*** AMD ThreadRipper ***

[Panos]

Sonic.

As it is for Zen's Infinitiy fabric both are important, but the actual Speed, which affects the bandwidth is what's important.
Since the I.F runs at 50% RAM speed, which directly affects the bandwidth between the CCXs.

With high speed RAM Infinity Fabric runs from 22GB/s all the way to 50GB/s, so you can see why it's so important.
I.F can scale to 500GB/s according to AMD, which is why they're also going to use it for APUs, and Navi/GPUs.

Yep.
I agree and that's what we know up to now. However, that for the mainstream dual channel market

We have no clue how the performance, positively, is affected on quad channel (Threadripper) or octachannel ram (EPYC), per socket.....

And a new video about EPYC and IF

 

[N19h7m4r3]

Yep.
I agree and that's what we know up to now. However, that for the mainstream dual channel market

We have no clue how the performance, positively, is affected on quad channel (Threadripper) or octachannel ram (EPYC)..

I've still got my eyes set on Threadripper. I game at 1440P so all these 1080p results reviewers keep putting out for Skylake X don't bother me.

I like my dual systems workstation + gaming.I wonder how XFR will work with with Threadripper.
With Ryzen it affects two cores, and Threadripper shows XFR and 4.0Ghz. Is that 2 cores total still, or 2 cores per Zeppelin die; which would mean 4 cores boost to 4.0Ghz.

If that's the case Threadripper will be fantastic for myself and others. Then we just need to see how well the memory clocks, and how quadchannel affects it.

 

[Rroff]

Also if it was bad design it wouldn't be scaling 100% nor both Intel & Nvidia having announced last week, that they are working to produce similar product design, in 4-5 years.

Intel's system can already be used to tie together multiple dies on a substrate though I doubt we will see it used this generation. The MCM stuff nVidia is talking about is a different direction - they have another interlink that is developing like IF - the recent nVidia news is for a very specialised interconnect purely for making it possible to dissect and lay out a monolithic GPU as separated packages on an interposer making it possible to scale up any area of the GPU at will without being limited by what you can fit in one die. IF currently is more general purpose - designed for communication between monolithic packages of either the same or different type.

 

[EsaT]

Sounds like arrogance goes hand in hand with complacency.

Considering it was Kodak's employees who invented way modern digital camera sensors work it remind me more about this:

General Lefcourt: We took care of the Dilgar, we can take care of the Minbari.
Londo Mollari: Ah, arrogance and stupidity all in the same package. How efficient of you.

 

[humbug]

Intel's system can already be used to tie together multiple dies on a substrate though I doubt we will see it used this generation. The MCM stuff nVidia is talking about is a different direction - they have another interlink that is developing like IF - the recent nVidia news is for a very specialised interconnect purely for making it possible to dissect and lay out a monolithic GPU as separated packages on an interposer making it possible to scale up any area of the GPU at will without being limited by what you can fit in one die. IF currently is more general purpose - designed for communication between monolithic packages of either the same or different type.

I think

Intel's system can already be used to tie together multiple dies on a substrate

They would have done.

Its like saying, "I could have more cores per CPU and at a much reduced cost.... but i don't want to increase my revenue margins"

 

[EsaT]

Intel got caught with their pants down. The market for more cores is there, it's the future both for HEDT and enterprise.

Intel are doing this the wrong way. Huge cores, hard to cool, expensive to make.

Actually with todays manufacturing processes CPU cores aren't that big in physical size...
https://www.techpowerup.com/215333/intel-skylake-die-layout-detailed
But Intel is badly skimping in their count.

Half of freaking line up is dual cores!
https://en.wikipedia.org/wiki/Kaby_Lake#Desktop_processors
When you compare that to Ryzen...
https://en.wikipedia.org/wiki/Ryzen#Summit_Ridge

 

[humbug]

Actually with todays manufacturing processes CPU cores aren't that big in physical size...
https://www.techpowerup.com/215333/intel-skylake-die-layout-detailed
But Intel is badly skimping in their count.

Half of freaking line up is dual cores!
https://en.wikipedia.org/wiki/Kaby_Lake#Desktop_processors
When you compare that to Ryzen...
https://en.wikipedia.org/wiki/Ryzen#Summit_Ridge

Death of the £100+ Dual Core...... lets hope so anyway, Christ... Intel are stuck in 2004.

 

[Rroff]

Actually with todays manufacturing processes CPU cores aren't that big in physical size...
https://www.techpowerup.com/215333/intel-skylake-die-layout-detailed
But Intel is badly skimping in their count.

I think Intel had been banking on not having to get into the core count race before 10nm given their current strategies.

 

[Panos]

Intel's system can already be used to tie together multiple dies on a substrate though I doubt we will see it used this generation. The MCM stuff nVidia is talking about is a different direction - they have another interlink that is developing like IF - the recent nVidia news is for a very specialised interconnect purely for making it possible to dissect and lay out a monolithic GPU as separated packages on an interposer making it possible to scale up any area of the GPU at will without being limited by what you can fit in one die. IF currently is more general purpose - designed for communication between monolithic packages of either the same or different type.

True, however the chip design should have been different from the start, and those things take years to develop. So back in 2012-13 when Intel was making those decision, didn't thought that AMD will come back with something so innovative 4 years down the line.
Ofc AMD did, and now Intel find themselves 4-5 years behind.
Also the biggest margins and profitability is on the server market, and the monolithic, very strict approach of Intel, going to hurt them having to deal with an AMD product that is far cheaper, and far superior in terms of performance, connectivity and security.

 

[Rroff]

and now Intel find themselves 4-5 years behind

I disagree with them being 4-5 years behind or at least needing that to catch up - AMD managed to dump the Zen platform on them too late for them to easily respond this generation but Intel have been doing the R&D behind the scenes for years on this - if it takes them 4-5 years to respond they've been doing something very wrong.

 

[humbug]

I disagree with them being 4-5 years behind or at least needing that to catch up - AMD managed to dump the Zen platform on them too late for them to easily respond this generation but Intel have been doing the R&D behind the scenes for years on this - if it takes them 4-5 years to respond they've been doing something very wrong.

I agree its not going to take that long to catch up, but then AMD are not going to stand still iether, Zen 2 is rumoured to be a 5Ghz processor with a 15% to 40% IPC increase, depending on which source you believe.

Oh, and Intel's Mesh is not at all like Infinity Fabric, they can't just do what AMD did, at least not without a significant performance drop, which is why they didn't, Intel need to put a lot more work into it, that's not going to come over night.

 

[Panos]

I disagree with them being 4-5 years behind or at least needing that to catch up - AMD managed to dump the Zen platform on them too late for them to easily respond this generation but Intel have been doing the R&D behind the scenes for years on this - if it takes them 4-5 years to respond they've been doing something very wrong.

Don't focus on mainstream CPUs and not the whole picture. Look at the server market where the profitability is, and why Intel spent whole Skylake Xeon release event to slander against AMD.
Check the videos I posted above and on the previous page, from the EPYC Tech Day. With pure numbers explains everything.

 

[Panos]

I agree its not going to take that long to catch up, but then AMD are not going to stand still iether, Zen 2 is rumoured to be a 5Ghz processor with a 15% to 40% IPC increase, depending on which source you believe.

Oh, and Intel's Mesh is not at all like Infinity Fabric, they can't just do what AMD did, at least not without a significant performance drop, which is why they didn't, Intel need to put a lot more work into it, that's not going to come over night.

Is Zen2 the 7nm one, which is different than Zen+ coming early next year yes?
Do we know, if DDR5 will be coming also with the Zen2? Because that would improve the performance even more.

 

[Rroff]

Don't focus on mainstream CPUs and not the whole picture. Look at the server market where the profitability is, and why Intel spent whole Skylake Xeon release event to slander against AMD.
Check the videos I posted above and on the previous page, from the EPYC Tech Day. With pure numbers explains everything.

I am looking at the broader picture - quite a few posts on this subject ignore the whitepapers and patents, etc. Intel have issued.

 

[humbug]

Is Zen2 the 7nm one, which is different than Zen+ coming early next year yes?
Do we know, if DDR5 will be coming also with the Zen2? Because that would improve the performance even more.

Yes and no idea about DDR5.

Zen+ is simply a process improved Zen core, that may come as early as early next year, Zen2 probably late 2018/early 19.

 

[Boomstick777]

Very tempted by threadripper, but ultimately it won't reach it's potential in a PC of mine xD

I think am going to stick with Ryzen Mainstream and wack in the top end Zen 2 chip when it drops. The thought of selling my own rig (Hassle involved) has put me off a new build. So Zen 2 it is.

Anybody else think Intel has done a great job of sabotaging itself lately? Meanwhile AMD launched and launching all these new products at the perfect time imho.

PS4, Xbox One / Xbox One X soon, RX 5XX Series (Sold out, minining / gaming) Zen, and soon VEGA. AMD really are quite amazing considering the size and budget in the face of competition. Kudos to AMD.

 

[Dirk Diggler]

Any news on a release date for this?

 

[humbug]

Very tempted by threadripper, but ultimately it won't reach it's potential in a PC of mine xD

I think am going to stick with Ryzen Mainstream and wack in the top end Zen 2 chip when it drops. The thought of selling my own rig (Hassle involved) has put me off a new build. So Zen 2 it is.

Anybody else think Intel has done a great job of sabotaging itself lately? Meanwhile AMD launched and launching all these new products at the perfect time imho.

PS4, Xbox One / Xbox One X soon, RX 5XX Series (Sold out, minining / gaming) Zen, and soon VEGA. AMD really are quite amazing considering the size and budget in the face of competition. Kudos to AMD.

+1, they deserve all the credit they get right now, and largely they are getting credit for their achievements.

They have also shut up all the nay sayers, especially those on investment forums like Seeking Alpha, self proclaimed experts pre Ryzen launch using words like cannot, will not, exaggerated claims... when talking about Frequencies no more than 3Ghz, IPC gains no more than 20% and must be high power levels..........

The truth is what AMD did here was target 40% IPC increase and get 52%, get all cores running 3.7Ghz 'base' on a Samsung 3.0Ghz Process while the power levels are sub 95 Watts, on a 8 core 16 thread chip.

The floating point performance on Threadripper (FP is usually AMD weakest aspect) is way above that of Intel's SkyLake-X, there are some analysts who have said Zen's FP performance (at least on the dual channel mainstream) is bottlenecked by memory performance, that might explain why gaming performance scales limier with memory performance on Zen and why the massive FP performance on the quad channel Threadripper.

Intel have been stagnant but what AMD did here is no less astonishing. 'Again'

 

[Panos]

+1, they deserve all the credit they get right now, and largely they are getting credit for their achievements.

They have also shut up all the nay sayers, especially those on investment forums like Seeking Alpha, self proclaimed experts pre Ryzen launch using words like cannot, will not, exaggerated claims... when talking about Frequencies no more than 3Ghz, IPC gains no more than 20% and must be high power levels..........

The truth is what AMD did here was target 40% IPC increase and get 52%, get all cores running 3.7Ghz 'base' on a Samsung 3.0Ghz Process while the power levels are sub 95 Watts, on a 8 core 16 thread chip.

The floating point performance on Threadripper (FP is usually AMD weakest aspect) is way above that of Intel's SkyLake-X, there are some analysts who have said Zen's FP performance (at least on the dual channel mainstream) is bottlenecked by memory performance, that might explain why gaming performance scales limier with memory performance on Zen and why the massive FP performance on the quad channel Threadripper.

Intel have been stagnant but what AMD did here is no less astonishing. 'Again'

Amen

 

[Rroff]

Anybody else think Intel has done a great job of sabotaging itself lately?

Wouldn't say sabotaged themselves - complacency and I think partly complacency that people would still buy Intel as a brand especially in established markets like servers, etc. even if AMD did come out with something competitive. AMD caught them on the wrong foot as well so they can't easily respond - despite some people having trouble understanding the concept - I suspect its more happy accident for AMD that they caught them at the point they did as you can't really strategise for something like that to that degree in the CPU market as there are limits how early you can bring a product to market and development time can be put back by things like having to respin, etc. which you can't really control and it makes no sense to delay a product in 99% of cases as it just gives the competition more time to develop their [next] product.