Intel's response to RyZen.

[CatMerc]

inb4 dual Summit Ridge for R7 1900X. (probably not)

 

[french toast]

not sure if it's a pun or referring to that TM? The reason I posted in this thread because it's my speculation on AMD's strategy to respond to Intel's response, be it soft or aggressive.

That was one of the names amd copywrighted.

 

[OatisCampbell]

It doesn't need to: AMD didn't "Conroe" Intel this time, they just barely catched up, with skylake still holding IPC and clock advantage. AMD is offering more cores at disruptive prices right now but that will be probably adressed with price cuts or different segmentation, Intel sold 6 cores haswell for not much more than 4 cores today so they could repeat that.



As I explained in my post both Cannonlake and Icelake µarch must be ready, heck they displayed a working 10nm Cannonlake so that's done and exist 100%.
It may be just 5%, but it adds to Skylake advantage, then 14nm++ and some clock gains... they won't come with Skylake-X because that already done (with its own possible gains against desktop arch.) but there's no reason to not bring all IPC gains to Coffe Lake if it comes end 2017/2018. Tick tock may be dead and 10nm late but cpu designers worked anyway these years.

I disagree. In a way, AMD "Conroe'd" here.

It has been over a decade since AMD had a performance leading cpu. They been teetering on bankruptcy, selling assets, laying off staff and losing staff to bigger companies all those years. Even when they had a better CPU (by far) they only had a 25% market share. (and last I saw they're in the single digits)

For AMD to not only compete with, but beat intel in 2017 was about as likely as Blackberry launching a better phone than Samsung this year.

When intel "Conroe'd" AMD they were a company 10X the market cap and 3X the marketshare of AMD, with inferior products.

I doubt people will even realize the magnitude of what AMD has done here, this is Biblical. David and Goliath kind of stuff. I have a 4790K and I might buy one of these things just on principle. I truly hope a lot of people buy RyZens, the competition in the market can't be bad for us.

Which is not me saying intel has been gouging, supply and demand has kept prices sane in AMDs absence. It is me saying I'd always rather have a choice than not, and kudos AMD.

 

[jihe]

Hopefully intel responds by dropping the f***ing awful TIM and goes back to soldered heat spreaders. The persistence of intel is a giant FU to the consumers.

 

[Spartak]

[OT] is there an established definition of 'performance crown' or we just have one for ST and another for MT? The way I judge performance personally is (ST + MT) / 2

 

[lolfail9001]

is there an established definition of 'performance crown' or we just have one for ST and another for MT? The way I judge performance personally is (ST + MT) / 2

We have 3, actually. One for ST, one for ST+MT (ala HEDT) and one for MT.

 

[PingSpike]

That idle power consumption is crazy good omg.

AMD has actually had pretty excellent idle numbers in a lot of benchmarks with their APUs on FM2, usually beating Intel actually. Its not really trumpeted because their load power draw was terrible and the performance was clearly worse. But they did look good before you started using them. Idle power is actually a pretty important number in a lot of cases IMO, not as important as performance per watt but not nothing either.

 

[Spartak]

We have 3, actually. One for ST, one for ST+MT (ala HEDT) and one for MT.

yeah makes sense I was just curious if the ST+MT rating is formalised in some way

 

[bjt2]

What mythical top part?

Maybe, except that there are still no signs of "top part" to speak of. And yes, Gen-Tier is the nomenclature.

If all 1800x top at +100 - +200 XFR boost and the OC are all similar, then it's not impossible that AMD is saving the higher clocking chips for a later 1900X release... We know that the process is highly variable at that stage. I don't think that every 1800X will top just above the default without outliers if they are not picked away from the retail chips...

 

[tamz_msc]

Ryzen was possible because Intel slacked off. We're not excited about Zen because of how it beats Bulldozer, we're excited while comparing with top of the line Intel consumer CPUs.

All that Intel had to do was to ride on the wave of the Nehalem-Sandy Bridge transition, and couple it with the incremental changes in architecture along with their (upto now) lead in process technology to get where it is today.

Compare it with AMD which was trounced by Conroe, to which they responded with a dud(Phenom), then caught up with a better Phenom II, but again had their thunder stolen by Nehalem. Finally when they could move to a new process, they started with a design that was flawed in many ways, and Sandy Bridge was the final nail in the coffin for them. They had to spend 3-4 years squeezing every last bit of performance from a weak architecture stuck on an obsolete node, while in the meantime Intel had made two node jumps AND got FinFET. On top of that, AMD was piling on losses, had to sell off assets, and were almost written off as a contender in the high performance during this period of time.

So the fact that they pulled it off, despite going through a turbulent decade and their relative size and resources available compared to Intel, tells more about AMD's engineers and their direction as a company than Intel who seem to look for short-term gains in their existing market strongholds, regardless of whether their engineers slacked off the past couple of years.

TL;DR by saying that Ryzen happened because Intel became lazy is too much of a discredit to AMD, in my opinion.

 

[coercitiv]

TL;DR by saying that Ryzen happened because Intel became lazy is too much of a discredit to AMD, in my opinion.

I'm not talking about the architecture and physical manifestation, that is entirely AMD's merit (and they deserve a lot of credit for ), I'm talking about the way Zen is perceived by the customer. You were also talking about perception (huge leap in relative performance over BD).

 

[Spartak]

If all 1800x top at +100 - +200 XFR boost and the OC are all similar, then it's not impossible that AMD is saving the higher clocking chips for a later 1900X release... We know that the process is highly variable at that stage. I don't think that every 1800X will top just above the default without outliers if they are not picked away from the retail chips...

I don't think they'll cherry pick away, since they'd want XFR to add to the hype and can sell all they produce in the first few months. More like a next stepping that will give them more headroom. It could be they rated the 1800X on the conservative side just to be sure they can add something in the 3.8/4.2 range in a month or 3. And have top notch XFR performance to boot.

 

[Glo.]

By adding 12 core CPU to Skylake-X lineup, Intel basically maintains very high margin on highest end part. 1700+$ 7950X will be 12 C/24T. 1099$ will be 10C/20T. 8C/16T - 599$, 6C/12T - 399$.

They have single threaded performance lead with Skylake, they still can maintain high margin prices, for their parts. From marketing point of view, countering both ends, its a masterpiece from Intel.

 

[Spartak]

By adding 12 core CPU to Skylake-X lineup, Intel basically maintains very high margin on highest end part. 1700+$ 7950X will be 12 C/24T. 1099$ will be 10C/20T. 8C/16T - 599$, 6C/12T - 399$.

They have single threaded performance lead with Skylake, they still can maintain high margin prices, for their parts. From marketing point of view, countering both ends, its a masterpiece from Intel.

I don't believe they'll cheapen their 8C/16T by that much. They'll mostly make 6 core the new 4 core and have the higher ST / slightly lower MT performance 6C part as the one to combat AMD's 8 core chips on price / performance.

Only if Zen+ proves a real threat to ST (and increase the MT lead) will they need to make 8C truely competitive. Right now I'd say 12C/24T $1700+, 10C/20T $1200+, 8C/16T $700+ and 6C/12T $300+

 

[Drazick]

This is amazing what is going on here.

Trying to see the large picture.
We were told for years that there is one thing Intel does best - Design x86 CPU's if not design CPU in general.

And what we see?
A company with very limited resources starting basically form scratch can get very very close to Intel's best.
What would have happened if a resourceful company as Apple would do that?

This is amazing and Intel response should be simple, be humble again.
Do real good engineering and improve.
Don't stick to the concepts, try to think things over.

This might be a defining moment for Intel.
Are we really as good as we think we are?
Or simply so far we had the best manufacturing and in the CPU design business we're very much like a limited resource on the verge of bankruptcy company?

Being humble and work hard, this should be Intel's response.

 

[LTC8K6]

I don't find it amazing at all. Taking so long to catch up is not really something to brag that much about.

The pricing is the real amazing thing.

Also, Intel's best has not been released yet. Intel is still on Broadwell and Skylake architecture, and will be for a little while yet.

 

[Shivansps]

Yeah, i can already see at $1700 12C/24T reeplacing 6950X, at $1000 10C/20T reeplacing 6900K. at $600 8C/16T reeplacing 6850K and 6C/12T at $400 reeplacing 6800K. They keep the prices, the big margins and performance crown. If they want to engage in a price war, they just need to price the 8C/16T $100 lower.

For the mainstream, the I3 and I5 are in trouble, the I7 will be the less affected since it has a fast ST clock, but for all of that Intel just need to launch Coffe Lake, and here it is where the AMD delays to launch 6 and 4C gonna play in Intel favor.

I really dont see a big issue here, Zen APU if priced right and Zen+ could be a much bigger problem thought.

 

[Atari2600]

Also, Intel's best has not been released yet. Intel is still on Broadwell and Skylake architecture, and will be for a little while yet.

So when will its best be released?

Saying that is a bit of a cop-out in the electronics market!!

If we have to wait 6+ months for Skylake-X (KL stepping) then someone could turn around and say AMD's best has not been released yet, Zen+ is "only" 6 months away.


I'd reckon its fair to use that excuse if your waiting <3 months.

 

[Spartak]

Yeah, i can already see at $1700 12C/24T reeplacing 6950X, at $1000 10C/20T reeplacing 6900K. at $600 8C/16T reeplacing 6850K and 6C/12T at $400 reeplacing 6800K. They keep the prices, the big margins and performance crown. If they want to engage in a price war, they just need to price the 8C/16T $100 lower.

They already sell a 6C/12T part for $389 and the 6800K for $434. They'll really need to step up their game with 6C and bring it to the $300+ range as this is the part that competes with R5/R7.

8C will need to be more competitive too but is vital to their margins. No way will they slash that price almost in half when they still have both the IPC and frequency lead. That's just wishful thinking.

 

[tamz_msc]

I don't find it amazing at all. Taking so long to catch up is not really something to brag that much about.

The pricing is the real amazing thing.

Also, Intel's best has not been released yet. Intel is still on Broadwell and Skylake architecture, and will be for a little while yet.

I wonder what Intel could have done if they were stuck on an inferior node for five years. The best of Intel is yet to come? Is it going to leapfrog SL/KL so that AMD plays catch-up for another five years? Wishful thinking.

 

[bjt2]

I don't think they'll cherry pick away, since they'd want XFR to add to the hype and can sell all they produce in the first few months. More like a next stepping that will give them more headroom. It could be they rated the 1800X on the conservative side just to be sure they can add something in the 3.8/4.2 range in a month or 3. And have top notch XFR performance to boot.

If higher clocking 1800x will surface, then you will be right. I was making an hypotethical claim.

 

[Glo.]

I don't believe they'll cheapen their 8C/16T by that much. They'll mostly make 6 core the new 4 core and have the higher ST / slightly lower MT performance 6C part as the one to combat AMD's 8 core chips on price / performance.

Only if Zen+ proves a real threat to ST (and increase the MT lead) will they need to make 8C truely competitive. Right now I'd say 12C/24T $1700+, 10C/20T $1200+, 8C/16T $700+ and 6C/12T $300+

They have to do so. Everything else will be considered very bad value. This is the earthquake that happened to this market, with Ryzen and its pricing.

 

[Spartak]

You dont understand. If you have the performance crown and you're the established market leader you will add a mark up to that. In the competitive majority part of the market you price competitive. As long as they can justify their 8C parts as equal or superior to the best of Zen they can afford a slight mark up. The enthusiast / hobbyist isnt the main market for these parts.

It doesnt make sense to predict 8C parts to be sliced in half while also stating 6C prices will barely nudge down. They are already in the $400 range and are the ones that will face about 80% of the R5/R7 market in direct competition. Only the 6850K will be outperformed and outclassed by the launch trio at its current price point.

 

[Atari2600]

This last bit may explain why Intel PR sent out a last-minute “call us before you write” email to most of the press

And so it begins....

I suppose we shouldn't be surprised...

 

[railroadmaster]

The 'war' between Intel and AMD has more or less ended. AMD will compete with some of Intel's product segments, but many segments will remain unopposed. Intel might lose some money, but Intel won't be massively affected.

Intel's problem is fundamentally itself. Intel is a large and profitable organization that has stagnated. Due to this stagnation they throw money at things that don't produce new knowledge/IP or ROI. Intel's graphics for example have only marginally improved, yet they add significant cost to Intel processors and take up significant space. Larrabee was an opportunity to bring needed improvement to their graphics, but gave up on it early. Intel Atom had great potential however Intel wasted the potential of it by subsidizing it rather than spending the money on improving it. Intel acquired companies like McAfee for no good reason. Intel has wasted money on 'diversity and inclusiveness' programs, which produce nothing for the company. Tech companies like Intel already are 'diverse' and have people from around the world, spending money on it for the sake of is not a good use of shareholders money. Intel is a company with lots of resources which it squanders rather than putting to good use. Intel is like that smart kid who has lots of potential but wastes it and ends up doing drugs and partying instead. Finally with Windows on ARM, ARM Chromebooks, ARM Linux Computers, Apple potentially adopting ARM, and ARM server Processors Intel has a challenger even bigger than AMD.

Intel isn't doomed, but unless they can find direction they will continue to stagnate and eventually decline.

Signed,

Signed disgruntled former Intel shareholder.

Also of note

http://www.pcworld.com/article/3022...three-pillarsand-the-pc-isnt-one-of-them.html