Intel has announced yet another tranche of 10th Generation mobile chips, this time based on 14nm. This is the third Intel 10th Generation announcement that the company has made recently and the first to show us how 10nm and 14nm products will live side-by-side in the same product families. The headline news here is that Intel is bumping its maximum mobile CPU core count to 6C/12T in a 15W power envelope, up from 4C/8T. The 14nm CPUs in the 10th Generation family are Comet Lake, paired up with Ice Lake to fill out the field.
On paper, this shift should be an excellent move for Intel. When the company launched 8th Generation chips, it delivered a significant overall performance improvement. Our initial concerns that high-clocked dual-cores might prove to be better options than lower-clocked quads were groundless; the low base clocks on 8th Generation mobile parts didn’t prevent them from delivering excellent gains in comparison.
There’s good reason to think that’s not the case any longer. Here’s one of the official Intel slides predicting the performance improvements customers who buy a new, 10th Generation CPU like the Core i7-10710U (that’s the six-core variant) can expect:
These are significant gains for a single generation of product. Up to 16 percent better overall performance compared with Coffee Lake, 41 percent better productivity in Office 365, and the same battery life? Not bad. But let’s check the fine print.
Click to enlarge.
This is from Intel’s official disclaimers page. Each numbered entry — 1, 2, 3, — deals with one of the claims we’ve just shown you. I’ve highlighted the listed TDP for each CPU in each entry. Note that #1 and #2 — the two performance claims — deal with two very different system configurations. In both cases, the six-core Core i7-10710U has been configured to run at a 25W TDP, while the Core i7-8565U has been handicapped to a 15W TDP.
The third data point, however, does not show this configuration. Here, the two chips are both running in a 15W envelope. The problem here is that users typically don’t have access to an OEM or Intel-provided method of switching between operating modes. That’s a decision that the laptop manufacturer makes. You can sometimes use third-party utilities or the Intel Extreme Tuning utility to tweak CPU configurations, but you can’t just flip between 15W and 25W configurations. Whatever configuration your laptop manufacturer used is the configuration you are stuck with, and they don’t typically advertise this information.
Intel Didn’t Do This for the 8th Gen Launch
We compared backward against the 2017 8th Generation launch to see how Intel had handled messaging in that situation. There’s a similar slide for the 8th Gen family comparing backward against the 7th Gen family.
We see a similar (though significantly larger) improvement and a similar footnote. Where’s that take us?
Nowhere good. In 2017, when Intel compared performance between the Core i7-8550U and the Core i7-7500U, it didn’t need to futz with TDP values in order to make its performance figures align. The comparison was performed with 15W allocated for both CPUs.
There’s only one reason we can think of for Intel to do this: power consumption. While TDP ratings are not equivalent to total CPU power consumption and should not be read that way, giving a CPU more TDP headroom allows it to draw more power. When reviewers spent time with Ice Lake earlier this month, we specifically noted how giving a CPU more TDP headroom allows it to run faster, as shown below:
We don’t know how much faster the Core i7-10710U is when running in a 25W TDP versus a 15W TDP. What matters is that Intel is misrepresenting the type of comparison it’s making on its 10th Generation launch slides. Comparing laptop performance in two different TDP ranges for your performance metrics, only to flip and compare what amounts to a fundamentally different machine configuration for battery life is disingenuous. The switch between 15W and 25W operating modes may not seem like a big deal, but that’s not a switch that an end-user can throw. When you buy one of these chips, you’ll be getting either the higher-performance 25W version or the lower-performing 15W flavor, and OEMs don’t typically communicate the ultra-fine points of their power management strategies or SKU selections.
The final reason to suspect that TDP is limiting CPU performance in this case? The gains aren’t large enough. Moving to a six-core CPU from a quad may not be as large an improvement as the jump from 2C/4T to 4C/8T, but it should still be worth 1.5x baseline improvement, and there are plenty of benchmarks that will show this type of gain — if the chip isn’t butting up against thermal limits already.
Meet the (Rest) of the 10th Generation 14nm Family
Intel is launching a full suite of U- and Y-class parts, as shown below:
Outside of the Core i7-10710U, improvements are kind of difficult to come by. The Core i7-105100U is a 1.8GHz base, 4.9GHz single-core Turbo, 4.3GHz all-core boost. Intel didn’t disclose its all-core boost frequencies for chips like the Core i7-8665U, but that CPU is a 1.9GHz base / 4.8GHz boost CPU. The total number of EUs for graphics and the graphics frequency are identical between the two parts. The Core i7-10710U does support LPDDR4X-2933, LPDDR3-2133, or DDR4-2666, while the Core i7-8665U only supports DDR4-2400 or LPDDR3-2133, but these improvements are going to be of limited value to users. Intel CPUs aren’t very RAM bandwidth-bound.
These chips will also carry the other 10th generation improvements Intel is shipping, like faster Wi-Fi and support for Intel’s Dynamic Tuning technology. They’ll collectively target the 7W envelope (Intel’s 10nm 10th Gen parts don’t fit into anything below 9W). They offer up to 4.9GHz of maximum frequency compared with 4.1GHz for 10nm Ice Lake CPUs. According to Intel, the U-series and Y-series are intended for customers that want top-notch CPU performance but care less about graphics on the whole. Outside of the single new 6-core SKU, all of the new chips are quad-core parts as well.
Our read on the situation is this: Intel is struggling to contain a resurgent AMD by doubling down on the one market where AMD has always been weakest: mobile. 10nm had to be in market by holidays 2020 for a host of reasons, but Intel isn’t manufacturing enough of the chips to just commit to a top-to-bottom 10nm refresh in that segment. So now we have a mix of 14nm and 10nm parts to address overall market needs, with the 10nm CPUs offering higher IPC and a dramatically improved graphics core, but significantly lower frequency. 14nm chips will theoretically anchor the product in-market with a “halo” six-core part.
But this time around, the situation is different. When Intel moved from 2C/4T to 4C/8T CPUs in mobile, it had held the line on 2C/4T configurations for multiple product cycles. Effectively, it had thermal headroom to spare. This time around, the company has telegraphed that its six-core 15W CPU is gasping for metaphorical air. We don’t know what the real improvements are between the Core i7-8565U and the Core i7-10710U, but we can bet they’re smaller than the 16 percent and 41 percent that Intel quoted. And if by some chance you do get a 25W laptop with a Core i7-10710U in it, it’s not going to offer commensurate battery life to that same configuration with a 15W CPU unless the OEM outfits it with a significantly heftier battery — which means you might get more cores and equivalent battery life, but you’ll pay for it with additional weight.