Intel’s Core i7 H-Series chips have long been the CPUs of choice to power high-end gaming and multimedia-content-creation laptops. The company’s latest 10th Generation versions improve upon this venerable chip design by boosting its maximum clock speed above the 5GHz mark. That’s remarkably high for any CPU, never mind a laptop CPU.
In early testing at PC Labs, some of the new 10th Generation H-Series Core i7s show they are well-positioned to be the new favorite for powerful laptops released this year. But these chips, based on the “Comet Lake” architecture, are no longer shoo-ins. The release of a new breed of powerful fourth-generation AMD chips in its own new H-series, dubbed the Ryzen “Renoir,” could loosen Intel’s stranglehold on the high-performance laptop market.
Gigabyte Aero 15 OLED XB, with a 10th Generation Intel Core H Series CPU
It’s far too early to call a winner, however. We’ve tested only two Intel 10th Generation Core i7 H chips (models from Gigabyte and MSI, more about which below), and one AMD fourth-gen H-series Renoir laptop so far, from Asus. Still, it’s a particularly exciting time if you’re quarantined at home with plenty of time to play, and itching to upgrade your years-old gaming rig. So let’s take a look at kind of performance Intel is advertising, and how those claims stack up against our own preliminary testing.
What’s Your Starting Point?
With Intel enjoying a dominant position in the market for powerful laptop CPUs for so long, the performance improvements between its laptop-chip generations have been relatively minor. The company’s reasoning is that people aren’t looking to upgrade their $2,000-plus gaming laptops every year, but rather every three to five years.
So Intel’s claims largely revolve around how much better the Core i7-10750H (the presumed chip for the price/performance sweet spot of the gaming laptop market) performs compared with its three-year-old cousin, the 7th Generation Core i7-7700HQ. Intel says the new chip should offer 44 percent more frames per second compared with its ancestor when playing graphics-intensive games. It should also be much more proficient at handling CPU-intensive tasks besides gaming: 70 percent faster at exporting 4K video, for instance.
Take the frame-rate claim with a huge grain of salt; that three-year-old sample laptop also has a three-year-old graphics chip, and that makes most of the difference in games.
Overall, though, Intel says upgraders can expect a 33 percent performance boost if they’re coming from a three-year-old system. The higher boost clock plays a part in this gain, with the Core i7-7700HQ reaching a maximum speed of just 3.8GHz, compared with the Core i7-10750H’s 5GHz. But the big-impact improvement for overall performance is adding more processor cores and threads.
The Core i7-7700HQ is a quad-core chip, while the Core i7-10750H has six cores. Both chips support thread-doubling Hyper-Threading, which means the Core i7-10750H can handle 12 instruction threads at once, compared with eight for the Core i7-7700HQ. That’s indeed going to make a huge difference in multi-threaded, CPU-dependent software, which includes many modern apps.
So, How Do Those Claims Stack Up?
Checking these performance claims with third-party tests isn’t straightforward. Other than proprietary tests like the ones that Intel uses, there aren’t a lot of impartial benchmarks that remain constant through the years and can help us evaluate Intel’s claims that the 10th Generation Core H chips are a huge leap over 7th Generation ones from 2017. The best tool we’ve got available is Cinebench R15, an aging benchmark, but a widely accepted and relevant one because it’s based on actual workflows in Maxon’s Cinema 4D rendering tool.
MSI GS66 Stealth, with 10th Generation Intel Core H Series
Cinebench is fully threaded to make use of all available processor cores and threads. It stresses the CPU rather than the GPU to render a complex image. The result is a proprietary score indicating a PC’s suitability for processor-intensive workloads.
The Cinebench scores below are from the MSI GS66 Stealth, which has the Core i7-10750H, and the early 2017 version of the Razer Blade, which has the Core i7-7700HQ. (We’re in the process of reviewing the GS66; the MSI link previous is to a preview of the laptop we did at CES 2020.)
For comparison’s sake, I’ve also included the Editors’ Choice Gigabyte Aero 15, which has the eight-core Core i7-10875H (a step up from the Core i7-10750H), as well as the Asus ROG Zephyrus G14, with the eight-core Ryzen 9 4900HS. Here’s a rundown of the basic CPU and GPU specifications for each of these machines…
And so, onward to our initial Cinebench R15 results. Again, emphasizing: These tests were run at at the benchmark’s All Cores setting, which stresses all cores and threads…
The 61 percent difference in the Cinebench score between the Razer Blade and the MSI GS66 Stealth is impressive, and it suggests that the improvements with highly threaded software like this should be at least as good as Intel says they should be. Since Cinebench is so dependent on the number of CPU cores and threads, it’s no surprise that the six-core Core i7-10750H performs better than the quad-core Core i7-7700HQ.
With software like this, you can expect noticeable performance gains if you’re upgrading a three-year-old laptop. But the chart above also poses a more nuanced question for those with a little extra cash to spend: Should you consider a laptop with the Core i7-10875H or the Ryzen 9 4900HS, instead, to eke out even better performance? To find out, let’s take a closer look at how both of these 10th Generation Core i7 laptops and the Ryzen 9 laptop perform on a range of CPU-intensive tasks.
Digging Deeper: Single- and Multi-Core Performance
A good place to start digging deeper into the performance differences is by isolating the contribution of each CPU core. Adding cores is almost always going to increase performance, but how a core performs on its own tells us a lot about the overall efficiency of the chip design. Here’s the Cinebench test again, but run on just a single core this time…
The Ryzen 9-powered Asus still comes out on top, though the relative margin is much, much smaller. In fact, all of the competitors are roughly equal here. It suggests a newly leveled playing field between AMD and Intel, when it comes to the ability of a single processor core to handle software instructions. Previous AMD CPU generations have consistently shown small but consistent deficiencies in single-core performance compared with their equivalent Intel competitors.
Now, of course, most people aren’t going to be turning off cores to admire the efficiency of the CPU in their laptop. Applications that can hit just a single core tend to be old, legacy ones, not rewritten for modern silicon realities. For a more real-world look at performance, let’s do some 4K video rendering using the open-source app Handbrake. This is a tough workout that’s highly CPU-dependent and, like Cinebench, scales well with cores and threads. We put a stopwatch on test systems as they transcode a standard 12-minute clip of 4K video (a demo movie called Tears of Steel) to a 1080p MP4 file. It’s a timed test, and lower results are better.
The Ryzen 9 laptop still comes out on top. Having eight cores and 16 threads, versus six and 12, is a major strength in programs like this.
What about image editing? Using an early 2018 release of the Creative Cloud version of Photoshop, we apply a series of 10 complex filters and effects to a standard JPEG test image. We time each operation and, at the end, add up the total execution time. As with Handbrake, lower times are better here. The Photoshop test stresses CPU, storage subsystem, and RAM, but it can also take advantage of most GPUs to speed up the process of applying filters, so systems with powerful graphics chips or cards may see a boost.
That’s not necessarily what happened here, however. The image-editing results demonstrate one of the key disadvantages of evaluating CPU performance by itself, as a pure drag race: A lot of complex software relies on other system resources (memory or graphics acceleration, especially) in addition to the CPU to accelerate performance. Yes, the eight-core Core i7-10875H in the Gigabyte finished in the least amount of time, but the MSI actually has the most powerful GPU, a GeForce RTX 2080 Super Max-Q. Some other factors are clearly at play. Photoshop isn’t optimized to take advantage of every single GPU on the market, and even the thickness of the laptop and how much air circulation there is can make a difference in how effectively it is able to perform.
Other common computing tasks are much less complex, like compressing and decompressing files into a Zip folder. That’s what our 7-Zip test measures, and like Cinebench, it’s mostly dependent on the number of CPU cores and threads.
The Ryzen 9-powered Asus returns to the lead, this time by a healthy margin of more than 20 percent over the Core i7-10875H in the Gigabyte.
Finally, many buyers of powerful laptops will use them to perform highly specialized workflows like compiling code or even ray tracing, a task that we approximate with the Persistence of Vision (POV-Ray) app. This free software makes tons of floating-point calculations to determine pixel colors and put them on screen. Similar to Cinebench and 7-Zip, it exercises CPU cores and threads to the maximum. POV-Ray gives the processor a workout during the off-screen rendering of a complex, photo-realistic scene with multiple light sources. The time this takes is listed in seconds (lower is better), and as with Cinebench, we can see different scores for one-core operation, as well as all the cores working together.
While the Ryzen 9-powered Asus is the fastest overall with all cores enabled, it’s interesting to see that the Core i7-10875H in the Gigabyte actually does better than the Ryzen 9 while the test is running on just a single core. This is different than what the single-core Cinebench test showed, and it’s possible that the 5GHz maximum boost clock for the Core i7 came into play here, versus the 4.3GHz maximum boost clock for the Ryzen 9. Whatever the case, the differences between this test and the Cinebench one suggest that the Intel-AMD playing field isn’t level by every measurement. They’re also a good reminder of the limitations of CPU benchmark testing—the capabilities of the software you run often matter just as much as the hardware you run it on.
Read the Fine Print
There a few more limitations to all of these test results. Most important, they only reflect resource-intensive workflows other than games. Unlike multimedia content creation, for which there are a few standard apps that many people use, there are a ton of different games, all of which perform differently and tax the CPU and GPU in different ways. This makes game testing and isolating the CPU’s contribution very, very challenging, and beyond the scope of a simple apples-to-apples comparison between CPUs. Even among the workflows mentioned above, there are several that are influenced by more than just the CPU (especially the image editing test).
If you’d like to see how the laptops perform while playing demanding games like Rise of the Tomb Raider and Far Cry 5, or in some industry-standard graphics benchmarks, head on over to our full reviews.
Asus ROG Zephyrus G14, with AMD’s Ryzen 9 4900S
Second, the comparison between the higher-end Core i7 in the Gigabyte and the Ryzen 9 in the Asus isn’t exactly fair. A more equal showdown would use the Core i9-10980HK, Intel’s eight-core H-series flagship, but we haven’t yet got our hands on a laptop that includes one. Based on specs alone, however, the difference likely won’t be significant. Both the Core i7-10875H and the Core i9-10980HK have the same number of cores and threads, and the only noteworthy differences between them are the Core i9’s additional (but very minor) 100MHz of base clock speed and 200MHz of boost clock speed.
On the other hand, the Core i7-10875H actually performed the best on the single-core POV-Ray test, and if the Core i9 performs even better, it suggests that the Core i9 might consistently offer more computing power on a per-core basis than the Ryzen 9 does, at least while handling certain intensive workflows. Also, the Ryzen 9 4900HS in our Asus test laptop is actually a throttled-down (35-watt) version of the true Ryzen laptop-chip flagship, the even-more-powerful Ryzen 9 4900H, which we haven’t yet tested.
Finally, even if benchmark tests did consistently reflect the real-world experience of using a PC (sometimes they don’t), you should never buy a laptop based solely on raw benchmark performance alone. Check out the keyboard comfort, feel the build quality, audit the connectivity options, and examine the display, just as if you were kicking the tires on a new car.
Should You Take the Plunge?
A powerful laptop isn’t for everyone. Even among the target market for laptops with H-class CPUs—hardcore gamers who need extra horsepower to fuel content-creation or streaming side hustles—the benefits of the 10th Generation Core i7 over its immediate 9th Generation predecessor probably aren’t worth it.
But if you’re upgrading from a five-year-old or even a three-year-old system, it’s a different story. The question, then, is not whether you should upgrade, but whether you should choose a laptop based on an Intel Core H or an AMD Ryzen H CPU. It’s a novel dilemma, since until this year, AMD CPUs haven’t really been an option for serious gaming and content-creation laptops.
All signs appear to point to the Ryzen 9’s supremacy, but we’re reserving our final judgment until we’ve seen more laptops to constitute a bigger sample size for each CPU, as well as what the flagship Ryzen 9 4900H and Core i9-10980HK can do. But for the moment, based on these three laptops, the early returns for AMD look rosy, considering that the chip underdog is a rookie player in this league.