for like an APU-kind-of-a product where you have microprocessor and graphics processor capabilities integrated together.

Lisa:
I think the customers have been deeply engaged with us since the first generation of EPYC. There is both software customization, as well as some hardware customization that we go through. And we feel very comfortable with our ability to address that across the cloud and the enterprise businesses. And then, as we go forward, I think we are also very excited about what we can do when we put our CPU and GPU portfolio together and really do system-level optimization for the data center.
So we view that as a early opportunity for us but one where there is a lot of opportunity to help customers really optimize for very high-performance computing applications.

TristanXppl stop buying Ryzens, waiting for next gen on 7nm.
Unfortunatelly, it won't be much better, up to 15% in best case

Troll much or are you just that ignorant?

TristanXppl stop buying Ryzens, waiting for next gen on 7nm.
Unfortunatelly, it won't be much better, up to 15% in best case

Wellll, AMD did show an 8-core that was comparable to a 9900k in cinebench, soooo.... Is that a bad thing?

Oc potential could be higher and efficiency should be better. Also, we should see more cores coming to the mainstream and we might even see more than a 15% increase from 2700x to a 12-core 5ghz model in per core performance!

If there's a 16-core I might even skip that for a better gaming 12-core with higher boost clocks, though 16-cores is better for streaming and recording, so I'm excited either way!

Aren't the numbers extremely disappointing?
120$b vs 38$b? That's a huuuuuge difference. And why is it so big?

ShadowjumpAren't the numbers extremely disappointing?
120$b vs 38$b? That's a huuuuuge difference. And why is it so big?

First of all, it's millions and second of all, it's operating income. Not a huge deal, especially considering AMD has both Navi and Zen2 to boost sales from june! Soon after servers will get 7nm too, so this slump is nothing out of the usual for a tech company that hasn't released any really new and exciting products since vega and polaris and Ryzen back in 2017.

Radeon vii is nice, but is mostly to keep up appearances and Zen+ will mostly sell better in laptops, not so much desktops where lots of people already had Zen processors.

At least AMD is nowhere near the levels of bulldozer and fiji-asco!

AMD's profit is pathetic, this needs to improve drastically when 7nm CPUs hit or there's no hope. They need to win more server clients. You'd think it would be easy given how far ahead ROME appears to anything Intel got. But I guess Intel has relationships in this field going back decades so a few phone calls and emails have no doubt been sent to clients to stop them jumping to Red.

Also, they do have 18-20m a year sold PS5 APU contract starting 2H 2020.

Shatun_BearAlso, they do have 18-20m a year sold PS5 APU contract starting 2H 2020.

Tbh I'd rather AMD use that 7nm capacity for server parts, however as you said it's not easy to get those big contracts & Intel still has a deathgrip there. However do note that it's loosening, slowly but surely!

Huawei Launches 2U 4-node Taishan X6000 V2 Server with Kunpeng 920 Arm Processor

R0H1THowever do note that it's loosening, slowly but surely!

Amazon going for it is a pretty big stab at Intel. So long as AMD stays ahead when it comes to cores to cost ratio, I think we'll continue to see this sort of shift. Phasing out old hardware takes time and doesn't happen overnight.
aws.amazon.com/ec2/amd/

Plus with AMD they improve their margins, with Chiplets AMD can give them big discounts even like Intel, since it's cost them like nothing compare to EPYC prices.
Even if Intel wants to make more 28 Xeons- their Yield are very low compared to tiny AMD's chiplets.
P.S: Did you see the power usage of the new XEON 8280? even agains EPYC 32C is very High, now consider 7nm EPYC.

HwGeekNo, not to compete Xeon G lineup, remember it's EPYC, the question was more like if they will add GPU chiplete so the Compute power will be improved, so CPU and GPU compute units on single package-and you got a lot of space on EPYC :)..
This is very interesting IMO.

Why not a competitor to Xeon G? They don't have anything in this segment.

Putting different chiplets in a package is an interesting option, but again: more likely for workstations. They could combine 3 faster CPU chiplets and one GPU or FPGA.

In a server you're less limited by space and cooling, so it's better (more flexible) to go full CPU in the package and add whatever PCIe cards you need.
With mixed chiplets it'll be very hard to use 100% of CPU and sockets are very expensive.

AquinusAmazon going for it is a pretty big stab at Intel. So long as AMD stays ahead when it comes to cores to cost ratio, I think we'll continue to see this sort of shift. Phasing out old hardware takes time and doesn't happen overnight.
aws.amazon.com/ec2/amd/

Amazon offers some instances based on EPYC. We don't know how many CPUs they actually bought, so we don't know how big the stab could be.

Amazon says EPYC instances are 10% cheaper for the same performance. Well, in some loads there are, in some they aren't.
Phoronix tested this not long ago:
www.phoronix.com/scan.php?page=article&item=amazon-ec2-m5a&num=1

Of course that's for tasks that use all the CPU potential available.
AMD is yet to show a server CPU that matches Xeons on single-core performance. This may happen with Zen2, but don't hold your breath.

Hugh MungusFirst of all, it's millions and second of all, it's operating income. Not a huge deal, especially considering AMD has both Navi and Zen2 to boost sales from june!

Suggesting that income is not a huge deal is rather brave. Would you say that on a recruiting interview? :)
AMD is not making a lot of money - that's a fact.
And it's not obvious whether 7nm will help that or not. It's still an expensive process. We roughly now how much Apple pays TSMC for a chip. If AMD's rates are similar, they'll have to raise prices compared to Zen+.

At least AMD is nowhere near the levels of bulldozer and fiji-asco!

True. But they still haven't got to neither revenue nor margins they had in the K10 period (~2010).

HwGeekP.S: Did you see the power usage of the new XEON 8280? even agains EPYC 32C is very High, now consider 7nm EPYC.

Roughly 10% difference (in line with AWS instance cost difference - shocking isn't it? :P).

I'm not sure if I'd call that "very high", but certainly it's an advantage of the more modern architecture.

HwGeekfrom conference call: Navi in Q3, will be placed under VII from pricing standpoint, Epyc Rome in Q3 (shipping in Q2) and AMD even thinking on EPYC APU in future(?)
On Navi raytracing- Lisa didn't want to say- more info only near launch

So, if we take into account that Rome is supposed to be released before, does it mean that there will be no Zen 2 consumer grade CPUs to at least Q4?

No. IMO Ryzen 3000 will be out on June 7th, this is why all the MB teasing X570.

notbI'm not sure if I'd call that "very high", but certainly it's an advantage of the more modern architecture.

Look it TH review, Epyc uses 16x memory dims vs 12 on Xeon 8280:
Now when you look at large scale Data Center- this is big, before even looking at 64C Rome.