Intel releases new versions of its Core laptop processors every year, and most years the company tends to talk up how much faster they are than the previous generation. But Wednesday’s launch of its newest silicon, the 11th generation “Tiger Lake” chips, is different. Intel’s main claim with Tiger Lake is not that it’s faster than the 10th generation “Ice Lake” family (though it is), but that it’s faster than the latest Ryzen laptop processors from AMD.
Why Intel’s shift, deigning to mention competing silicon from its archrival? The obvious answer is that the latest Ryzen 4000-series “Renoir” chips have been trouncing Ice Lake in raw performance on classic benchmarks ever since the first Renoir laptops went on sale earlier this year. So regaining the lead is important, though we’ll have to take Intel’s word for it since no Tiger Lake laptops are available for independent tests just yet.
But the more interesting answer is that the persistent delays in the research, development, and manufacturing of Intel silicon that helped AMD get a leg up in the first place have led Intel to explore other ways of making their processors faster that don’t rely on pure silicon-muscle improvements. Perhaps the most significant of these methods is adding artificial intelligence (AI) capabilities, which are responsible for much of what backs up Intel’s claim that Tiger Lake is once again the best-performing processor family for thin-and-light laptops.
Speeding Up Those Crops and Blurs
Enabling AI capabilities on a PC or Mac is a team effort involving the chip maker, the computer manufacturer, and the developer of the app that stands to benefit from AI. In Tiger Lake’s case, Intel has added support and optimizations for numerous AI platforms that manufacturers and developers can use to speed up the performance of their apps.
With cryptic names like Open Vino, Gaussian & Neural Accelerator, and Deep Learning Boost, Tiger Lake’s AI suite is difficult to comprehend for non-techies. And much of it actually relies on refinements to technologies already available on Ice Lake and previous generations. But the implications for performance are clear: Tiger Lake’s AI achievements speed up common workflows even as they use the same or fewer computing resources and battery power as previous generations and AMD competitors, Intel says.
The best example of Tiger Lake’s AI advantage is touching up images and videos after you’ve shot them. Take CyberDirector’s PowerDirector video-editing suite, which now offers several AI-assisted Style packs. They can add a bit of French Impressionist flair (or another artistic period of your choosing) to your videos. When we reviewed CyberDirector last year using an older Asus PC with a 6th-generation Core i7, a 16-second clip took 2 minutes to transform. That’s a fair amount of time spent twiddling our thumbs.
Because PowerDirector uses Open Vino, it can harness Tiger Lake’s AI capabilities to drastically speed up that process. A Tiger Lake Core i7 can now apply that AI style effect 11 times faster than a comparable Renoir Ryzen 7 can. A lot of footage that people are editing is a lot longer than our 16-second test clip, which can translate into some truly useful speedups for prosumers and professional editors.
The PowerDirector example also helps explain why Intel is comparing Tiger Lake performance to Renoir instead of to Ice Lake: Tiger Lake is less than 3 times faster than Ice Lake in the CyberDirector example, offering a still-significant, but much smaller, relative advantage to someone who just bought a brand new Intel-powered laptop.
With All That AI, Many Caveats
In fact, there are many more caveats when you take a close look at other AI examples that Intel has published to tout the strengths of Tiger Lake. One of the most popular among pros and casual users alike is Adobe Photoshop’s relatively new Content Aware filters and effects. They use the Adobe Sensei AI suite, which is based on Open Vino. In Intel’s internal testing, the Tiger Lake Core i7 was again faster than the Renoir Ryzen 7, but by only about 10 percent.
Even more significant, the Ice Lake Core i7 was actually 10 percent slower than the Ryzen 7. Why is that important? It suggests that even without any of Intel’s AI features, the Renoir chips are still faster based on their superior silicon architecture alone. (Renoir uses a 7-nanometer architecture, while Intel has said its 7nm architecture won’t be ready for another year, at least.)
Other image and video AI workflows Intel demonstrated at Tiger Lake’s launch on Wednesday offer similarly significant gains over comparable Renoir processors. They include a 4.5-times speedup using Topas Labs’ Gigapixel AI, and a 1.9-times improvement on Nero’s Photo Tagger.
It bears repeating that no independent Tiger Lake performance tests are available yet. Right now, we’re taking Intel’s internal tests at face value. It’s also worth mentioning that while AMD doesn’t have the robust AI ecosystem that Intel has cultivated both in-house and with partners like Adobe and CyberDirector, Renoir processors have plenty of AI capabilities of their own. They mostly involve things like predictive algorithms that boost performance by steering apps and games down the most efficient paths inside the processor.
Still, Intel’s Tiger Lake improvements are clear evidence that making future laptops faster isn’t just about who has the most advanced silicon. AMD might have leapfrogged Intel in terms of processor architecture, but artificially intelligent software and hardware is potentially more important at determining the performance of workflows that prosumer PC users rely upon the most. We’ll be interested to see how that shakes out in the real world once Tiger Lake-based laptops start hitting PC Labs in the coming weeks and months.