Speculation: AMD's response to Intel's 8-core i9-9900K

Page 35 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

How will AMD respond to the release of Intel's 8-core processor?

  • Ride it out with the current line-up until 7nm in 2019

    Votes: 128 71.9%
  • Release Ryzen 7 2800X, using harvested chips based on the current version of the die

    Votes: 30 16.9%
  • Release Ryzen 7 2800X, based on a revision of the die, taking full advantage of the 12LP process

    Votes: 17 9.6%
  • Something else (specify below)

    Votes: 3 1.7%

  • Total voters
    178

Spartak

Senior member
Jul 4, 2015
273
128
86
Looks like we get a new intel architecture next year like I stated. And also as stated, the elements we already knew indeed pointed to a wider redesign.

They've been very clear today this is indeed their intent for 2019 (and they are probably confident enough about it's feasability to publicly state)
 
Last edited:
Feb 23, 2017
421
268
96
Sounds like "don't buy Zen 2 coz we'll have something new too" but with a whole lot of tumbleweed afterwards.
Intel is about on time as any publicly funded building project; 10nm is already 3 years late, and this latest fuzz doesn't really add anything to the conversation.
 

coercitiv

Platinum Member
Jan 24, 2014
2,962
254
136
Sounds like "don't buy Zen 2 coz we'll have something new too" but with a whole lot of tumbleweed afterwards.
It looks more like the new blood infusion they got through new hires helped whatever inside forces that were already trying to get Intel out of their ivory tower and meet the true consumer landscape. This is more of a return to normality rather than an attempt to downplay Zen.

Intel as an organization has gotten so used to winning that they built their entire business process around this key step. Their people may be able to work in adverse conditions, may be able to face defeat and rebuild on a personal level, but their organization as a whole, their collective decision making process was so hung up on the idea of obliterating their competition that the concept of building something from ground up, with all the hurdles and setbacks and maybe years of continuous commitment before you end up being truly competitive and (maybe) the best is lost to them.

This is the key I chose to see the new hires Intel made, partially with Raja and Keller, but more importantly with marketing execs from AMD: they need people with a different mind set, who understand what it takes to compete from behind, from being 2nd or even 3rd. Keeping the lead tends to hone a different set of skills than achieving it.
 

moinmoin

Senior member
Jun 1, 2017
625
145
96
It looks more like the new blood infusion they got through new hires helped whatever inside forces that were already trying to get Intel out of their ivory tower and meet the true consumer landscape. This is more of a return to normality rather than an attempt to downplay Zen.

Intel as an organization has gotten so used to winning that they built their entire business process around this key step. Their people may be able to work in adverse conditions, may be able to face defeat and rebuild on a personal level, but their organization as a whole, their collective decision making process was so hung up on the idea of obliterating their competition that the concept of building something from ground up, with all the hurdles and setbacks and maybe years of continuous commitment before you end up being truly competitive and (maybe) the best is lost to them.

This is the key I chose to see the new hires Intel made, partially with Raja and Keller, but more importantly with marketing execs from AMD: they need people with a different mind set, who understand what it takes to compete from behind, from being 2nd or even 3rd. Keeping the lead tends to hone a different set of skills than achieving it.
Indeed. I really liked how Keller (or was it Koduri? Answer is cited to both, and both were there) acknowledged their competition in his answer as if that was always a normal thing to do for Intel:
Our products will be decoupled from our transistor capability. We have incredible IP at Intel, but it was all sitting in the 10nm process node. If we had had it on 14nm then we would have better performance on 14nm. We have a new method inside the company to decouple IP from the process technology. You must remember that customers buy the product, not a transistor family. It’s the same transformation AMD had to go through to change the design methodology when they were struggling. At Apple it was called the ‘bus’ method.
 

Spartak

Senior member
Jul 4, 2015
273
128
86
Sounds like "don't buy Zen 2 coz we'll have something new too" but with a whole lot of tumbleweed afterwards.
Intel is about on time as any publicly funded building project; 10nm is already 3 years late, and this latest fuzz doesn't really add anything to the conversation.

it's quite funny to see the tinfoil AMD brigade accept everything AMD says at face value yet a very public and detailed roadmap for next year and the years after stated by Intel is completely dismissed.

You guys are hilarious.
 

Markfw

CPU Moderator, VC&G Moderator, Elite Member
Super Moderator
May 16, 2002
16,950
373
136
it's quite funny to see the tinfoil AMD brigade accept everything AMD says at face value yet a very public and detailed roadmap for next year and the years after stated by Intel is completely dismissed.

You guys are hilarious.
When AMD has been delivering and Intel is just playing games (like the 28 core fiasco), you believe them ? "Fool me once, shame on me, fool me twice, shame on you" come to mind ?
 

epsilon84

Senior member
Aug 29, 2010
931
104
136
When AMD has been delivering and Intel is just playing games (like the 28 core fiasco), you believe them ? "Fool me once, shame on me, fool me twice, shame on you" come to mind ?
I think you have the quote the wrong way around Mark ;)

Anyway, it's by far the most solid info we have and even has a 2H 2019 launch timeframe. Do people expect Intel to be stuck at 14nm forever?
 

epsilon84

Senior member
Aug 29, 2010
931
104
136
When AMD has been delivering and Intel is just playing games (like the 28 core fiasco), you believe them ? "Fool me once, shame on me, fool me twice, shame on you" come to mind ?
I think you have the quote the wrong way around Mark ;)

Anyway, it's by far the most solid info we have and even has a 2H 2019 launch timeframe. Do people expect Intel to be stuck at 14nm forever?
 

jpiniero

Diamond Member
Oct 1, 2010
5,932
98
126
Anyway, it's by far the most solid info we have and even has a 2H 2019 launch timeframe. Do people expect Intel to be stuck at 14nm forever?
For desktop, it's going to be awhile.
 

Markfw

CPU Moderator, VC&G Moderator, Elite Member
Super Moderator
May 16, 2002
16,950
373
136
I think you have the quote the wrong way around Mark ;)

Anyway, it's by far the most solid info we have and even has a 2H 2019 launch timeframe. Do people expect Intel to be stuck at 14nm forever?
You could be right about which way, but you get the idea. I am very dubious of Intel until they actually DELIVER something fresh, not like the 9900k.
 
Apr 27, 2000
10,496
349
126
Intel has to demonstrate the ability to successfully launch a client-group CPU on a process denser than their 14nm group of processes. Period. Comet Lake was a warning sign that a meaningful release might be awhile. Not to speak of the silly rebranding of chips on the mobile side.
 

epsilon84

Senior member
Aug 29, 2010
931
104
136
For desktop, it's going to be awhile.
https://www.anandtech.com/show/13699/intel-architecture-day-2018-core-future-hybrid-x86

For the high performance Core architecture, Intel lists three new codenames over the next three years. To be very clear here, these are the codenames for the individual core microarchitecture, not the chip, which is an important departure from how Intel has previously done things.

Sunny Cove, built on 10nm, will come to market in 2019 and offer increased single-threaded performance, new instructions, and ‘improved scalability’. Intel went into more detail about the Sunny Cove microarchitecture, which is in the next part of this article. To avoid doubt, Sunny Cove will have AVX-512. We believe that these cores, when paired with Gen11 graphics, will be called Ice Lake.
 

ub4ty

Senior member
Jun 21, 2017
749
307
96
It looks more like the new blood infusion they got through new hires helped whatever inside forces that were already trying to get Intel out of their ivory tower and meet the true consumer landscape. This is more of a return to normality rather than an attempt to downplay Zen.

Intel as an organization has gotten so used to winning that they built their entire business process around this key step. Their people may be able to work in adverse conditions, may be able to face defeat and rebuild on a personal level, but their organization as a whole, their collective decision making process was so hung up on the idea of obliterating their competition that the concept of building something from ground up, with all the hurdles and setbacks and maybe years of continuous commitment before you end up being truly competitive and (maybe) the best is lost to them.

This is the key I chose to see the new hires Intel made, partially with Raja and Keller, but more importantly with marketing execs from AMD: they need people with a different mind set, who understand what it takes to compete from behind, from being 2nd or even 3rd. Keeping the lead tends to hone a different set of skills than achieving it.
This is not rocket science... Their prices are to high for what they offer compared to AMD.
Their prices were too high for what they offered before AMD began competing and their customers were pissed about it because they knew they were being Milked. The same goes for Nvidia. Eventually there's a cost to this once your competitor catches up which AMD did and surpassed them IMO. A company doesn't compete or win back those people until their price/value offering are even better than their competitor as they already lost them and have to win them back. If a company thinks they haven't and continue to price their inferior product absurdly, they will continue to lose even more customers. People are now very technically competent and know how to analyze/spec products. You have mainstream reviewers discussing specs/features reserved only for "nerds" some time ago. In this, you better have a better value product.

What's value at this point? 7nm/PCIE 4.0 ... expanded PCIE slot offerings beyond Intel's gimped i/o pipe to the CPU and a dedicated nvme slot to the CPU. They lost people at 6 core, 8 core, 12 core, and 16 core to AMD. I'm not converting back to Intel's platform at these core counts. My next investment is either a 32 core Epyc or 64 core rome. Intel has no value offering here either. Thus, when AMD delivers 7nm, it's going to be a nail in the coffin. I'm not scrapping my motherboards and familiarity with AMD's platform and dynamics to go back to Intel. They'd have to undercut AMD substantially on price.

They can hire 400 Jim Kellers and it wont change a thing beyond optics.
They have to deliver an equal product and value as AMD. Until they do that, as a consumer, Intel is dead to me and that goes for any unbiased sensible consumer... And no I'm not paying double for a processor for 10-20% more performance from Intel or so that I can brag about it to people who don't care. So no, Intel gains no points by any potential performance crown. Once you go north of 6 cores, you already have a 'performance crown'. What matters then is core count. Given that I can easily buy a 16 core from AMD for less than their flagship 8 core, they still don't even win the performance crown.

Raja/Jim Keller nor any silly marketing gimmick they can slap together will cause me to be an idiot and buy an inferior product/ridiculously overpriced nonsense. So, the solution is quite simply for Intel :

Cut your silly pricing. You're no longer the top dog. So, its best you behave as though you're trying to win people back by undercutting the superior competition. If you need Jim Keller/Raj to understand that, you're already lost. That's more about optics than anything ... Something which does nothing for me as a consumer after the best value.

And for the the love of God Intel, Stop w/ all the stupid codenames. K.I.S.S, no one's buying into your utterly confusing market gimmick of 500 goofy platform names. This is another thing you can learn from AMD that got you into trouble... Simplify your offerings and stop trying to obscure your shortcomings w/ silly behind gimmicky 'MUh quantified pristine suborbital lake' desktop processor. Or muh crystalized unobtanium bronze Xeon processor. Get your head out of your behind and simplify your product/platform nomenclature. No one's buying it.... o_O
 
Last edited:

ub4ty

Senior member
Jun 21, 2017
749
307
96
Check out the Comet Lake thread. That's what the desktop is getting instead.
From the name, 'comet lake'... It seems like that is : nowhere...

Comet Lake
Who is coming up w/ these stupid names? They should be fired :eek:
 

jpiniero

Diamond Member
Oct 1, 2010
5,932
98
126
From the name, 'comet lake'... It seems like that is : nowhere...

Comet Lake
Who is coming up w/ these stupid names? They should be fired :eek:
I imagine they just look at a list of Lakes and randomly pick one.
 

ozzy702

Senior member
Nov 1, 2011
776
104
136
They can hire 400 Jim Kellers and it wont change a thing beyond optics.
They have to deliver an equal product and value as AMD. Until they do that, as a consumer, Intel is dead to me and that goes for any unbiased sensible consumer... And no I'm not paying double for a processor for 10-20% more performance from Intel or so that I can brag about it to people who don't care. So no, Intel gains no points by any potential performance crown. Given that I can easily buy a 16 core from AMD for less than their flagship 8 core, they still don't even win the performance crown.
Is there an AMD CPU that has better overall performance in non-server workloads than the 9900k? Is Intel having a hard time selling their 9900ks?

I agree, AMD has a great value proposition right now and should have a phenomenal value and potentially performance proposition in 2019/2020. I hope that their 2019/2020 lineup is stellar and that they hold both the performance and value crown.

That said, these claims that Intel somehow is behind, can't innovate/compete, etc are hilarious. When Intel's bottom line is impacted seriously they'll drop prices and not a moment sooner. They're selling more CPUs than ever and can't keep up with demand.
 

Markfw

CPU Moderator, VC&G Moderator, Elite Member
Super Moderator
May 16, 2002
16,950
373
136
Is there an AMD CPU that has better overall performance in non-server workloads than the 9900k? Is Intel having a hard time selling their 9900ks?

I agree, AMD has a great value proposition right now and should have a phenomenal value and potentially performance proposition in 2019/2020. I hope that their 2019/2020 lineup is stellar and that they hold both the performance and value crown.

That said, these claims that Intel somehow is behind, can't innovate/compete, etc are hilarious. When Intel's bottom line is impacted seriously they'll drop prices and not a moment sooner. They're selling more CPUs than ever and can't keep up with demand.
How many people say "well, I just like Intel better" and buy it regardless of value, performance or innovation. They are behind, they are not innovating and for anyone concerned with price/perf they are not competing. They may be making money, but the question is how long will the world be asleep ?
 

Zucker2k

Senior member
Feb 15, 2006
689
33
136
They are behind in what? Certainly not performance, and that's what matters most, especially to enthusiasts. All this talk about Intel standing still and not innovating, yet AMD is still behind in performance per core. In addition, Intel's mainstream platform chips offer the better AVX, and graphics for encode/decode/acceleration. That's additional value that those in the AMD camp pretend they don't see, or doesn't matter.
Newsflash: the reason why AMD's chips are cheaper is by necessity. They are 2nd tier and trying to win decent marketshare. From 2006, right up till last year year their products were relatively junk - in a number of ways. Even now, AMD still trails Intel's 4 year old architecture by a significant margin. We should be happy Intel hit a snag because AMD would've been toast by now.
 

Topweasel

Diamond Member
Oct 19, 2000
4,563
185
126
They are behind in what? Certainly not performance, and that's what matters most, especially to enthusiasts. All this talk about Intel standing still and not innovating, yet AMD is still behind in performance per core. In addition, Intel's mainstream platform chips offer the better AVX, and graphics for encode/decode/acceleration. That's additional value that those in the AMD camp pretend they don't see, or doesn't matter.
Newsflash: the reason why AMD's chips are cheaper is by necessity. They are 2nd tier and trying to win decent marketshare. From 2006, right up till last year year their products were relatively junk - in a number of ways. Even now, AMD still trails Intel's 4 year old architecture by a significant margin. We should be happy Intel hit a snag because AMD would've been toast by now.
This is a shortsighted view. The doom and gloom for Intel also fits as another shortsighted view. The problems with Intel don't really exist in this exact setting at this exact moment. So you are right but Intel is going to be doing a highwire act for the next two years and this will give AMD an oppotunity to catch up something Intel should have never let happen.

So problems.
1. No 10nm yet. Billions thrown into a process switch that will be nearly outdated by the time they can actually mass produce products.
2. Increased Demand on server front. Intel has to push out more Server dies and consumer dies will take a hit from this. Increased i7 costs.
3. AMD has pushed Core requirements on 14nm increasing on consumer level. Dies get larger meaning less margin, less dies, worse yields (yields would still be good, just less so). All under constrained 14nm production Increased i7 costs.
4. AMD has pushed core requirements on 14nm on enterprise level. Intel is now going to double up on server dies per CPU, still ending up short of AMD's total. Which means worse effectively halved yields. All on a constrained 14nm production. Intel will either rise everything up a decent amount in price or they will double up cost on these chips.
5. A decade of minor tweaks to arch has left Intel with very predictable performance benchmarks to target.
6. Bottom will fall out on ST clock growth. Making it harder to maintain specific metrics that enthusiasts look at.

Things Intel has to look forward to.
1. Significant single clore clock advantage. More value in Advertising and allows Intel to maintain a performance lead in many applications that are driven by clocks over MT capabilities.
2. Has so far kept pace on desktop core count.
3. Huge manufacturering capabilities.
4. One of if not the largest case of Mindshare for a corporation.
5. Eventual 10nm mass production should allow Intel to go back to previous margins and profits as long as they keep mindshare.
6. Largest R&D budget of anybody really. Should be able to develop a faster arch if needed.

But what this all means is for the next year or two. Intel will be making less and less CPU's out of the same production capability on an already strained capacity. Forced to spend even more money on production on an outdated by their standards process. Their Dekstop, Workstation/HEDT, and Server MT superiority was already on a razors end of failing will be illiminated. Assuming Zen2 maintains current "ipc" and clocks and doesn't get better, their core count, efficiency allowing for better clocks will shatter all but specific use-cases. Because of that and long delivery projections will allow AMD to make great leaps in Prosumer and Professional use-cases. AMD doubling up on AVX2 per core performance being the icing on the cake there. Some of the Doom and gloom has a place, for shareholders this means less margins, less volume, heavy growth by a competitor. It doesn't mean they are doomed. It doesn't mean AMD has an easy road to upstage them, they don't have any chance of that. Back when the K6-2 came out Sanders was asked if they could take the fight to Intel. He responded with something along the lines of if Compaq called and asked them to be their sole supplier for CPU's AMD would have to decline that they were a long way from having the ability to really take the battle to Intel. That applies here as well. AMD can make significant inroads into Intel's income and profits. But they can't kill Intel not now, not ever really. Only Intel can do that.
 

Mopetar

Diamond Member
Jan 31, 2011
4,293
245
126
Yes, Intel isn’t going to die. That’s ridiculous, but the chiplet approach from AMD means they can gain a lot more ground than they otherwise could. It makes much more efficient use of the resources they have available and allows them to offer some high core count products that Intel will struggle to compete with in the server space and probably outright kills Intel’s HEDT.

It almost is a perfect storm, but Intel has a lot of cushy margins they can eat into. They won’t have problems matching AMD on cost for value if they absolutely need to even if shareholders may not like it.
 

Similar threads



ASK THE COMMUNITY