Intel's response to RyZen.

[Lodix]

GF will have 7nm in mass production by mid 2018. So AMD can perfectly have a late launch that year on that process node. Just to clarify some people saying that they won't have access to 7nm until 2020 ( when GF has scheduled 5nm ).

 

[inf64]

Have you taken cooling and differences in ram speed and timings into account between reviews?
Lets take Anandtech review, they used a closed loop water cooler for their 6900k review and it STILL posted lower MT scores than Ryzen on stock air cooler! what kind of boost would ryzen see with the same cooling solution with XFR? AMD seems to have used stock 140w coolers for BOTH platforms, is it AMDs fault that intel cheaped out on their own cooler?
Also AMD used turbo boost 3 @4ghz when Anandtech didn't due to fiddling around in the bios (cant remember) so they already did something to help 6900k scores when some reviews didn't, some consumers are not going to touch the setting in bios, AMD gave them the benefit of the doubt here.
Come on be fair, lets not start spreading fud ok?

If anything it would be even better for AMD NOT to use Turbo 3.0 since 1800X would score higher @ stock than 6900K in ST test . 1800X is beating 6900K in MT test no matter what so MT test is not that important as Ryzen already has an advantage in this benchmark.

 

[tamz_msc]

Dude i explained it many times already...

Amd showed Cinebench R15 score

Ryzen R7 1800x scoring 1601

I7 6900k scoring 1474? Actually is 1578 from independent reviews..

thats the problem..

Linus reran the test at their demo booth and the i7 6900K scored 1479 vs 1612 for the 1800X. In line with the AMD slides. Tom's Hardware has got the specs of the demo systems as well.

4:33

 

[inf64]

I wonder what happens when we use DDR4 @ 3200Mhz with NB clock adjustments on 1800X.

Specs of both systems (from THG).

 

[french toast]

If anything it would be even better for AMD NOT to use Turbo 3.0 since 1800X would score higher @ stock than 6900K in ST test . 1800X is beating 6900K in MT test no matter what so MT test is not that important as Ryzen already has an advantage in this benchmark.

Agree totally,AMD have been more than fair.
It seems some peeps would never be happy, perhaps they expect i7 6900k bench to use water cooling vs ryzen stock air? Perhaps they would want quad channel for performance benchmarks - but then switch to dual channel when doing perf/watt measurements? Alot of people cried foul over the blender benchmarks, but when invited everyone to test themselves and it was validted we hear no more from the same people?

You cant please everyone it seems, AMD have been very fair and transparent imo.

 

[SAAA]

Whatever Intel's response is it won't be another Conroe. An Athlon 64 or Core 2, or in this case Ryzen happens once in a decade.

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.

At RWT they are talking about Zen's SpecCPU2006 scores.

The Base version of the benchmark(not-so-threaded) without the benchmark-breaking libquantum shows that Skylake has about 10% advantage. Meaning its on par with Broadwell, if not better.

Intel has no true answer for at least few years. Man this is a repeat of 1999's Athlon again. The financial impact to Intel is practically zero for couple of years just like with the Athlon. The real ramifications come in a few years when modification to the core arrives like Athlon 64 did. Then they have chance to take real marketshare(AMD briefly took 50% marketshare in retail).

It is a bit more complex than that of course. In terms of marketshare, they need a good APU with Raven Ridge. You NEED an iGPU to sell into most computers.

Oh, and I highly doubt SKL-X and KBL-X will do more than 5%, if that. What secrets do you think they are hiding? It's a fallacy to believe magical big gains are out there.

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.

 

[IntelUser2000]

I dont think it is even remotely the same as back in 99. From 99 onwards is was a slow ramp up and then down right up until neleham and QPI killed AMD's last advantage.

In the space of 6 months AMD will go from not being competitive in any market to heavy competing in all Intel markets.

The details are different, but the overall scenario is similar. When Athlon launched the belief was also that Intel was unsurmountable. Previous to Zen, the feeling was that AMD had zero chance.

In the laptop space Zen + Vega will be a better SOC then kabylake
In Server 32 core naples will better then broadwell-EP not only as a processor but as a platform. It will also be very competitive to Skylake-EP (outside people who want AVX-512) in performance and as a platform.

That we will see because Laptop and Enterprise I would say isn't really Intel's weakness, but desktop. Traditionally in laptops its where Intel's R&D budget paid off even during its low point because it translated into better power management. One thing is sure, AMD now has a good baseline.

Even with "crappy" excavator AMD has more OEM product then during the golden years thax to the lack of intel bribes.

Absolute numbers wise that may be true, but the PC market is also much greater than that.

That really only leaves intel with one area of advantage and that is the high clocked 4 core.

This is also the market that has most complaints. Because while absolute clock wise its pretty good, compared to predecessors its quite mediocre.

This time its going to actually hurt intel because of market conditions and i think far quicker then most people are predicting. How intel will react will be interesting and i think will be determined by what their next products actually look like, price cuts? "cost cutting"? ignoring until its undeniable ?

I am quite wary of saying such fancy statements, because its few, and rare. Real world is much more progressive rather than a "breakthrough". General population will also take few years to warm up to rising AMD brand, which is why I still stand on taking few years for mass success people are looking for.

 

[JDG1980]

Intel doesn't do price cuts. Although dropping prices on Intel's existing Broadwell-E HEDT lineup would help it compete more effectively with Ryzen, I don't see that happening because it would represent a loss of face. Most likely, Intel would conclude that the hit on their brand equity (by legitimizing Ryzen as a competitor) outweighed than whatever money they would gain from the additional sales.

What's more likely is that Skylake-X pricing will be reevaluated, and better products moved into the same price points.
What would be reasonably competitive pricing for Skylake-X?

• Flagship 16-core at $1723, for those who absolutely must have the top consumer processor no matter what
• 12-core at $999
• 8-core at $599
• 6-core (full PCIe lanes) at $399
• 6-core (castrated PCIe lanes) at $299

I could see Intel doing something like that. By doing it for the new generation rather than cutting prices on the old one, they would be able to pretend that this is just the usual process of generational improvement and has nothing to do with competition from AMD.

AMD's best countermove, in turn, would be to do a consumer 16-core CPU as soon as possible (via two MCM'd Summit Ridge dies), and clock speed bumped versions of the existing core configurations ~6 months from now once the production quirks are all worked out.

 

[IntelUser2000]

The mobile chips are a bit misleading. They are able to stay at higher boost levels for longer periods than previous chips. The gains in mobile are clock speed gains, not ipc gains.

Yes, and I know that.

But in mobile where thermals are important as everything else, unlike in desktops, clock speed gain is pretty much same as IPC gain. If Kabylake was 50% faster per clock but at 1/2 the clock it would be the failure on the level of Prescott-on-laptops.

They did not, and Kabylake on mobile gets that increase with lower thermals and better battery life as well.

On desktops though if Kabylake 7700K was at 5GHz base to get that 15%, it would basically be factory-overclocked chip that can't overclock and essentially same as the actual 7700K.

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.

Kabylake did not. I mean most of us believed it would bring some, even little. Because they claimed they had some "key" performance improvements. Heck, there were some rumors that seemed very credible that Kabylake would bring Cannonlake's uarch changes.

Again, the fault is that people will always conclude rationally and that engineers are always at fault. You can have the best engineers, but managers direct where to go. I am pretty sure engineers at RIM(Blackberry) and Nokia were competent. But company mood and management was at its low. The ARM threat started the derailing at Intel. Starting with the iPhone.

 

[french toast]

Intel doesn't do price cuts. Although dropping prices on Intel's existing Broadwell-E HEDT lineup would help it compete more effectively with Ryzen, I don't see that happening because it would represent a loss of face. Most likely, Intel would conclude that the hit on their brand equity (by legitimizing Ryzen as a competitor) outweighed than whatever money they would gain from the additional sales.

What's more likely is that Skylake-X pricing will be reevaluated, and better products moved into the same price points.
What would be reasonably competitive pricing for Skylake-X?

• Flagship 16-core at $1723, for those who absolutely must have the top consumer processor no matter what
• 12-core at $999
• 8-core at $599
• 6-core (full PCIe lanes) at $399
• 6-core (castrated PCIe lanes) at $299

I could see Intel doing something like that. By doing it for the new generation rather than cutting prices on the old one, they would be able to pretend that this is just the usual process of generational improvement and has nothing to do with competition from AMD.

AMD's best countermove, in turn, would be to do a consumer 16-core CPU as soon as possible (via two MCM'd Summit Ridge dies), and clock speed bumped versions of the existing core configurations ~6 months from now once the production quirks are all worked out.

rumours of Skylake X being another 10 core, other than that i agree.

 

[french toast]

I wonder what happens when we use DDR4 @ 3200Mhz with NB clock adjustments on 1800X.

Specs of both systems (from THG).

Yes , Ryzen supports 3600mhz + at leadt on one mobo

 

[tamz_msc]

What I meant by the Conroe reference that generational gains this big only happen once in a long time. Ryzen was possible because Bulldozer was an inherently bad design and AMD were stuck forever on 32nm. With the minor speed bumps Intel brings each year and process technologies running into limitations with the laws of physics, it seems clear to me that the rate of progress, at least when it comes to desktop CPUs, has slowed down considerably over the past 5 years or so. Both AMD and Intel will from now onward settle to a cadence cycle which is mostly about refinement of the existing architecture.

 

[inf64]

Intel doesn't do price cuts. Although dropping prices on Intel's existing Broadwell-E HEDT lineup would help it compete more effectively with Ryzen, I don't see that happening because it would represent a loss of face. Most likely, Intel would conclude that the hit on their brand equity (by legitimizing Ryzen as a competitor) outweighed than whatever money they would gain from the additional sales.

What's more likely is that Skylake-X pricing will be reevaluated, and better products moved into the same price points.
What would be reasonably competitive pricing for Skylake-X?

• Flagship 16-core at $1723, for those who absolutely must have the top consumer processor no matter what
• 12-core at $999
• 8-core at $599
• 6-core (full PCIe lanes) at $399
• 6-core (castrated PCIe lanes) at $299

I could see Intel doing something like that. By doing it for the new generation rather than cutting prices on the old one, they would be able to pretend that this is just the usual process of generational improvement and has nothing to do with competition from AMD.

AMD's best countermove, in turn, would be to do a consumer 16-core CPU as soon as possible (via two MCM'd Summit Ridge dies), and clock speed bumped versions of the existing core configurations ~6 months from now once the production quirks are all worked out.

Good post . Are you sure intel will bring 12C in HEDT segment, let alone 16C? Do we have any roadmap or leak that confirms that?

 

[IntelUser2000]

What I meant by the Conroe reference that generational gains this big only happen once in a long time. Ryzen was possible because Bulldozer was an inherently bad design and AMD were stuck forever on 32nm. With the minor speed bumps Intel brings each year and process technologies running into limitations with the laws of physics,

It seems clear to me there still can be differences, but in-line with other technological areas rather than fancy schmancy 100x gain in 10 years Moore's Law changes brought. Apple creates a very capable core. It's basically 15W Iris Core i5 chips that can be fit into a very slim and very portable Tablet.

AMD was given an opportunity to catch up with Ryzen thanks to physics limitations. It seems success isn't a given or for the privileged, you just need to give it some time. Same with ARM cores. ARM never "lost" because Intel's meteoric rise in the 90s. You only lose if you die.

It will be very depressing to technology folks that aspire to see that sci-fi like future. But there is a genuine limitation that won't be overcome anytime soon. The problems are so numerous in so many areas.

 

[coercitiv]

Ryzen was possible because Bulldozer was an inherently bad design and AMD were stuck forever on 32nm.

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.

 

[revanchrist]

I wonder what happens when we use DDR4 @ 3200Mhz with NB clock adjustments on 1800X.

Specs of both systems (from THG).

That idle power consumption is crazy good omg.

 

[dark zero]

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.

Actually... They did Conroe BIG time here. Intel didn't expected a leapfrog like that. And to stay not that behind for them and even better than them is worsening the situation of the Blue team...

And they were busy on the modem bussiness that QUALCOMM is leading hard... Oh boy... Intel will have a hard time.

 

[bononos]

The first response would be to pick up their jaws off the ground.

 

[Spartak]

It's pretty obvious to me AMD is keeping a top part behind (1900X) to introduce a bit later. Both for needing some extra time to improve bins / performance and as a money maker. If they introduced that part right away at $600 everybody would be focussed too much on that high price.

The nomenclature is probably such that the 1 stands for the generation and the 9 for where it falls in the lineup. They might even do a 1950X and move everything down a notch in pricing if Intel responds aggressively with pricing.

Maybe this is already clear to everyone here, but it's a thought that just occurred to me lol.

 

[lolfail9001]

It's pretty obvious to me AMD is keeping a top part behind (1900X) to introduce a bit later.

What mythical top part?

The nomenclature is probably such that the 1 stands for the generation and the 9 for where it falls in the lineup.

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

 

[Spartak]

It's clear that they reserved a spot in their line-up to introduce a better top part when their bins and performance have improved with successive steppings. There is nothing 'mythical' about it.

It just occured to me they have a vacant spot in their nomenclature that actually makes a lot of sense form a strategic and pricing POV. I'm not implying they're withholding something right now ('keeping top part behind' was maybe a bit poorly worded), just planning for the near future. The 'sign' is the fact that 9 > 8 and it reflects performance.

 

[french toast]

It's clear that they reserved a spot in their line-up to introduce a better top part when their bins and performance have improved with successive steppings. There is nothing 'mythical' about it.

It just occured to me they have a vacant spot in their nomenclature that actually makes a lot of sense form a strategic and pricing POV. I'm not implying they're withholding something right now, just planning for the near future.

You speak of a threadripper?

 

[CatMerc]

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.

Fottemberg appears to hint at such a part: https://twitter.com/BitsAndChipsEng/status/830204223698636800

 

[Spartak]

You speak of a threadripper?

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. Meta and all.

Maybe I should've made a new thread called AMD's response to Intel's response to Ryzen. And just post the Kermit pic. Missed chance!

 

[Spartak]

Fottemberg appears to hint at such a part: https://twitter.com/BitsAndChipsEng/status/830204223698636800

haha. Exactly. BTW I strongly identify with Kermit (if people would ask me to name a fictional character) so this is me spilling the beans right here in this thread