No one should be surprised that the Ryzen Mobile 4000—which brings AMD’s Zen 2 7nm architecture to the laptop world—outperforms Intel’s laptop CPU offerings on multi-threaded performance or graphics performance. Even single-threaded performance—which has finally achieved par—isn’t the big stumbling block we’ve been waiting to see if AMD could overcome.
The egregious black mark against AMD’s mobile line over the last year or two hasn’t been about performance at all—at first glance, it’s about battery life. At a second, closer look, it’s even more about OEM integration. We’re pleased to see that AMD has taken giant strides to improve both issues with this year’s mobile CPU lineup.
Although last year’s Ryzen Mobile 3000 line didn’t use AMD’s newest Zen 2 architecture or 7nm process size, it was still a solid match for its Intel competition on paper. Unfortunately, a good mobile CPU isn’t enough to make a good laptop. Despite raw CPU performance that ranged from “solid competition” to “outright winner” compared to Intel designs, most Ryzen 3000 laptops had poor battery life, poor thermal design, and mediocre system performance at best due to thermal throttling.
The problem is that laptop OEMs weren’t taking the Ryzen Mobile 3000 seriously. They generally lavished attention to detail on Intel-powered designs for higher-end laptops while shoveling Ryzen CPUs into generic designs aimed at cost-is-the-only-object buyers.
Hardware integration isn’t a really crucial factor with desktops or even servers. They can afford to devote extra space, weight, and power to offset minor inefficiencies—but in a laptop, any sacrifice in design and integration quality translates to obvious, user-visible penalties.
The most common complaint with Ryzen 3000 laptops is poor battery life. But performance takes a hit from sub-par integration work as well. Modern laptop designs generally spec a TDP (Thermal Design Power) budget for cooling that’s insufficient to handle the total heat the system’s components can produce when running flat-out for long periods.
Since the system can’t handle full system load for long periods, it throttles performance on the components as necessary to avoiding cooking them in their own waste heat. This is why poor laptop design results in poor performance—a less-efficient system generates more heat, exceeds the TDP budget more quickly, and the system must slow itself down as a result.
This year, it seems clear that AMD has gotten OEMs’ attention and has them on board with producing detailed and fully integrated Ryzen 4000 system designs. AMD had a Lenovo executive take the stage to talk about the upcoming Yoga Slim 7, an ultrathin 14-inch laptop with high performance and long battery life—but Lenovo is unlikely to be the only OEM paying attention.
Although AMD is being cagey for the moment about exactly who is or isn’t on board, the company specifically talked about partnerships with Microsoft and Google as well as new design certification and continuous validation programs aimed at fixing any problems with poor integration going forward.
We went into AMD’s Tech Day—the press-invite-only launch for the Ryzen Mobile 4000 line—excited about the potential performance to be found in the new 7nm design. But we had real concerns about battery life. To overcome user dissatisfaction with earlier AMD powered laptops, the company needed both to improve OEM and OS vendor-integration efforts and improve the direct hardware efficiency in its own designs significantly.
Happily, AMD seems to have done exactly that. Both the general presentations and a specific breakout on power and thermal efficiency made very clear that AMD took this aspect of the Ryzen Mobile 4000 very seriously. The company claims overall System-on-Chip power reduction of 20 percent and a multi-threaded, heavy-load performance-per-watt twice that of earlier designs.
Shrinking die size from 12nm to 7nm alone might account for as much as half of the increase in efficiency, but the rest comes from significant design improvements. In particular, the Renoir design incorporates the newest ACPI power-state model—which includes three separate low-power c-states to shut down portions of the CPU, instead of the earlier design’s one.
Renoir also adds new OS hinting features. Operating systems supporting the new ACPI model—including both Windows and Linux—can inform the CPU what performance levels are expected for upcoming workloads and how best to choose between higher frequencies and lower power draw.
Finally, entry and exit latency from the lower-power C-states is drastically improved, allowing the CPU to spend more time in lower-power idle modes, with less impact on user-perceived task latency.
With all this architecture explanation done, we can look at a battery-life analysis comparing two real-world systems: the Ryzen 4800U-powered Lenovo Yoga Slim 7, and an Ice Lake i7-1065G7-equipped Dell XPS 13 two-in-one.
Although the analysis here is entirely AMD’s and has not been third-party verified, we think it’s worth pointing out how careful the company has been to provide realistic assessments of its earlier Zen 2 parts. We would also like to highlight the extreme, thorough detail into which AMD goes to explain both how and why it measures battery life the way it does.
Although there is a third-party benchmark aimed at evaluating mobile battery life—MobileMark14—AMD doesn’t like it, finding it too lenient. Instead, AMD created a “blended battery life” routine that involves testing battery life for multiple workloads individually, then averaging them together according to a ratio designed to model “a day in the life of a typical end user.”
We like the approach to real-world use cases, and we like the fact that AMD published its results for each individual use case, not just the blended average, even more. This isn’t a sales-room powder-puff chart; we can see the places where Ice Lake has a leg up on Ryzen 4000 as well as the clear wins for Team Red.
AMD even offers a third column, normalizing results for battery capacity difference between the Yoga and the XPS 13—a change that is not to AMD’s favor, since the Lenovo’s battery capacity is 20 percent higher than the Dell’s.
Ultimately, AMD’s Ryzen 4000 looks like it will provide more active use time for typical users, with the Ice Lake series offering longer idle times and somewhat longer runtime for heavy content creation.
After normalizing the Lenovo and Dell designs for battery capacity, the Ryzen had higher runtimes for productivity, Web browsing, and gaming (as measured by PCMark 10 and 3DMark Time Spy). Intel’s i7-1065G7 had higher runtimes for Cinebench R20, Windows Idle, and Windows Connected Standby.
Last but not least, the entire Ryzen 4000 line’s performance levels look great. Although there were lots of comparisons demonstrating generation-on-generation improvement, we think most readers will want to dive straight into the Team-Red-on-Team-Blue comparisons.
We’re not sure there’s much point in dragging this one out—AMD wins the performance crown handily. If we take AMD at its word—which has been trustworthy so far—Ryzen Mobile 4000 continues the same Zen 2 story already told with Ryzen 3000 desktop CPUs, Threadripper 3000 workstation CPUs, and Epyc Rome server CPUs.
It doesn’t much matter whether you’re comparing Ryzen 7 to i7, Ryzen 5 to i5, or Ryzen 3 to i3. It also doesn’t matter whether you’re comparing Ryzen 4000 U-series to Ice Lake or Comet Lake, or comparing Ryzen 4000 H-series to Comet Lake’s i9.
Single-threaded Cinebench R20 is the only head-to-head performance comparison Intel did not lose outright. In that test, the Ryzen 7 4800U wins against the Ice Lake i7-1065G7 but “only” achieves parity with the Comet Lake i7-10710U.
We had some limited time during AMD’s February launch event in Austin to play with Ryzen 4000-series laptops hands-on, and they did not disappoint. Although there has not yet been an opportunity for any kind of third-party formal benchmarking process, we had the chance to play AAA games on ultrathin laptops and watch engineers demonstrate direct-electrical-circuit testing of Ryzen 4000-equipped laptops run through various workloads.
None of the hands-on experience we had with the laptops in the lab—including but by no means limited to the Yoga Slim 7—seemed in any way “off” from the expectations set by the day’s presentations from managers and engineers.
We expect the new designs to be immensely popular—and in particular, they may very well cause a spike in laptop gaming. Most people will find AAA gaming on lightweight, ultrathin Ryzen 4000 designs not merely “possible” but practical, with many titles exceeding 60fps at 1080P with maximum-quality settings, on 15W CPUs with integrated graphics.