Intel's Future Chips: News, Rumours & Reviews

[-Fran-]

So this begs the question: what's more important in the current Server world? Per-core performance or moar coars?

Neither; Performance Per Watt dominates. Server loads do tend to be more sensitive to more cores however, but let's not pretend Xeon's don't exist or anything.

The implied "writing on the wall" for that was "at a given TDP".

Cheers!

 

[goldstone77]

Ice Lake 2020, and Cooper Lake 2019 share the platform.
https://twitter.com/Ashraf__Eassa/status/1027236588911124480

 

[goldstone77]

More confirmation about Coffee Lake refresh compatibility with 300 series motherboards.

 

[goldstone77]

Here are the first Intel Core i5-8265U (Whiskey Lake) benchmarks – they don’t tell us much, though
LaptopMedia.comNewsHere are the first Intel Core i5-8265U (Whiskey Lake) benchmarks - they don't tell us much, though
From: LaptopMedia Team 9 August 2018 / 05:32

The CPU we have is Intel Core i5-8265U and it’s pretty interesting how it fares against its Kaby Lake Refresh predecessors i5-8250U and i7-8550U.

The new Core i5-8265U is almost identical to i5-8250U – the sole difference is in the Turbo Boost frequency. While i5-8250U can reach 3.40 GHz when boosted, the new i5-8265U reaches 3.90 GHz. Both CPUs have a base frequency of 1.6 GHz.

As you can see, 8250U has an advantage in 2 of 3 tests which means that the new CPUs are not ready for the market and have to go through more optimizations. However, the performance in Adobe Photoshop is pretty impressive and if we can make our conclusion based on it, we can expect around 10% more performance from i5-8265U (Whiskey Lake), compared to 8250U (Kaby Lake Refresh).

https://laptopmedia.com/news/here-are-the-first-intel-core-i5-8265u-whiskey-lake-benchmarks-they-dont-tell-us-much-though/

 

[goldstone77]

https://videocardz.com/77188/intel-core-i9-9900k-will-boost-up-to-4-7-ghz-with-all-eight-cores

The 8 core's are going to be monsters!

 

[goldstone77]

Before I get too carried away about the new turbos on the 8 cores, some word from Ian:

 

[aldaia]

ICE Lake Xeon-SP will have up to 38 cores.
https://www.heise.de/newsticker/meldung/Xeon-Phi-ist-tot-es-lebe-der-Xeon-H-3891026.html?

 

[goldstone77]

ICE Lake Xeon-SP will have up to 38 cores.
https://www.heise.de/newsticker/meldung/Xeon-Phi-ist-tot-es-lebe-der-Xeon-H-3891026.html?

Nice!

2019/20 will then follow the Ice Lake Scalable Xeons (ISX-SP) in 10 nm +, with up to 38 cores, eight memory channels and up to 32 GB of High Bandwidth Memory (HBM2) on board.

 

[logainofhades]

ICE Lake Xeon-SP will have up to 38 cores.
https://www.heise.de/newsticker/meldung/Xeon-Phi-ist-tot-es-lebe-der-Xeon-H-3891026.html?

And will probably be absurdly priced. All these extra cores/threads are practically useless, in most scenarios, as software hasn't caught up to take advantage of them.

 

[aldaia]

ICE Lake Xeon-SP will have up to 38 cores.
https://www.heise.de/newsticker/meldung/Xeon-Phi-ist-tot-es-lebe-der-Xeon-H-3891026.html?

And will probably be absurdly priced. All these extra cores/threads are practically useless, in most scenarios, as software hasn't caught up to take advantage of them.

Server applications and desktop applications are very different. Xeon-SP is a server processor, most data-center applications are heavily threaded and can use as many cores as you can throw in.
Intel problem is the opposite, too few cores and too late. AMD EPYC 2 will arrive early 2019 and is rumored to have 64 cores/128 threads (although some rumors point to 48 cores and other rumors point to more than 2 threads per core)

 

[jimmysmitty]

ICE Lake Xeon-SP will have up to 38 cores.
https://www.heise.de/newsticker/meldung/Xeon-Phi-ist-tot-es-lebe-der-Xeon-H-3891026.html?

And will probably be absurdly priced. All these extra cores/threads are practically useless, in most scenarios, as software hasn't caught up to take advantage of them.

Server applications and desktop applications are very different. Xeon-SP is a server processor, most data-center applications are heavily threaded and can use as many cores as you can throw in.
Intel problem is the opposite, too few cores and too late. AMD EPYC 2 will arrive early 2019 and is rumored to have 64 cores/128 threads (although some rumors point to 48 cores and other rumors point to more than 2 threads per core)

The Xeon Phi and Epyc are two very different products. Epyc is designed currently for low socket count systems, I think current is up to 2S systems. Intels Xeon is up to 8S systems. AMD will have the advantage at 2S systems but anything more is still all Intels market until AMD expands Epyc to it.

As for Xeon Phi, it has very different applications. Notice that it has 32GB of HBM2 memory built in? Epyc will not be running in systems Xeon Phi is designed for.

 

[goldstone77]

This article puts the server market in fairly good perspective with respect to Epyc's potential.

Consider this excerpt (6/17/18) from a Motley Foolpost:

Nomura Securities recently published a research note in which it says Intel (NASDAQ:INTC) CEO Brian Krzanich “was very matter-of-fact in saying that Intel would lose server share to AMD (NASDAQ:AMD) in the second half of 2018.”

“This wasn’t new news,” the analysts said, “but we thought it was interesting that Krzanich did not draw a firm line in the sand as it relates to AMD’s potential gains in servers.”

Apparently, according to the analysts, Krzanich merely “indicated that it was Intel’s job not to let AMD capture 15-20% market share.”

Just wow (Krzanich resigned today over different matters).

In April, AMD reported quarterly revenue totaling $1.65 billion, up 40 percent from the same quarter last year. The big gains prompted observers to suggest AMD has grabbed market share from Intel, which also reported strong quarterly results as both chip makers benefit from heightened big data processing demand in datacenters (see HPCwire article, AMD Continues to Nip at Intel’s Heels). That’s not all EPYC. AMD’s Radeon GPU line is also faring well and the chip market is strong overall. Intel reported $16.1 billion for the quarter citing heavy datacenter demand for its Xeon Scalable processors.

It’s important to restate that AMD’s market share is tiny compared to Intel. That said, given the size of the market, a small market share at this time is probably one of AMD’s greatest strengths – it has enormous head room for awhile. If AMD delivers as promised, Intel will be hard pressed to prevent AMD from taking significant market share. How much remains to be seen.

“We see the data center as a $21 billion-plus opportunity with a combination of CPUs and GPUs. Talking only the CPU side, and really overall x86 is the dominant architecture, we are only one of two vendors that have access to the x86 technology that’s most relevant to the server market today. We estimate just the silicon portion of that market to be well over $15 billion. Today we have roughly one percent market share, a little more, but a vast opportunity ahead of us.” Said Narrod.

Whether this is a one-time victory lap or a sign of persistent renewed strength for AMD in the server CPU business remains to be seen. Stay tuned…

https://www.hpcwire.com/2018/06/21/amds-epyc-road-to-redemption-in-six-slides/

AMD's quarterly revenue was totaling $1.65 billion, which is ~10% of what Intel made for a quarter at $16.1 billion. That said there is huge room to grow, $21 billion annully, with just the section of the data center market AMD is planning on attacking with a mix of CPUs and GPUs. AMD has roughly 1% of the market, so it doesn't take a genius, given BK's 15-20% statement, that AMD is going to make a significant bite into that market gaining billions in revenue. IMO it will be a slow process, but the end result is going to be billions of dollars in revenue, with Lisa Su estimating mid single digit at the end of the year. If 7nm Rome is a 64 core monster, that rumors suggest, AMD will be in position to make billions of dollars next year at Intel's expense.

 

[aldaia]

ICE Lake Xeon-SP will have up to 38 cores.
https://www.heise.de/newsticker/meldung/Xeon-Phi-ist-tot-es-lebe-der-Xeon-H-3891026.html?

And will probably be absurdly priced. All these extra cores/threads are practically useless, in most scenarios, as software hasn't caught up to take advantage of them.

Server applications and desktop applications are very different. Xeon-SP is a server processor, most data-center applications are heavily threaded and can use as many cores as you can throw in.
Intel problem is the opposite, too few cores and too late. AMD EPYC 2 will arrive early 2019 and is rumored to have 64 cores/128 threads (although some rumors point to 48 cores and other rumors point to more than 2 threads per core)

The Xeon Phi and Epyc are two very different products. Epyc is designed currently for low socket count systems, I think current is up to 2S systems. Intels Xeon is up to 8S systems. AMD will have the advantage at 2S systems but anything more is still all Intels market until AMD expands Epyc to it.

As for Xeon Phi, it has very different applications. Notice that it has 32GB of HBM2 memory built in? Epyc will not be running in systems Xeon Phi is designed for.

4S and 8S represent only 9% of the server market, and with increased core counts the share keeps diminishing. The bulk of the market is 1S and 2S

As for Xeon Phi, well it's being discontinued, mostly because its targeting a market where GPUs do a better job.

 

[jimmysmitty]

ICE Lake Xeon-SP will have up to 38 cores.
https://www.heise.de/newsticker/meldung/Xeon-Phi-ist-tot-es-lebe-der-Xeon-H-3891026.html?

And will probably be absurdly priced. All these extra cores/threads are practically useless, in most scenarios, as software hasn't caught up to take advantage of them.

Server applications and desktop applications are very different. Xeon-SP is a server processor, most data-center applications are heavily threaded and can use as many cores as you can throw in.
Intel problem is the opposite, too few cores and too late. AMD EPYC 2 will arrive early 2019 and is rumored to have 64 cores/128 threads (although some rumors point to 48 cores and other rumors point to more than 2 threads per core)

The Xeon Phi and Epyc are two very different products. Epyc is designed currently for low socket count systems, I think current is up to 2S systems. Intels Xeon is up to 8S systems. AMD will have the advantage at 2S systems but anything more is still all Intels market until AMD expands Epyc to it.

As for Xeon Phi, it has very different applications. Notice that it has 32GB of HBM2 memory built in? Epyc will not be running in systems Xeon Phi is designed for.

4S and 8S represent only 9% of the server market, and with increased core counts the share keeps diminishing. The bulk of the market is 1S and 2S

As for Xeon Phi, well it's being discontinued, mostly because its targeting a market where GPUs do a better job.

% of the market vs margins is very different. That 9% could be a much larger margin.

That said it doesn't change what I said. Xeon Phi is not a product that is comparable to Epyc. As said there is HBM 2 and beyond that support for Optane DIMMS. Its in a different market all together.

As for its end, that is the current news but there has been news as well of an Ice Lake-H which would replace this one. Without the full internal documents of Intel, AMD or any company only time will tell what will happen.

 

[goldstone77]

An Interview with Lisa Spelman, VP of Intel’s DCG: Discussing Cooper Lake and Smeltdown
by Ian Cutress on August 15, 2018 8:00 AM EST
https://www.anandtech.com/show/13204/an-interview-with-lisa-spelman-vp-of-intels-dcg-discussing-cooper-lake-and-smeltdown

One take away I picked up on, while reading the article.

So when we say Cooper will start shipping in 2019 and then Ice Lake as a ‘fast follow-on’, rather than think of Ice Lake as a 12-18 month cadence you should think of it as much shorter than that. We’re talking the middle of 2020.

Note: Lisa Spelman, VP of Intel’s DCG, saying mid 2020 for 10nm+
Edit: I take that back we she doesn't clarify later on if it's 10nm or 10nm+. I just assumed to much.

IC: Ice Lake on the chart was shown for 2020, still two years away, but on 10nm. Can you say if that is 10nm or 10+?

LS: I think that’s just getting into a mess of naming things that doesn’t serve anyone’s interests.

 

[YoAndy]

This is not Intel related news but it will change intel's future thinking and business.. NO MORE 7NM for AMD

GlobalFoundries Stops All 7nm Development: Opts To Focus on Specialized Processes
https://www.anandtech.com/show/13277/globalfoundries-stops-all-7nm-development

 

[gamerk316]

This is not Intel related news but it will change intel's future thinking and business.. NO MORE 7NM for AMD

GlobalFoundries Stops All 7nm Development: Opts To Focus on Specialized Processes
https://www.anandtech.com/show/13277/globalfoundries-stops-all-7nm-development

Not really; we've known AMD is using TSMC, and this pretty much shows us why.

 

[jimmysmitty]

This is not Intel related news but it will change intel's future thinking and business.. NO MORE 7NM for AMD

GlobalFoundries Stops All 7nm Development: Opts To Focus on Specialized Processes
https://www.anandtech.com/show/13277/globalfoundries-stops-all-7nm-development

Not really; we've known AMD is using TSMC, and this pretty much shows us why.

My assumption would be that GF is finding it much more difficult to move beyond 14/12nm than they though.

I always say it. If the leader in process technology who spends billions a year in R&D (more than some companies make in gross revenue yearly) is having issues moving past 14nm there are few who wont.

 

[gamerk316]

This is not Intel related news but it will change intel's future thinking and business.. NO MORE 7NM for AMD

GlobalFoundries Stops All 7nm Development: Opts To Focus on Specialized Processes
https://www.anandtech.com/show/13277/globalfoundries-stops-all-7nm-development

Not really; we've known AMD is using TSMC, and this pretty much shows us why.

My assumption would be that GF is finding it much more difficult to move beyond 14/12nm than they though.

I always say it. If the leader in process technology who spends billions a year in R&D (more than some companies make in gross revenue yearly) is having issues moving past 14nm there are few who wont.

Remember GF needed to license Samsungs 14nm tech; they simply don't make enough money to be able to invest in cutting edge process nodes. They're probably better off being the foundry of choice for everything that doesn't need cutting edge performance, given their size.

 

[volkgren]

What's the latest news and rumors regarding Intel's i7-9700K? 8C/8T but when should we expect it? Soldered IHS is what I hear also? Is it called Ice Lake or something else? I think I want one.

 

[goldstone77]

Geekbench 4
Intel 0000 3.10/4.50GHz 8C/16T
GenuineIntel Family 6 Model 158 Stepping 12
https://browser.geekbench.com/v4/cpu/9718580
ASRock Z390M-ITX/ac

 

[goldstone77]

This is not Intel related news but it will change intel's future thinking and business.. NO MORE 7NM for AMD

GlobalFoundries Stops All 7nm Development: Opts To Focus on Specialized Processes
https://www.anandtech.com/show/13277/globalfoundries-stops-all-7nm-development

Not really; we've known AMD is using TSMC, and this pretty much shows us why.

My assumption would be that GF is finding it much more difficult to move beyond 14/12nm than they though.

I always say it. If the leader in process technology who spends billions a year in R&D (more than some companies make in gross revenue yearly) is having issues moving past 14nm there are few who wont.

It could easily be that it was more difficult, they failed at virtually every node. 14nm LPP was a direct license from Samsung. They are also, money constrained. Even if the were able to build 7nm it would cost them ~4 billion dollars, which they didn't have enough customer engagement to compensate the losses. That was GloFo explanation.

Jim has a video on GlobalFoundries
https://www.youtube.com/watch?v=t-nHjBLh7Ro

 

[-Fran-]

I can't find a proper generic run of the 2700X, but at least, from what I can see and assuming that is an ES-base result, those look pretty darn good.

Cheers!

 

[aldaia]

Intel can’t supply 14nm Xeons, HPE directly recommends AMD Epyc
https://semiaccurate.com/2018/09/07/intel-cant-supply-14nm-xeons-hpe-directly-recommends-amd-epyc/

 

[adamsleath]

Ran across this video today, great video if you have the free time 41 minutes long, computer buffs might recognize the name John Hennessy 😉
[video="https://www.youtube.com/watch?v=Azt8Nc-mtKM&feature=youtu.be"][/video]
The future of computing: a conversation with John Hennessy (Google I/O '18)

interesting watch. re Dennard scaling...