Intel's Architecture Day 2018: The Future of Core, Intel GPUs, 10nm, and Hybrid x86
by Dr. Ian Cutress on December 12, 2018 9:00 AM EST- Posted in
- CPUs
- Memory
- Intel
- GPUs
- DRAM
- Architecture
- Microarchitecture
- Xe
Ice Lake 10nm Xeon Scalable On Display
One of the more sedate talks at the event was discussing Intel’s approach in the datacenter. We’ve covered this story in detail, especially at Intel’s Data-Centric Summit only a few months ago. Intel has stated that Cascade Lake and Cooper Lake are the next two products for the enterprise market, both built on 14nm, focusing on enhanced security as well as AI instructions to help with acceleration. We also know that after these two Intel will have Ice Lake Scalable built on 10nm, but that’s about it.
To be honest, we don’t actually know much more than what we did back then. Intel confirmed that Ice Lake will be built using Sunny Cove cores. But Intel also showed off what they said was an Ice Lake Xeon 10nm processor and package, as shown in the image above.
Color me skeptical, but what was held up is likely either not ICL-SP or just silicon that doesn’t work. In order to make those products, Intel would have to have pumped out at least one large (350mm2+?) die that worked and then put it into a package with a heatspreader. Intel finally seems to be happy discussing a few products on 10nm, as shown at this event, but all the 10nm hardware is based on tiny 100mm2 or smaller silicon. Given Intel’s documented problems, I would have loved that CPU that was held up in the air to be Ice Lake-SP. But I’ll need to see something more concrete to believe it at this point; it’s too much of a jump.
Ending Intel’s Architecture Day
As I’m writing this, it is 3am PT and only a couple of hours away from Intel’s listed embargo time. The event finished 10 hours ago (a few of us skipped the end event drinks to get to writing) and despite the short time to write it all up, it was a good event overall. For the first time in a good while, Intel decided to talk shop, and in an honest way with very little hand waving. One could argue that in every discussion point, Intel raised more questions than they answered, but the positive here is that questions are being answered, and Intel is willing to share things like roadmaps into 2021, demonstrations of some exciting new products for 2019/2020, and a taste of how they are progressing in both manufacturing and microarchitecture. Hopefully Intel will feel the same and this can become a yearly cadence. The trio of Keller, Koduri, and Murthy, is a strong team to field to the press, and this event fits that bill.
To end this piece, I’m going to put in the Q&A section from day’s presentations, as well as some of the questions put in my particular round-table. It’s an interesting read, and it helps that Jim is full of memorable quotes.
Facebook
Twitter
RSS
148 Comments
View All Comments
Raqia - Thursday, December 13, 2018 - link
Your point is taken and Keller did say it was in its infancy, but I am interested in whether what we're seeing here will be a competitive product or will remain an interesting science experiment. There are theoretical benefits of stacking high performance dies on low leakage ones like this but also substantial challenges and deficiencies which the current iteration doesn't show that it has overcome. What we might see in benefit in terms of better overall area, lower package level fab rejection rates, and better net power characteristic could be offset by a worse concentration of heat and hence more throttling when both elements are running or more expensive packaging. Perhaps in the end, a monolithic die is a better compromise despite losing out on some metrics for mobile.nico_mach - Wednesday, December 12, 2018 - link
So the GPU is going to be called ... Ten to the Eeth power? Is that right?I reject all these Xes used in unpredictable ways. The iPhones are pronounced exar and excess. This is ecksee, and I still use oh ess ecks on my emm bee eh at home.
Jon Tseng - Wednesday, December 12, 2018 - link
>Intel actually says that the reason why this product came about is because a customer>asked for a product of about this performance but with a 2 mW standby power state.
Huh wonder who the customer for that Core/Atom hybrid is. Seems a bit overpowered for a tablet. A bit underpowered for a MacBook (or for a car). Chromebooks maybe but most are too low volume to demand a custom part (maybe the education market is taking off?). PC OEMs don't normally take such custom parts for their laptops. But the graphics loadout implies some kind of PC-type application?
Any ideas??
HStewart - Wednesday, December 12, 2018 - link
From the diagram, it appears that hybrid cpu - has single Core CPU with 4 small (Atom) CPU's - such technology is done with Samsung Processors - this would mean it still lower power - but still have primary single thread core speed.Most interesting would be how the smaller core are used in scheduling system. Most like means and enhancement in OS for proper usage.
A5 - Wednesday, December 12, 2018 - link
There aren't a ton of companies big enough to make Intel create a new product line just for them.The whole list is probably Apple/HP/Dell. Maybe Microsoft.
The_Assimilator - Wednesday, December 12, 2018 - link
Microsoft Surface, obviously. It's become a very profitable line for MS but the current models are either too battery-hungry (Core CPUs) or too slow (Atom CPUs). Fovoros will give the best of both these worlds while also being x86... priced right, a Fovoros-based Surface will essentially end any argument for iPads in a business environment, especially considering most software remains firmly single-threaded. But it remains to be seen whether (a) Intel can get the power down even further (7W is still double most smartphones) and (b) whether their big.LITTLE implementation is good enough.Raqia - Wednesday, December 12, 2018 - link
Windows on ARM will do just fine now that Visual Studio emits ARM native code. Once Chrome gets ported (and that will be soon https://www.neowin.net/news/both-chromium-and-fire... the platform should address 95% of typical daily use cases and provide substantial compatibility with legacy software / file formats. This is better value than iPads and upcoming dedicated SoCs like the 8cx should offer better performance and battery / heat characteristics than what Intel has planned for next year in the same power envelope.The_Assimilator - Thursday, December 13, 2018 - link
I think you missed the part where Windows on ARM is horribly slow and therefore shitty. As a result, Microsoft has no plans to port anything useful (e.g. Office) to ARM, which means Windows on ARM is stuck being the lowest of the low-end. And that's not a space that Surface is intended to play in; Surface is an iPad competitor, and an iPad competitor can't be slow and shitty. Business devices can't be slow and shitty, and they absolutely need to be able to run Office.I expect that either Windows on ARM will be allowed to wither and die once Fovoros ships, or it will languish in a dead zone whereby only the cheapest of the cheap devices by no-name-brand OEMs (think $100 Lenovo tablets) use ARM chips and hence need it.
So unless Qualcomm's 8cx is a game-changer in terms of performance, Fovoros should be the end of ARM on desktop, and thank fucking God for that.
Spunjji - Thursday, December 13, 2018 - link
Microsoft already have an Office code base on ARM, so I'm not sure what you're talking about there.What would worry me about an Intel BIG.little style design is that if Windows doesn't assign your performance-critical application to the correct (big) core, performance will mostly suck just as hard as if all your cores were Atom.
As such, I'd be cautious on calling a winner just yet.
gamerk2 - Thursday, December 13, 2018 - link
Agreed with this; Microsoft has been let down by Intel not having a good mobile platform. If it were up to them, they wouldn't bother with ARM, but they have to due to battery/power/heat requirements.