Intel’s first 10nm Ice Lake CPUs revealed: here’s your decoder ring

Photo by Sean Hollister / The Verge

After years of delays and false starts, Intel is finally shipping its first batch of 10nm processors to the world, the company announced last week. As we’ve previously discussed, the new Ice Lake CPUs are only for thin-and-light laptops no more powerful than a 13-inch MacBook Pro — but we’re now getting a rough idea how the lineup actually breaks down.

Today, Intel has revealed the first 11 processors based on its Ice Lake architecture, and some differences might not be obvious when you’re buying a laptop from a store — so I thought I’d whip up a little decoder ring to help you understand what a phrase like “Intel Core i5-1035G4” is all about.

Particularly because not all these chips are equal: a Core i7-1060G7, Core i7-1065G7 and Core i7-1068G7 might sound roughly the same, but they’re really not.

Intel’s naming scheme, broken down

You may want to reference this handy visual aid as we go, but don’t worry about digesting it all just yet:

As you can see, your typical processor name will be something like “Core i5-1035G7.”

The big caveat: processing power isn’t as easy to identify anymore

Those are the basics, but I want to spend a little more time talking about digit number four and what that processor wattage (aka TDP) actually means — it stands for “thermal design power,” or the amount of heat a processor is expected to produce on average when it’s being used, and thus how much cooling it needs to run at full speed.

See how the Y-series chips are rated for 9W or 12W, or how the U-series chips are rated for 15W or 25W? That’s a tremendous range within each chip, and it means the exact same chip can run at faster frequencies for longer if manufacturers stick them in a larger laptop with better radiators and fans.

But also the fact that the Y-series TDP has crept up from 5W or 7W to what’s now 9W or 12W, while the U-series’ base clockspeeds have slunk down to around a Y-series-like 1GHz instead of hovering around 1.6GHz, suggest that U and Y are more alike than ever before — suggesting you may no longer be able to rely on a U-series processor to give you fast sustained performance unless it’s got enough thermal headroom to do so. Which you won’t find on a spec sheet, of course.

Here’s what Intel told us about the speed difference between Ice Lake Y and U:

[The] main performance difference between the 12W and 15W will be seen on multithreaded applications, which are more power limited. 3W of extra power will be used to increase average frequency into higher performance. We have seen between 5-15% performance increase on some benchmarks such as Spec06, SYSmark and 3DMARK.

That’s why that “8” in “Core i7-1068G7” is going to be the most important digit to look for later this year: it’s the only chip in the entire lineup that, for better or for worse, guarantees it’ll give you that minimum 2.3GHz in all apps. Just don’t expect the difference to be huge here either: Intel tells The Verge it’s seen a similar 5 to 15 percent performance increase from that 28W part compared to the 15W ones.

Intel says 35 laptops are lined up to deliver Ice Lake chips this holiday season, some of them Project Athena designs that’ll offer over 9 hours of real-world battery life. We’ll be eagerly waiting to see how they perform.

Comments

4 core Y series, yuss

This makes the timing of the recent MBP13/MBA update a little awkward if this update is out any time soon. Maybe Apple is waiting till there’s something appropriate for the 15/16, but I like this new Apple that updates as they come along instead.

Or maybe that’s why the 15-turned-16 is getting the bigger redesign, while the 13 gets a technically more advanced chip (albeit lower power).

If not…That Ice Lake XPS 13 2 in 1 is awful tempting now that it has more power than my 15" rMBP…

Apple will probably wait until next year before updating the 13" MBP seeing as how they updated the higher wattage systems 5 months ago and just came out with the new model. They usually don’t update a system within 9 months of its last update.

That XPS 13 2-in-1 does look tempting. I like the auto aware feature, slimmer bezels (it looks more like the XPS 13 than the first 2-in-1 did), and some of the underlying features from Ice Lake. Hopefully they price it competitively. I wouldn’t want to see it go above $15000 for a Core i5 version with 8GB RAM, 256GB SSD, G7 graphics, auto aware, and (at least) a QHD display with 15 hours of real world battery life (not just a 15-hour rating with 8-9 hours of actual usage).

I doubt Dell will have any trouble pricing it below $15,000, lol. Seriously, though, even the MacBook Pro 13" is priced at $1,499 for those specs.

I wouldn’t be surprised if Apple skipped this generation of chips and waited for Willow Cove. The benchmarks don’t seem that improved. The only thing that matters is Gen11 graphics and Apple got something really close with their special/custom 8th gen chips. The performance on the integrated graphics side just isn’t enough to upgrade. But Willow Cove should have a whole new graphics architecture. That should mean 1080p gaming at a reasonable and consistent frame rate like 30fps… compared to the crap 1080p at 15fps on Gen 11.

What I’m interested in is the H-series parts on 10nm… I think Intel should be releasing those in August/September-ish… and I’d love to see a 16in MBP w/ Navi graphics coupled with those.

Also, I double checked Intel’s roadmap. Tiger Lake (featuring Willow Cove) is set to launch in Q2 of next year. We’re in the middle to end of Q3 right now… if there is a 9 month upgrade cycle… might as well wait and announce laptop upgrades and what not at WWDC or at the October event next year. Especially since Gen12 Xe graphics are supposed to be way better.

Doubling the graphics performance in the same TDP seems entirely like something Apple would be interested in updating for. There have been generations where they gave up CPU upgrades for a GPU one instead. Would particularly help the 13"ers with media loads.

Apple will want the increase in battery life. Also, Apple never used Iris in the Air, so switching to 10th gen would significantly increase graphics perf in that line.

OEMs and Apple be like

That naming scheme is fucked.

It’s a pretty good naming scheme. About the same as it was before, but with a new graphics number. Have you seen a processor naming scheme that’s better somewhere?

This is actually a great explanation of the naming scheme. Wasn’t really expecting an article from The Verge to get so relatively technical.

0.7 GHz Base Frequency….

Uhhhhhhhhhhhhhhhhhhhhhh are they serious about those frequencies? Holy shit. I bet these don’t perform up to 14nm parts.

According to AnandTech it’s a bit of a wash with the lower clocks canceling out the IPC improvement. The real difference on mobile will be battery life.

This is awesome and all but wake me up when Intel’s Super 14nm ++++++ EX alpha comes out.

I’m surprised a Y-series chip is getting Iris Plus graphics… that’s usually reserved for 28W U-series chips… I have a feeling buying a MacBook Air/Pro is going to get complicated again…

They appear to have redefined Iris with Ice Lake. They haven’t mentioned the presence of any eDRAM which used to be required for Iris graphics.

Anyone paying more attention than me that can estimate which config the new Surface Pro will have, and if it will have a USB-C TB3 port instead of mDP?

The packaging of the 12 watt chips is significantly smaller than that of the 15 watt chips. I wouldn’t be surprised if Microsoft goes with the Y series just to save on space, especially now that the performance difference is just 10-15%.

Even today, Microsoft clocks their Surface stuff relatively conservatively.

Also, if the Surface Book is anything to go by, mDP is definitely being replaced by TB.

A surface go refresh with this would be pretty cool.

I’m betting that one’s reserved for ARM. It’s the redesigned surface pro that I’m dying for.

I don’t know how Intel can keep coming out new products with minimum performance gains. Maybe the x86 architecture has nothing left to improve?

I bought an old HP 8300 (8 years old) for $150 and turned it into a Hackintosh and it is faster than my 2 year old MacBook!

So you’re saying a quad-core desktop chip is faster than a laptop chip, in spite of its’ age? Crazy!

Moore’s law is in its twilight, you mean?

View All Comments
Back to top ↑