The notebook PC is undoubtedly the meat and potatoes for PC makers these days, but the basic processing resources that laptops have offered for the better part of a decade have been just as stagnant as they have been on the desktop. Ever since Intel’s Sandy Bridge microarchitecture made its debut, getting a thin-and-light Intel machine has meant getting a CPU with a TDP in the range of 15W. That chip has had two cores, as many as four threads, and a serviceable integrated graphics processor as the cherry on top.
Sure, power consumption has fallen as process sizes have shrunken, allowing for thinner and lighter laptops and longer battery life, among other benefits. But for folks who want more processing power from their mobile systems, period, it’s been necessary to step up from thin-and-light machines to beefier systems with bigger batteries, thicker chassis, and above all, more pounds on the scale.
|i7-8650U||1.9 GHz||4.2 GHz||4/8||8MB||2||Up to DDR4-2400
|Up to 1150 MHz||15W|
|i7-8550U||1.8 GHz||4.0 GHz|
|i5-8350U||1.7 GHz||3.6 GHz||6MB||Up to 1100 MHz|
|i5-8250U||1.6 GHz||3.4 GHz|
2017 has been the hottest year for CPU performance increases in quite some time, though, so it’s only fitting that Intel would finally shake things up with its first eighth-generation Core mobile processors, code-named Kaby Lake Refresh (or Kaby Lake-R). The first wave of those CPUs, announced back in August, encompassed two Core i5s and two Core i7s. Intel’s latest 15W chips are quad-core parts with Hyper-Threading enabled, for a total of eight threads.
The Kaby Lake Refresh die. Source: Intel
Until now, four cores and eight threads in an Intel mobile CPU has been the exclusive domain of the company’s H-series chips. Those parts generally carry 45W TDPs that require big honking cooling systems and thick chassis to operate at peak performance, and it’s been rare to find one of those chips in a Windows machine outside of gaming laptops with dedicated graphics cards on board. (Apple’s MacBook Pros are one notable exception.)
A block diagram of the Skylake client core. Source: Intel
The basic CPU core for Kaby Lake Refresh is the same Skylake microarchitecture we’ve known since 2015. Intel’s 14-nm process has now gone through two cycles of improvement in that time, but Kaby Lake Refresh doesn’t rely on 14-nm++ to deliver its extra cores. Instead, these chips are still fabricated on the good old 14-nm+ process that underpins Kaby Lake parts.
A typical Kaby Lake Refresh CPU package. Source: Intel
Although its continuing encores are a bit of a let-down for chip nerds, Skylake has aged remarkably well as CPUs go. AMD’s Ryzen CPUs might not trail Skylake by that much on a clock-for-clock basis, but Intel can usually clock its chips much higher than AMD can. Skylake’s pair of wider SIMD units gives it a further edge compared to the Zen architecture in some tasks, too.
Future generations of Ryzen CPUs may require Intel to dig more single-core performance out of its labs somehow, but for now, Skylake is still the best thing going in x86 CPUs. More of those cores in a given CPU is nothing to get disappointed about.
AMD’s recently-announced family of Ryzen Mobile APUs poses another threat to Intel’s performance crown, though: the potential of integrated Vega graphics processors. Intel hasn’t changed the basic graphics architecture on board Skylake CPUs since 2015, either, so Kaby Lake Refresh uses what we might classify as Gen9.5 graphics execution units in a GT2 configuration. Gen9.5 made its debut with Kaby Lake, and it includes improved hardware support for encode and decode of 10-bit 4K HEVC content plus hardware decode support for 4K VP9 cat videos.
Ryzen Mobile APUs, on the other hand, boast much beefier IGPs with eight or 10 Radeon Vega compute units on board. It’s hard to tell how those integrated graphics processors might stack up, since we don’t have full specs for those parts just yet and their shared-memory architecture makes bandwidth calculations difficult. Still, the theoretical performance numbers we can tease out of AMD’s specs for its Vega integrated graphics processors so far put Vega IGPs leagues ahead of an Intel GT2 IGP.
The hope for AMD, then, is that the combo of competitive per-core performance from Zen CPU cores and the potentially class-leading performance of Vega processor graphics is enough to sway buyers to its corner. The company already boasts three design wins with Ryzen Mobile chips, and those systems generally have appealing thin-and-light designs and reasonable spec sheets to go with their APUs.
Intel’s OEMs have a possible answer to Vega processor graphics, though, courtesy of Nvidia’s GeForce MX150 graphics chip. Including this Pascal GPU in a system may not be as elegant as the all-in-one Ryzen APU, but the GeForce’s 384 stream processors and 2GB of dedicated GDDR5 memory seem poised to deliver an experience similar to, if not better, than AMD’s parts. The discrete graphics chip has some advantages, too, namely that it’s not contending with the processor for memory bandwidth in one socket.
For entry-level gaming, the MX150 deserves to be taken seriously. At its 1532 MHz boost clock, the MX150 offers 1.18 TFLOPS of single-precision number-crunching performance, 23.5 Gpixels/sec of pixel fill rate, and 35.3 Gtexels/sec of texture-filtering capability. Those numbers more than make up for the shortcoming of the Intel GT2 IGP, and the MX150 does it in just 30W of board power. The performance-per-watt potency of the Pascal architecture cannot be denied. Of course, AMD’s Ryzen APUs deliver the sum of their performance from a 15W TDP, but one should game on AC power for the best experience to begin with. The extra power draw of the MX150-and-i5-8250U combo isn’t that much of a liability in that light.
The challenge for AMD’s partners, then, is to deliver Ryzen Mobile systems at the right prices. The HP Envy X360 15z Touch, the first Ryzen Mobile system to go on sale, offers a Ryzen 5 2500U APU with Vega 8 processor graphics. That SoC pairs up with a 1TB mechanical hard drive, 8GB of dual-channel DDR4-2400 RAM, and a 15.6″ IPS touch screen with Windows and a convertible hinge. The whole package runs $749.99 at Best Buy right now, although HP’s Black Friday deals have driven that starting price down to as little as $575 lately. Considering the ephemeral nature of those discounts, we’ll use the $750 figure as a solid starting point.
Breathing life into Kaby Lake Refresh
So what does $700 to $800 buy you if you’re shopping Core i5-8250Us? The company sent me three of Acer’s 14″ Swift 3 systems to provide as even a playing field as possible for discerning the advances offered by its eighth-generation Core i5s.
If someone asked me to describe the typical laptop computer in 2017, I’d probably sketch out something like a Swift 3. For about a $650 e-tail price, the base Swift 3 I’m testing offers 8GB of DDR4-2400 memory and a 256GB Intel 600p NVMe SSD to go with its Core i5-8250U CPU. It also boasts a 14″ 1920×1080 IPS display, a backlit keyboard, and a Microsoft Precision Touchpad.
Intel also sent over a GeForce MX150-equipped Swift 3 with slightly different specs. Aside from the upgraded graphics card, that system comes with 8GB of LPDDR3-1866 RAM instead of DDR4-2400. Considering the reduced pressure on the integrated graphics processor to perform, that lower memory spec makes sense to me. LPDDR3 likely lets Acer both include the discrete graphics card and keep prices down all at once. The MX150-equipped Swift 3 runs about $730 on Amazon right now.
The final system Intel sent us to test with is a slightly older Swift 3 with a Core i5-7200U CPU inside. That chip comes paired with 8GB of DDR4-2133 RAM and the same 256GB SSD as its more recent brethren. Since this is an older machine, it’s available at a discount from some retailers. Walmart has this golden wonder for $580 right now, for example.
While this review is primarily concerned with the performance of the CPUs inside the Swift 3s, I should take a moment to talk about the eighth-generation Swift 3 hardware itself. These machines are, in a word, solid. Their non-touch displays are neither the brightest nor contrastiest IPS panels I’ve ever laid eyes on, but their color accuracy seems reasonable out of the box.
The keyboards on the latest Swift 3s offer decent feedback with shallow travel, and there’s no unwanted flex or slop in their action. The Swift 3s’ all-metal chassis didn’t offer a hint of flex, either. The Microsoft Precision Touchpads in these systems are responsive and accurate, and they handled multi-finger scrolling and gestures without a hitch. The only downside to the trackpad on all of these machines is a bit of a hollow sound when they’re tapped, but that’s a minor complaint.
In short, the basic Swift 3 design does everything you’d want from an ultrabook, and it does it all well. For their prices, these systems are thin, light, well-built, responsive, and a general pleasure to work with. They’re pretty nice to look at, too. If this is the state of general-purpose computing devices in 2017, we’re living in good times.
Our testing methods
As always, we did our best to deliver clean benchmark numbers. We ran each benchmark test at least three times and took the median of the results.
Here are the specifications of our test systems:
|Acer Swift 3
|Acer Swift 3
|Acer Swift 3
|Alienware 13 R3|
|CPU||Intel Core i5-8250U||Intel Core i5-7200U||Intel Core i7-7700HQ|
|Memory||8GB (2x4GB) DDR4-2400||8GB (2x4GB) LPDDR3-1866||8GB (2x4GB) DDR4-2133||16GB (2x8GB) DDR4-2666|
|GPU||Intel UHD Graphics 620||Nvidia GeForce MX150||Intel HD Graphics 620||Nvidia GeForce GTX 1060 6GB|
|Graphics memory||N/A||2GB GDDR5, 6 GT/s effective||N/A||6GB GDDR5, 8 GT/s effective|
|Storage||Intel SSD 600p 256GB NVMe SSD||Samsung PM961 512GB NVMe SSD|
|Battery||3220 mAh Li-ion||76 Wh Li-ion|
Our thanks to Intel for providing the three Acer Swift 3 systems for our testing. The Alienware 13 R3 playing host to our Core i7-7700HQ is my personal system and was not provided by a manufacturer for evaluation.
Some additional notes regarding our testing methods:
- All systems were configured to use Windows’ default “Balanced” power plan over the course of our testing.
- Unless otherwise noted, all tests were conducted with display resolutions of 1920×1080 and refresh rates of 60 Hz. Vsync was disabled using the graphics-driver control panel, where possible.
- All drivers and system firmwares were updated to the most recent versions publicly available before testing.
Our testing methods are generally publicly available and reproducible. If you have questions regarding our methods, feel free to leave a comment on this article or join us in our forums.
Memory subsystem performance
Let’s kick off our testing with a quick look at the main memory performance from these systems using the built-in benchmarks from the AIDA64 utility.
Given the architectural similarities between these three chips, it’s no surprise that the differences between them mostly come down to memory speeds (and a higher TDP and boost clock, in the case of the i7-7700HQ). The i5-8250U has slightly higher memory speeds and unsurprisingly outpaces the i5-7200U in these bandwidth-focused benchmarks.
Although we initially encountered a bit of weirdness when testing memory latency on the i5-8250U, further retests put this chip in the lead here by a few nanoseconds. The i5-8250U’s memory latency hasn’t suffered for its extra cores.
Some quick synthetic math tests
AIDA64 also includes some useful micro-benchmarks that we can use to sketch out broad differences among CPUs on our bench. The PhotoWorxx test uses AVX2 instructions on all of these chips. The CPU Hash integer benchmark uses AVX, while the single-precision FPU Julia and double-precision Mandel tests use AVX2 with FMA.
PhotoWorxx doesn’t offer a dramatic start for the i5-8250U and its extra cores versus the i5-7200U. The less-thermally-constrained i7-7700HQ can boost its clocks higher than the 15W chips in this test and holds those speeds under load, so it opens a wide lead.
CPU Hash lets the i5-8250U stretch its legs a bit. Here, the refreshed Kaby quad nearly doubles the i5-7200U’s performance.
In these AVX-intensive floating-point tests, the i5-8250U doesn’t quite double its predecessor’s numbers as we might expect it to given its extra cores and threads. For that kind of increase, you need to step up to the 35W version of the i7-7700HQ in my Alienware 13 R3. Still, these tests show how four cores and eight threads running at relatively low clock speeds can still offer a major performance boost over two Hyper-Threaded cores.
Dota 2 (Fastest)
Let’s kick off our gaming test with one of the most popular games on the planet today. Dota 2 has a reputation for running on everything, even Intel integrated graphics, thanks to a wide and forgiving range of graphics options. More importantly, Dota 2 has a powerful built-in replay system that allows us to precisely replicate the most intense moments in a match from a given player’s point of view, letting us reliably repeat the game’s chaotic battles for benchmarking.
We’ll begin our Dota 2 trek with the Fastest preset at 1920×1080, which cranks down the resolution scaling setting to 52% or so and disables virtually all of the game’s eye candy. It ain’t much to look at with these settings, but budding e-sports pros can at least get in on the fun with Fastest.
At Dota 2‘s least-demanding quality preset, both Intel IGPs deliver playable-enough experiences. Despite reasonable average frame rates, though, the IGPs turn in high 99th-percentile frame times that could be indicative of a rougher experiences than one might expect of 50-FPS or 60-FPS averages alone. To figure out just how bumpy an experience we’re talking, we need to look at our next set of charts.
These “time spent beyond X” graphs are meant to show “badness,” those instances where animation may be less than fluid—or at least less than perfect. The formulas behind these graphs add up the amount of time our graphics card spends beyond certain frame-time thresholds, each with an important implication for gaming smoothness. Recall that our graphics-card tests all consist of one-minute test runs and that 1000 ms equals one second to fully appreciate this data.
The 50-ms threshold is the most notable one, since it corresponds to a 20-FPS average. We figure if you’re not rendering any faster than 20 FPS, even for a moment, then the user is likely to perceive a slowdown. 33 ms correlates to 30 FPS, or a 30-Hz refresh rate. Go lower than that with vsync on, and you’re into the bad voodoo of quantization slowdowns. 16.7 ms correlates to 60 FPS, that golden mark that we’d like to achieve (or surpass) for each and every frame, while a constant stream of frames at 8.3-ms intervals would correspond to 120 FPS.
At the 50-ms mark, both Intel IGPs put some numbers on the board. Surprisingly, the UHD Graphics 620 turns in a significantly rougher experience than its predecessor by this measure. We’d expect the more recent IGP to outpace its predecessor, thanks to its DDR4-2400 memory and significantly larger pool of L3 cache. That time spent past 50 ms could result in perceptible, unpleasant hitching during gameplay.
At the 33.3-ms mark, which corresponds to time spent under 30 FPS—a widely-accepted definition of “playable” for this class of product—neither Intel IGP puts significant numbers of milliseconds in its bucket. The UHD Graphics 620 still performs worse than the HD Graphics 620 of the i5-7200U, though.
The UHD Graphics 620 finally trades places with its forebear at the 16.7-ms mark, where increasing time past that threshold would mean time spent below 60 FPS. Here, the UHD Graphics 620 spends five fewer seconds on tough frames than does the i5-7200U and its HD Graphics 620 processor.
Meanwhile, our GeForce MX150-equipped machine spends hardly any time past 33.3 ms, only a second of our one-minute test run past 16.7 ms, and less than half the time past 8.3 ms of even the UHD Graphics 620 part. If high frame rates at low Dota 2 settings are a must, so then is an MX150.
Dota 2 (Best Looking)
You can’t have fast without slow, and Dota 2 offers a preset called Best Looking that turns up all of the game’s eye candy and returns resolution scaling to 100%. Before I dive into these numbers, I should be clear that we’re not trying to slag any of these chips by going with this setup. We’re not expecting playable frame rates. Rather, the point of this testing is to figure out the point past which these graphics processors run out of gas, and how hard they stumble when they do.
Best Looking really puts the hurt on all of these chips. Maybe a notch off “Fastest” would have been fairer. Both Intel IGPs deliver frame rates that simply aren’t conducive to playing Dota 2. The MX150 loses half of its average frame rate compared to Fastest, and its 99th-percentile frame times climb to potentially worrisome levels. Dota 2 can have teeth when it wants to.
I’ll focus on the MX150’s results here, since both Intel IGPs basically need to ride the sag wagon. The MX150 spends enough time past 50 ms and 33.3 ms that you’ll notice it, but the hitches and bumps that show up there probably aren’t numerous enough to prove ruinous to the gameplay experience. Once we get to the 16.7-ms mark, though, it’s clear that a maxed-out Dota 2 is still a bit too much for a consistently fluid experience on the MX150 at 1920×1080. If you simply can’t abide lower graphics settings than all-out, though, the MX150 copes OK.
Rocket League (1280×720)
Rocket League is another wildly popular title with a runs-anywhere reputation. We dialed in some fairly high settings at a low resolution to see how it played on this trio. Like Dota 2, Rocket League has a wonderful replay interface that lets us precisely play back a given match in-engine and from a player’s viewpoint.
In the integrated-graphics-only bracket, the UHD Graphics 620 actually delivers a big boost over the plain HD Graphics IGP. A 42-FPS average moves the i5-8250U’s IGP from “plays the game” to “playable game,” and the eighth-gen IGP’s 99th-percentile frame time suggests it can keep the experience playable most of the time.
Indeed, only the HD Graphics 620 chip posts any time spent past 50 ms in our reckoning. The UHD Graphics 620 only spends a heartening third-of-a-second past 33.3 ms over our one-minute test run, too, meaning its 99th-percentile frame time is closer to an exception than a rule.
The 16.7-ms threshold shows that the eighth-gen IGP isn’t quite up to delivering a consistently fluid experience, though. While it shaves a whole 10 seconds off the result of its predecessor, the UHD Graphics 620 stlil spends nearly a third of our test run under 60 FPS. Compare that to the MX150 and its mere 48 ms spent past 16.7 ms. Still, if you want to play Rocket League and don’t mind low-res visuals, the UHD Graphics 620 can certainly do it at playable frame rates.
Rocket League (1920×1080)
Our second round of e-sports stress testing cranks up the resolution to 1920×1080 while leaving all of our previous graphics settings in place.
Try as it might to get there, the UHD Graphics 620 can’t quite clear the 30-FPS average that defines “playable” for many folks at these settings. Its 24.7-ms 99th-percentile frame time is what we might call downright cinematic on the i5-8250U if it spends too much time past that figure. Clearly the Intel IGPs could benefit from a pruning of the eye candy at this resolution.
The GeForce MX150 loses plenty of performance when the resolution rises, as well, but an average frame rate in the 50s is still nothing to complain about. Its 23.7-ms 99th-percentile frame time isn’t quite as low as we’d like, but that still translates to a 42.1-FPS instantaneous rate 99% of the time. Not bad at all, if resolution is important to you.
Our “time-spent-beyond-X” charts actually paint a happier picture for the UHD Graphics 620. Where the HD Graphics 620 puts enough time up past 50 ms to prove less than enjoyable, the UHD Graphics IGP only spends 50 ms past 50 ms on those tough frames. You still might notice a couple lengthy hitches with it, but Intel’s latest GT2 IGP keeps out of the 50-ms danger zone much better than its predecessor.
At 33.3 ms, the UHD Graphics 620 keeps up the good work. It spends over 70% less time on tough frames that might result in drops below 30 FPS than the plain HD Graphics IGP does. At 16.7 ms, though, both Intel IGPs are pretty much winded. Still, if you want to push a 1920×1080 laptop display with Rocket League, the UHD Graphics 620 might perform pretty darn well with less eye candy, considering.
The MX150, on the other hand, only has a vanishing few frames that need more than 50 ms or 33.3 ms to complete, and it only spends six-and-a-half seconds on frames that take more than 16.7 ms to render. It remains a major step up in performance.
Tomb Raider (1366×768)
In our final all-hands test, I dug the 2013 reboot of Tomb Raider out of the catacombs of my Steam library to serve as a general-purpose example of how an older AAA title might play on these systems. I started my expedition at 1366×768 using the game’s High preset.
Compared to plain ol’ HD Graphics, the UHD Graphics 620 takes Tomb Raider from “unplayable” to “exceeds expectations.” Given that this is a full-fat AAA game with graphics settings to match, clearing a 32-FPS average on integrated graphics isn’t anything to sneeze at. That 99th-percentile frame time might be cause for alarm, but we can see just how much it affects our test run before passing judgment using our time-spent-beyond-X graphs.
As that 99th-percentile frame time suggests, you’re going to have a bumpy ride in Tomb Raider with the UHD Graphics 620, even if it can outperform its predecessor. Still, the i5-8250U’s IGP cuts more than a second off the HD Graphics 620’s time spent beyond 50 ms, and it spends less than half the time past 33.3 ms that its integrated cousin does. We’re not talking awesome gaming experiences from either chip by any stretch, but I’d certainly burn up some time in an airport terminal with Tomb Raider on the i5-8250U.
Of course, one could also step up to the GeForce MX150 and enjoy a truly fluid and responsive gameplay experience at this resolution. Tomb Raider doesn’t look bad at all at 1366×768, and if its visuals aren’t badly hurt by running at borderline-HD resolutions, well, no complaints.
Tomb Raider (1920×1080)
One more round of torture testing. As we did with Rocket League, I bumped Tomb Raider up to 1920×1080 while leaving all of our graphics settings at 1366×768 unchanged. Let’s see how these graphics processors handle it.
In the case of both Intel IGPs, it’s back to the sag wagon. Tomb Raider is jerky and unresponsive at these frame rates, and it’s difficult to even point Lara Croft in the right direction with this much lag between frames. 1920×1080 is simply too much to ask of these IGPs, which is where the MX150 steps in. With a 48-FPS average and a 29.2-ms 99th-percentile frame time, Nvidia’s entry-level discrete chip certainly clears the “playable” bar by a wide enough margin here.
If you prioritize resolution over fluidity, the MX150 is happy to oblige with Tomb Raider. The chip posts next to no time past 33.3 ms, and its 50-ms chart is squeaky-clean. I’ll take a consistent 30 FPS at 1920×1080 over no gaming at all.
The Witcher 3 (1600×900)
After I finished testing the first three games we’ve discussed, I felt the need to find a game that the Intel IGPs couldn’t run well, period. The Witcher 3 did the job. Even at 1280×720 and the lowest possible settings, neither Intel IGP could produce playable frame rates with CD Projekt Red’s fantasy RPG. I then turned to the GeForce MX150 and found that it did surprisingly well with this title at a blend of medium and low settings at 1600×900, so I pitted it against the mobile GTX 1060 6GB in the Alienware 13 R3 that plays host to the Core i7-7700HQ in our testing for comparison.
For as pleasing as the MX150’s performance is next to integrated graphics, the GTX 1060 6GB in my Alienware 13 puts its performance in perspective a bit with nearly four times the frames per second, on average. Of course, the desktop GTX 1060 6GB is almost four times as expensive as the MX150’s desktop cousin, the GT 1030, and the mobile version I’m showing here is wrapped up in a $1900 laptop. Not exactly fair play.
At 1600×900, the MX150 falls just short of delivering 99% of its frames under the 33.3-ms mark that would resemble a “console-like” gaming experience. I still enjoyed playing The Witcher 3 on a machine barely half an inch thick, so let’s see how long the MX150 spent on tough frames.
Not much at all, it turns out. At the 33.3-ms mark, the MX150 spends just about two-tenths of a second on frames that take longer than that to render. My gut impression of a playable Witcher 3 experience on this tiny chip wasn’t just fantasy.
With that, we can conclude our gaming tests on these systems. Although Intel’s UHD Graphics 620 IGP certainly provides more performance than its predecessor in the titles we tested, that performance usually goes toward crossing the line from “unplayable” to “tolerable” instead of fulfilling some greater ambition. For truly satisfying entry-level gaming performance, you really need a discrete graphics card like the MX150 to get there.
Another good reason to consider an ultrabook with Nvidia graphics inside goes beyond performance. I was able to download and install the latest Nvidia Game Ready drivers for the GeForce MX150 from Nvidia’s site without any intermediary interference, while attempting to download and install the latest Intel graphics drivers from the horse’s mouth led to a frustrating message to the effect of “this driver is not certified for your device.”
I have to wonder whether some of the performance issues I experienced with the titles I tested weren’t related to out-of-date graphics drivers. I turned to Acer’s official support site in hopes of finding an approved version of that driver for the two machines whose integrated graphics processors I was testing, but I came up empty-handed. Acer’s certified graphics drivers for its Swift 3 machines are months old in the case of the Kaby Lake-R Swift 3s and about a year old in the case of the i5-7200U-equipped system. If there are gaming performance improvements waiting in those driver updates, they’re locked behind Acer’s apparently glacial evaluation-and-approval process.
That tardiness in approving graphics drivers isn’t just of concern for gamers; in fact, it may not even be primarily of concern to gamers. The latest Intel drivers offer support for Netflix HDR and YouTube HDR content, as well as Windows Mixed Reality, the headlining feature of the Windows 10 Fall Creators Update. Given that Acer is now two versions behind in its driver-approval cycle, there’s no telling when Swift 3 owners whose machines only have Intel integrated graphics will be able to take advantage of those features.
Now, let’s roll up our sleeves and see if these machines work as well as they play.
We don’t usually include Futuremark’s wide-ranging PCMark benchmark in our test suite, but I’m making an exception for these general-purpose laptops. PCMark 10 tests some things that we don’t have tools for, like video conferencing, app launch times, and general productivity performance with common office apps like LibreOffice Writer and Calc. Considering that PCMark is a one-click test, it was hardly a burden at all to rotate it in for this review. Thanks to Futuremark for providing us with access to a PCMark 10 Professional license for these tests.
Along with the regular eighth-gen Swift 3, I’ve included the MX150-equipped model so that it can register results in the gaming portion of the PCMark test, as well as any OpenCL-enabled portions of the four benchmarks comprising the extended version of this test.
Let’s start off with these systems’ overall scores, including gaming results. These standings should come as no surprise, given the Alienware 13 R3’s potent processor and graphics card combo. The MX150-equipped Swift 3 pulls ahead of the system without a discrete GPU, and the i5-7200U ends up at the back of the pack.
The essentials benchmark tests web-browsing performance, video conferencing, and a range of app start-up times. The Speed Shift-enabled eighth-gen laptops might have an edge in this apparently latency-sensitive test. What’s surprising is that the most powerful PC of this bunch falls to the back of the pack.
The productivity test checks word-processing acuity and spreadsheet performance with both lightweight and heavy-duty number-crunching. Some portions of the test are even OpenCL-enabled, although it’s a bit odd that the i7-7700HQ-and-GTX-1060 combo doesn’t perform better as a result.
The digital-content-creation portion of this test covers three phases: photo editing tasks with large raw files from a range of DSLRs (some parts of which are OpenCL-enabled), video editing across both CPU and OpenCL code paths, and rendering and visualization of 3D models. Given those workloads, it’s no shock that the Alienware 13 R3 comes out on top again, but I would have expected more of a gap between the MX150-equipped Swift 3 and its IGP-only counterpart. C’est la vie.
Although we already know the story the PCMark gaming test tells, it’s nice to see it confirmed in yet another benchmark.
Overall, the PCMark suite paints a favorable picture for the performance of our i5-8250U-equipped machines in a range of common workloads. Gaming is, again, the only task where these chips struggle. Let’s see if our usual benchmark suite supports this broad impression of the performance of these systems.
These short, latency-sensitive benches show an interesting quirk of two of our test systems. You see, neither Kaby Lake system has Speed Shift enabled, while the i5-8250U does. Speed Shift allows the processor to ramp up to its peak clocks faster than chips without, and that pedal-to-the-metal approach lets the i5-8250U outpace the i5-7200U and shadow the i7-7700HQ for single-threaded performance and responsiveness.
In fact, I’m a bit annoyed that Dell didn’t decide to support Speed Shift for the Alienware 13 R3, given its high price and purported performance focus. Given that manufacturers don’t really talk about which processor power management minutiae like Speed Shift they support on spec sheets, though, it’s hard to figure out this kind of thing without a given machine in one’s hands.
Compiling code with GCC
Our resident code monkey, Bruno Ferreira, helped us put together this code-compiling test. Qtbench records the time needed to compile the Qt SDK using the GCC compiler. The number of jobs dispatched by the Qtbench script is configurable, and we set the number of threads to match the hardware thread count for each CPU.
Two more cores and four more threads give the i5-8250U a major boost in this multithreaded workload. It doesn’t quite halve the time the i5-7200U needs in this test, but a 35% reduction is still nothing to sneeze at.
File compression with 7-zip
Impressive. For file compression, the 15W i5-8250U is just about as competent as its 35W stablemate, and it leaves the dual-core i5-7200U completely in the dust.
Disk encryption with Veracrypt
Veracrypt, a continuation of the TrueCrypt project, offers a built-in benchmark that tests both the AES algorithm (which many of today’s CPUs can accelerate) and a variety of other algorithms that require good old CPU elbow grease to crunch. In the accelerated AES portion of the benchmark, the i5-8250U surprisingly comes out ahead of the i7-7700HQ.
The i5-8250U’s Twofish performance is no less impressive. It more than doubles the throughput of the i5-7200U and hangs right with the i7-7700HQ.
The evergreen Cinebench benchmark is powered by Maxon’s Cinema 4D rendering engine. It’s multithreaded and comes with a 64-bit executable. The test runs with a single thread and then with as many threads as possible.
The Speed Shift-equipped i5-8250U has a 3.4 GHz maximum Turbo speed, compared to the 3.8 GHz the i7-7700HQ can claim. Without Speed Shift and within a 35W TDP, however, the i7-7700HQ only ekes out a small advantage over the 15W part.
Cinebench is really about its multithreaded portion, though. Here, the higher TDP and beefy Alienware cooling system of the i7-7700HQ become an asset, as it allows the chip to boost higher than the i5-8250U can and for longer periods. The i5-8250U delivers a nice boost over its 15W cousin, but it can’t keep up with the i7-7700HQ.
Blender is a widely-used, open-source 3D modeling and rendering application. The app can take advantage of AVX2 instructions on compatible CPUs. We chose the “bmw27” test file from Blender’s selection of benchmark scenes to put our CPUs through their paces.
Blender makes liberal use of AVX instructions, and that fact plays to the i7-7700HQ’s advantage again. As with Cinebench, the 7700HQ’s better cooling and greater thermal headroom allows it to run way ahead of the more constrained i5-8250U.
Here’s a new benchmark for our test suite. Corona, as its developers put it, is a “high-performance (un)biased photorealistic renderer, available for Autodesk 3ds Max and as a standalone CLI application, and in development for Maxon Cinema 4D.”
The company has made a standalone benchmark with its rendering engine inside, so it was a no-brainer to give it a spin on these CPUs. The benchmark reports both a rays-per-second and time-to-completion figure, and we’re reporting the time-to-completion result.
Sorry to repeat myself, but the i7-7700HQ can’t help but run though the Corona benchmark much faster than the i5-8250U. Still, the 15W Kaby Refresh part handily outpaces the i5-7200U.
Handbrake is a popular video-transcoding app that recently hit version 1.0.7. To see how it performs on these chips, we’re switching things up from some of our past reviews. Here, we converted a roughly two-minute 4K source file from an iPhone 6S into a 1920×1080, 30 FPS MKV using the HEVC algorithm implemented in the x265 open-source encoder. We otherwise left the preset at its default settings.
The i5-8250U makes waiting for a transcode to finish much less onerous than it is with the i5-7200U.
CFD with STARS Euler3D
Euler3D tackles the difficult problem of simulating fluid dynamics. It tends to be very memory-bandwidth intensive. You can read more about it right here. We configured Euler3D to use every thread available from each of our CPUs.
This multithreaded and bandwidth-hungry benchmark yokes both the faster memory and extra threads of the i5-8250U to take a lead over the i5-7200U. The DDR4-2666 RAM and higher TDP of the i7-7700HQ still gives it the overall win in Euler3D, though.
Digital audio workstation performance
One of the neatest additions to our test suite of late is the duo of DAWBench project files: DSP 2017 and VI 2017. The DSP benchmark tests the raw number of VST plugins a system can handle, while the complex VI project simulates a virtual instrument and sampling workload.
We used the latest version of the Reaper DAW for Windows as the platform for our tests. To simulate a demanding workload, we tested each CPU with a 24-bit depth and 96-KHz sampling rate, and at two ASIO buffer depths: a punishing 64 and a slightly-less-punishing 128. In response to popular demand, we’re also testing the same buffer depths at a sampling rate of 48 KHz. We added VSTs or notes of polyphony to each session until we started hearing popping or other audio artifacts. We used Focusrite’s Scarlett 2i2 audio interface and the latest version of the company’s own ASIO driver for monitoring purposes.
A very special thanks is in order here for Native Instruments, who kindly provided us with the Kontakt licenses necessary to run the DAWBench VI project file. We greatly appreciate NI’s support—this benchmark would not have been possible without the help of the folks there. Be sure to check out their many fine digital audio products.
The DAWBench VI test at 96 KHz and 64 samples is the most punishing of this bunch, and none of our mobile CPUs can handle it. Even the i7-7700HQ delivered crackling and popping with no voices of polyphony in play. Relax the buffer size, and the i5-8250U more than doubles the performance of the i5-7200U.
The DAWBench DSP test is less about agility and more about pure multithreaded grunt. The i5-8250U triples the i5-7200U’s VST instances at 96 KHz and 64 samples. Relaxing the buffer size narrows the i5-8250U’s lead, but it still enjoys a pure doubling of performance versus the 7200U.
With our sampling rate duly reduced, we return to the VI portion of DAWBench. At 64 samples, the i5-8250U once again doubles the i5-7200U’s performance. At 128 samples, the i5-8250U’s advantage over the i5-7200U shrinks, but a 50% increase in performance still isn’t bad.
Returning to the more straightforward DSP test at a 48 KHz sampling rate, the 8250U holds its wide lead over its 15W predecessor at both buffer depths. Open and shut.
Overall, the Core i5-8250U represents a major leap forward for mobile digital audio workstation performance with 15W CPUs. A thin-and-light laptop with one of these chips inside won’t replace a beefy mobile workstation, but it opens up greater possibilities for DAW work in the ultrabook form factor than ever before.
To get a quick idea of how well the i5-8250U handles when it’s away from an outlet, we fired up the trusty TR Browserbench. Browserbench refreshes an older, Flash-heavy version of our website at one-minute intervals until the host machine shuts down. We’re working on a new version of Browserbench that more closely resembles modern browsing habits, but the old standby will have to do for now.
For consistency, each machine was charged to 100% before our test began. Display brightness was normalized relative to the 50% brightness setting of the eighth-generation Swift 3’s screen. Special power plans (like Windows 10’s battery saver mode) were disabled, and each machine was allowed to run the test until it hit 5% battery, at which point Windows automatically shut down. We then plugged the machine back in, fired it up, and recorded the final elapsed time that Browserbench indicated.
For our two Swift 3s with IGPs alone, battery life with Browserbench is practically identical—within 2%, generation-to-generation. That’s encouraging performance, given the i5-8250U’s two extra cores versus the i5-7200U. The MX150-equipped Swift 3 didn’t keep up with either of those machines in the battery life department, though. It conked out an hour earlier than the IGP-equipped notebooks. Even with Nvidia’s power-saving Optimus technology, the discrete graphics chip in the MX150-equipped Swift 3 still has a power cost when the laptop is away from a wall and in light use.
Watt for watt, the Core i5-8250U thoroughly outclasses its Kaby Lake predecessor in every test we can throw at it. Heck, Intel’s own 40% claim for generation-on-generation improvements from its 15W CPUs seems modest in light of our test results. The i5-8250U delivers a whopping 59% mean performance improvement over its predecessor when I take all of my tests into account. Intel can more than credibly claim a generational leap in performance from these chips, even if we’re “just” getting more Skylake cores in the same thermal envelope.
Heavy multithreaded workloads, and especially heavy AVX workloads, will drive the Core i5-8250U to its base clock faster than you can say “SIMD,” but even with its low 1.6-GHz base speed, Kaby Lake-R offers a major boost in performance over the i5-7200U in those tasks. For as much skepticism as that base-clock figure has gotten in some corners of the ‘net, I don’t think it’s worth worrying about. Folks who depend on heavy-duty applications like Blender and Handbrake to get their work done quickly shouldn’t throw away their mobile workstations just yet, but typical PC users will enjoy a nice helping of extra speed in heavy-duty work from ultrabooks with these chips inside.
Somewhat serious gamers are the only crowd I would expect to be unimpressed by the i5-8250U. By themselves, eighth-generation Core CPUs don’t do much to move the integrated-graphics bar. Low graphics settings and resolutions will be the order of the day with popular e-sports titles like Dota 2 and Rocket League on eighth-gen notebooks with integrated graphics alone.
If you want more graphics horsepower from an eighth-gen ultrabook, though, you can get one with a GeForce MX150 graphics chip and 2GB of dedicated GDDR5 RAM inside for not much more money than models without. If gaming performance matters, the Nvidia chip offers solid enough frame rates and frame times in some of today’s most popular games, sometimes even at 1920×1080 with some eye candy on. Lowering resolutions or relaxing the eye candy can result in downright smooth performance across the board.
Acer’s Swift 3 family of notebooks puts a good face on Kaby Lake Refresh chips in general. Although I’m spoiled by the premium laptops I have at hand in the TR labs, it’s hard not to be impressed by the plastic-free construction, rock-solid chassis, and responsive touchpads on these systems. If every $700-ish Ultrabook is as finely made, we’re truly living in a golden age for thin-and-light mobile computing.
All told, Intel’s eighth-generation mobile processors are a resounding win. We get a lot of the same world-class responsiveness from the Core i5-8250U that we do from Kaby Lake and Coffee Lake desktop chips, and we get more multithreaded power than ever before from a 15W Intel chip. Systems with these processors inside only seem to demand modest premiums over seventh-gen Core laptops, too.
Like our experience with the Core i7-8700K on the desktop, the i5-8250U’s improvements come with virtually no downsides. It’s a no-brainer to seek out an ultrabook with an eighth-generation Core chip inside if you’re on the hunt for a new thin-and-light PC.