For months now, we’ve been hearing about how great Ivy Bridge ultrabooks are going to be. We’ve heard they will reach lower price points, use enclosures made of either metal or composite materials, and feature a slew of other goodies, both standard and optional, like touch-screen displays. Intel hasn’t been shy about adding to the hype, predicting price tags as low as $599 and touting 2012 as the year ultrabooks will go mainstream.
Well, we’re about to see if reality lives up to the hype, because today Intel unwraps its first ultrabook-bound Ivy Bridge processors: 22-nm chips with 17W power envelopes tailored for uber-slim, ultra-light systems.
Unlike the quad-core mobile Ivy Bridge CPUs that arrived earlier this year (which we reviewed not long ago), these newcomers are based on a new version of Ivy Bridge with half the cores and half as much last-level cache. Normally at this point, we’d tell you the basic specs of the dual-core Ivy Bridge chip, but strangely, Intel PR rep Thomas Kaminski refused to answer our questions about the chip’s transistor count and die size. Intel virtually always divulges such information up front, and its reticence here may be an indicator that the firm is recovering problematic quad-core chips by disabling half of their cores and cache. We’ll have to crack open our review laptop in order to confirm, and we haven’t had time to do so yet. We do expect Intel to produce a natively dual-core variant of Ivy Bridge eventually, though.
Ivy Bridge DC, as Intel calls it, retains some of the amenities of its quad-core sibling. Intel has outfitted the processor with the full-featured version of its HD Graphics 4000 integrated graphics processor, or IGP. (Graphics clock speeds are substantially lower on the 17W parts, though.) Functionality like hardware virtualization, AES, TXT, and vPro support all remain fully enabled. Ultrabook-bound Ivy Bridge processors feature Turbo Boost 2.0, so clock speeds will scale dynamically based on available thermal headroom, and Hyper-Threading, so you’ll see four graphs instead of two in the Windows Task Manager. Also, of course, these puppies benefit from all of the other architectural improvements of Intel’s Ivy Bridge microarchitecture. For more details on those, be sure to check out Scott’s review of the Core i7-3770K.
Here’s a full list of the first 17W, dual-core Ivy Bridge models:
For reference, the quickest previous-generation 17W chip is the Core i7-2677M, which has a 1.8GHz base clock speed, a 2.9GHz peak Turbo speed, 4MB of L3 cache, and a $317 price tag in thousand-unit quantities. The Ivy Bridge-based Core i5-3427U is largely similar, but it has a substantially lower price tag: just $225. The i5-3427U does have a lower peak Turbo speed and less L3 cache, as well, but remember that Ivy Bridge is faster clock-for-clock than Sandy Bridge. We’d expect the two chips to perform similarly, or perhaps for the Ivy model to have a slight edge.
You don’t have to take our word for it, though. Intel sent us a prototype ultrabook with a Core i5-3427U inside, and we’ve run it through our revamped mobile benchmark suite alongside a first-gen ultrabook based on the Core i7-2677M. You’ll find the results in the next several pages.
But before we do that, let’s briefly introduce our guinea pig.
Introducing Intel’s Ivy Bridge ultrabook prototype
Intel was adamant that this prototype isn’t a production machine and shouldn’t be treated as such. But… well, we couldn’t resist snapping a couple of pictures anyway. After all, this is the first Ivy Bridge ultrabook we’ve gotten to play with, and it’s a looker:
This bad boy has a 13.3″ display (with a 1600×900 resolution), measures about 0.79″ at its thickest point, and tips the scales at a scant 3.22 lbs. Don’t let the shiny gray palm rest fool you: there isn’t a single metallic surface to be found anywhere on the outside. Still, the system feels surprisingly tough and rigid. Intel didn’t get into specifics, but we figure this might be one of those ultrabooks with composite enclosures we’ve heard so much about.
Inside that chassis dwells 4GB of DDR3 memory, a 240GB Intel 520 Series solid-state drive, and a 49.4 watt-hour battery. Connectivity includes mini HDMI, analog headphone, and dual USB 3.0 ports. SuperSpeed USB is, of course, standard on all 7-series chipsets, including this system’s UM77 Express.
According to Intel, systems similar to this one will retail for $1,000-1,100 when they hit stores. “When will that be,” you ask? Intel tells us the Ivy ultrabook launch is scheduled for June 5, but a “couple of” systems will already be out by then. I guess you can consider this the pre-launch… or something.
In any case, we hope production systems are more polished. This one had a few issues, including a buggy touchpad, odd noises coming from under the palm rest when the system was running, and a bit of play between the the display bezel and LCD panel. We’ll have to forgive those oversights, obviously, since this doesn’t even look like a pre-production system from a major vendor. It’s plastered with Intel logos and has “Ultrabook” etched in large letters on the lid.
Our testing methods
We’d like to thank Asus for sending us a Sandy Bridge-based UX31E ultrabook to compare with the Intel system. We actually reviewed (and benchmarked) the UX31E back in October of last year, but we’ve since refreshed our benchmark suite, and so we couldn’t use the old data. This time, Asus sent us a faster model featuring the Core i7-2677M, which happens to be the most comparable chip to the Core i5-3427U in the Intel reference machine.
We ran every test at least three times and reported the median of the scores produced. The test systems were configured like so:
|System||AMD A8-3500M test system||AMD A10-4600M test system||Asus N56VM||Asus N53S||Asus UX31E||Intel Core i5-3427U test system|
|Processor||AMD A8-3500M APU 1.5GHz||AMD A10-4600M 2.3GHz||Intel Core i7-3720QM 2.3GHz||Intel Core i7-2670QM 2.2GHz||Intel Core i7-2677M 1.8GHz||Intel Core i5-3427U 1.8GHz|
|North bridge||AMD A70M FCH||AMD A70M FCH||Intel HM76 Express||Intel HM65 Express||Intel QS67 Express||Intel UM77 Express|
|Memory size||4GB (2 DIMMs)||4GB (2 DIMMs)||8GB (2 DIMMs)||8GB (2 DIMMs)||4GB (2 DIMMs)||4GB (2 DIMMs)|
|Memory type||DDR3 SDRAM at 1333MHz||DDR3 SDRAM at 1600MHz||DDR3 SDRAM at 1600MHz||DDR3 SDRAM at 1333MHz||DDR3 SDRAM at 1333MHz||DDR3 SDRAM at 1600MHz|
|Audio||IDT codec||IDT codec with 126.96.36.19977 drivers||Realtek codec with 188.8.131.5237 drivers||Realtek codec with 184.108.40.20663 drivers||Realtek codec with 220.127.116.1177 drivers||Realtek codec with 18.104.22.16812 drivers|
|Graphics||AMD Radeon HD 6620G + AMD Radeon HD 6630M
with Catalyst 12.4 drivers
|AMD Radeon HD 7660G with Catalyst 8.945 RC2 drivers||Intel HD Graphics 4000 with 22.214.171.12496 drivers
GeForce GT 630M with 296.54 drivers
|Intel HD Graphics 3000 with 126.96.36.1992 drivers
GeForce GT 630M with 296.54 drivers
|Intel HD Graphics 3000 with 188.8.131.529 drivers||Intel HD Graphics 4000 with 184.108.40.20625 drivers|
|Hard drive||Hitachi Travelstar 7K500 250GB 7,200 RPM||WD Scorpio Black 500GB 7,200 RPM||Seagate Momentus 750GB 7,200-RPM||Seagate Momentus 750GB 7,200-RPM||SanDisk U100 256GB SSD||Intel 520 Series 240GB SSD|
|Operating system||Windows 7 Ultimate x64||Windows 7 Ultimate x64||Windows 7 Professional x64||Windows 7 Home Premium x64||Windows 7 Home Premium x64||Windows 7 Home Premium x64|
Thanks to Asus for volunteering a quad-core Sandy Bridge laptop, as well, and thanks to AMD and Intel for providing the other machines.
We used the following versions of our test applications:
- Stream 5.8 64-bit
- SiSoft Sandra 2012.SP3
- 7-Zip 9.20 64-bit
- TrueCrypt 7.1a
- Chromium 20.0.1096.0
- SunSpider 0.9.1
- The Panorama Factory 5.3 x64 Edition
- LuxMark 2.0
- CyberLink MediaEspresso 6.5
- WinZip 16.5
- AMD APP SDK 2.6
- Intel SDK for OpenCL Applications 2012
- x264 HD benchmark 4.0
- Battlefield 3
- The Elder Scrolls V: Skyrim
- Batman: Arkham City
- FRAPS 3.5.0
The tests and methods we employ are usually publicly available and reproducible. If you have questions about our methods, hit our forums to talk with us about them.
Memory subsystem performance
Per our tradition, we’re going to start off by comparing the memory subsystems of our CPUs in a few synthetic tests.
Please note that the A10-4600M, Core i5-3427U, and Core i7-3720QM have higher-clocked memory than the other two offerings. Because of the discrepancy, the results below won’t paint a clear, unadulterated picture of memory controller efficiency. But they will show us something else. You see, the A10-4600M, Core i5-3427U, and Core i7-3720QM all support faster RAM than their predecessors. (They can accommodate DDR3-1600 memory, while the A8-3500M, i7-2677M, and i7-2670QM are limited to DDR3-1333.) So we’re going to be able to see what dividends the faster memory support pays from one generation to the next.
The Core i5-3472U has slightly less memory bandwidth than its higher-clocked, quad-core sibling, the 45W Core i7-3720QM. Nevertheless, the jump from DDR3-1333 to DDR3-1600 gives both Ivy Bridge CPUs an edge over their Sandy Bridge counterparts.
Next up: SiSoft Sandra’s more elaborate memory and cache bandwidth test. This test is multithreaded, so it captures the bandwidth of all caches on all cores concurrently. The different test block sizes step us down from the L1 and L2 caches into L3 and main memory.
The 17W Ivy and Sandy CPUs are closely matched in this test, which kind of flies in the face of what we saw in Stream. The Ivy Bridge chip does have higher bandwidth at block sizes under 128KB, presumably thanks to quicker L1 cache, but that’s the starkest difference, and it isn’t very large. We see a much greater leap from the quad-core Sandy Bridge CPU, the Core i7-2670QM, to its quad-core, Ivy Bridge-based successor. In that case, the quad-core Ivy has a ~500MHz clock speed advantage over its Sandy Bridge-based counterpart.
Sandra also includes a new latency testing tool. SiSoft has a nice write-up on it, for those who are interested. We used the “in-page random” access pattern to reduce the impact of prefetchers on our measurements. We’ve also taken to reporting the results in terms of CPU cycles, which is how this tool returns them. The problem with translating these results into nanoseconds, as we’ve done in the past with latency measurements, is that we don’t always know the clock speed of the CPU, which can vary depending on Turbo responses.
Here, the 17W Ivy chip has slightly higher cache and memory latencies than its predecessor, in terms of absolute clock cycles. Interesting.
TrueCrypt disk encryption
TrueCrypt supports acceleration via Intel’s AES-NI instructions, so the encoding of the AES algorithm, in particular, should be very fast on the CPUs that support those instructions. We’ve also included results for another algorithm, Twofish, that isn’t accelerated via dedicated instructions.
7-Zip file compression and decompression
The Core i5-3427U is quicker than the Core i7-2677M in these tests, though only by a slim margin—one that may or may not be palpable without a stopwatch on hand. Remember that the i5-3427U has a 100MHz lower peak Turbo speed than the i7-2677M, though, and that it’s $92 cheaper. Intel seems to be offering slightly more performance for less, and that’s a good thing.
It’s also worth pointing out how close both of these 17W CPUs are to AMD’s fastest 35W Trinity offering, the A10-4600M. We’ll look at graphics in a little bit, but for CPU cores, Intel looks to have a serious performance-per-watt advantage so far.
The Panorama Factory photo stitching
The Panorama Factory handles an increasingly popular image processing task: joining together multiple images to create a wide-aspect panorama. This task can require lots of memory and can be computationally intensive, so The Panorama Factory comes in a 64-bit version that’s widely multithreaded. We asked it to join four pictures, each eight megapixels, into a glorious panorama of the interior of Damage Labs.
x264 HD benchmark
This benchmark tests one of the most popular H.264 video encoders, the open-source x264. The results come in two parts, for the two passes the encoder makes through the video file. I’ve chosen to report them separately, since that’s typically how the results are reported in the public database of results for this benchmark.
The Core i5-3427U pulls farther ahead of its predecessor in our x264 video encoding test. And, again, both 17W Intel processors outrun AMD’s top 35W Trinity chip.
Trinity, Llano, Sandy Bridge, and Ivy Bridge all dedicate a substantial chunk of their die area to graphics. With the exception of Llano, they all have special-purpose video transcoding logic, as well. We sought to unleash all of those extra transistors in a few general-purpose applications, to see if the competitive picture would change at all.
LuxMark OpenCL rendering
We’ve deployed LuxMark in several recent reviews to test GPU performance. Since it uses OpenCL, we can also use it to test CPU performance—and even to compare performance across different processor types. And since OpenCL code is by nature parallelized and relies on a real-time compiler, it should adapt well to new instructions. For instance, Intel and AMD offer integrated client drivers (ICDs) for OpenCL on x86 processors, and they both claim to support AVX. The AMD APP ICD even supports Bulldozer’s distinctive instructions, FMA4 and XOP.
A note about those missing bars in the graph. Sandy Bridge’s HD 3000 integrated graphics lack OpenCL support, so we couldn’t run LuxMark on the IGPs of the Core i7-2677M and Core i7-2670QM. Also, the AMD processors don’t support Intel’s ICD driver, so we were only able to run LuxMark on their integrated Radeon HD graphics and on their CPU cores using the AMD APP ICD. Ivy Bridge is the only processor that supports both AMD and Intel ICDs and has the ability to execute OpenCL code using its integrated graphics.
By combining its CPU cores with its OpenCL-supporting IGP, the Core i5-3427U can achieve substantially higher performance than the Core i7-2677M—87%, to be exact. The newcomer is even quicker than AMD’s A10-4600M.
The latest version of WinZip features a parallel processing pipeline with OpenCL support. The pipeline allows multiple files to be opened, read, compressed, and encrypted simultaneously, all with hardware acceleration. Right now, though, WinZip’s OpenCL capabilities seem to be off-limits to Intel processors—again, regardless of what ICD is installed. The OpenCL switch in the WinZip settings would only appear on our AMD systems.
We tested WinZip by compressing, then decompessing, a 1.17GB directory containing about 150 small text and image files, a couple dozen medium-sized PDF files, and 14 large Photoshop PSD files. We timed each operation with a stopwatch.
The results of our WinZip compression test are roughly in line with what we’ve seen so far: Ivy beats Sandy, Intel quad-core beats Intel dual-core by a big margin, and the A10-4600M doesn’t really distance itself from the 17W Intel CPUs—not even with its IGP chipping in via OpenCL.
Things are a little more puzzling in the decompression test. There, the Sandy ultrabook is oddly slow, while the Ivy ultrabook nips at the heels of its quad-core big brother. One would suspect a storage bottleneck, but both ultrabooks have solid-state drives, while our other machines have 7,200-RPM mechanical hard drives. Strange.
This user-friendly video transcoder supports AMD’s VCE and Intel’s QuickSync hardware transcoding blocks. Those are effectively black boxes without much programmability, so their output isn’t necessarily comparable—and neither is their performance, strictly speaking. From a practical standpoint, though, it’s helpful to see which solution will transcode videos the quickest. So that’s what we’re going to do.
For our test, we fed MediaEspresso a 1080p version of the Iron Man 2 trailer, and we asked it convert the clip to a format suitable for the iPhone 4. We tested with full hardware acceleration as well as in software mode. Where the setting was available, we selected encoding speed over quality. The A8-3500M was only run in software mode, since it lacks hardware H.264 encoding.
With QuickSync, the Core i5-3427U is nearly as fast as its 45W quad-core sibling. Pretty impressive. Encoding times balloon up in software mode, but the i5-3427U still beats its forebear, the i7-2677M, handily.
Note that the different encoding methods didn’t yield identical results. We didn’t see much of a difference in output image quality between VCE and QuickSync, but the output files had drastically different sizes. QuickSync spat out a 69MB video, while VCE got the trailer down to 38MB. (Our source file was 189MB.) Using QuickSync in high-quality mode extended encoding times slightly, but the resulting file was even larger—around 100MB. The output of the software encoder, for reference, weighed in at 171MB.
The Elder Scrolls V: Skyrim
Our Skyrim test involved running around the town of Whiterun, starting from the city gates, all the way up to Dragonsreach, and then back down again.
We tested at 1366×768 using the “medium” detail preset.
Now, we should preface the results below with a little primer on our testing methodology. Along with measuring average frames per second, we delve inside the second to look at frame rendering times. Studying the time taken to render each frame gives us a better sense of playability, because it highlights issues like stuttering that can occur—and be felt by the player—within the span of one second. Charting frame times shows these issues clear as day, while charting average frames per second obscures them.
For example, imagine one hypothetical second of gameplay. Almost all frames in that second are rendered in 16.7 ms, but the game briefly hangs, taking a disproportionate 100 ms to produce one frame and then catching up by cranking out the next frame in 5 ms—not an uncommon scenario. You’re going to feel the game hitch, but the FPS counter will only report a dip from 60 to 56 FPS, which would suggest a negligible, imperceptible change. Looking inside the second helps us detect such skips, as well as other issues that conventional frame rate data measured in FPS tends to obscure.
We’re going to start by charting frame times over the totality of a representative run for each system—though we conducted five runs per system to sure our results are solid. These plots should give us an at-a-glance impression of overall playability, warts and all. (Note that, since we’re looking at frame latencies, plots sitting lower on the Y axis indicate quicker solutions.)
We can slice and dice our raw frame-time data in other ways to show different facets of the performance picture. Let’s start with something we’re all familiar with: average frames per second. Though this metric doesn’t account for irregularities in frame latencies, it does give us some sense of typical performance.
Next, we can demarcate the threshold below which 99% of frames are rendered. The lower the threshold, the more fluid the game. This metric offers a sense of overall frame latency, but it filters out fringe cases.
Of course, the 99th percentile result only shows a single point along the latency curve. We can show you that whole curve, as well. With integrated graphics or single-GPU configs, the right hand-side of the graph—and especially the last 10% or so—is where you’ll want to look. That section tends to be where the best and worst solutions diverge.
Finally, we can rank solutions based on how long they spent working on frames that took longer than 50 ms to render. The results should ideally be “0” across the board, because the illusion of motion becomes hard to maintain once frame latencies rise above 50-ms or so. (50 ms frame times are equivalent to a 20 FPS average.) Simply put, this metric is a measure of “badness.” It tells us about the scope of delays in frame delivery during the test scenario.
The new, ultrabook-bound Ivy chip has a quicker IGP than its predecessor, but that doesn’t count for much in Skyrim at these settings. Both solutions trudge along with high average frame times and quite a bit of variance. From a seat-of-the-pants perspective, the two systems feel choppy, stuttery, and not really playable. Only the A10-4600M, with its Radeon HD 7660G integrated graphics, delivers a reasonably smooth, playable experience in this scenario.
Batman: Arkham City
We grappled and glided our way around Gotham, occasionally touching down to mingle with the inhabitants.
Arkham City was tested at 1366×768 using medium detail and medium FXAA, with v-sync disabled.
Looking at our frame time plots for Arkham City, we can see that the Core i5-3427U falls surprisingly close to the A10-4600M, and that its frame latencies exhibit almost as little variance.
Analzying the data more thoroughly confirms that the Core i5-3427U’s IGP performs much better here than it did in Skyrim, but it also shows the 35W AMD chips are faster overall by a decent margin. The 17W Ivy processor does spend less time above 50 ms than the 35W A8-3500M, but as we saw in our Trinity review, the A8’s score was skewed by a couple of big latency spikes.
Subjectively speaking, the Core i5-3427U seems to sit on the threshold of playability in Arkham City—the game is almost fluid enough, but maybe not quite. At least it’s an improvement over the previous-gen Core i7-2677M, which suffers from a larger number of latency spikes and higher average frame times.
We tested Battlefield 3 by playing through the start of the Kaffarov mission, right after the player lands. Our 90-second runs involved walking through the woods and getting into a firefight with a group of hostiles, who fired and lobbed grenades at us.
BF3 wasn’t really playable at anything but the lowest detail preset using these IGPs—so that’s what we used.
Even at the lowest detail preset with no antialiasing, the Ivy ultrabook struggles with Battlefield 3. Frame latencies aren’t terribly inconsistent—not compared to the A10-4600M, at least—but they’re high across the board, which makes the game feel sluggish and laggy. The game does render accurately, though, which is more than we can say for the the Core i7-2677M. With that processor, BF3 performs even worse and exhibits ugly visual artifacts.
Battery run times
We tested battery run times twice: once running TR Browserbench 1.0, a web browsing simulator of our own design, and again looping a 720p Game of Thrones episode in Windows Media Player. (In case you’re curious, TR Browserbench is a static version of TR’s old home page rigged to refresh every 45 seconds. It cycles through various permutations of text content, images, and Flash ads, with some cache-busting code to keep things realistic.)
Before testing, we conditioned the batteries by fully discharging and then recharging each system twice in a row. We also used our colorimeter to equalize the display luminosity at around 100 cd/m². That meant brightness levels of 70% for the Llano machine, 45% for the N53S, 40% for the Trinity system, 25% for the N56VM, 25% for the Ivy ultrabook, and 20% for the UX31E. The N53S and N56VM had larger panels than the other machines, though, which might have affected power consumption.
We should note one other caveat: these machines didn’t all have the same battery capacities. The batteries in the two quad-core Intel notebooks both had 56 Wh ratings. The Llano laptop had a 58 Wh battery, and the Trinity system’s battery was rated for 54 Wh. As for our ultrabooks, the Ivy system was rated for 49.4 Wh, and the Asus UX31E had a 50 Wh battery rating.
Our Ivy ultrabook managed to stay up 22 minutes longer than the Sandy Bridge ultrabook in our web-surfing test, but it shut down one minute earlier in our video playback test. Overall, this looks like a win for Ivy Bridge: even with slightly higher CPU performance and substantially better integrated graphics performance, battery life is just as good as or better than with Sandy Bridge.
So, there you have it. In its ultrabook-friendly incarnation, Ivy Bridge gives you higher performance and longer battery run times for less money than Sandy Bridge.
Not exactly a tough sell, is it?
That said, I’m a little disappointed Intel didn’t give the 17W Ivy Bridge variant a little more graphics oomph. The quad-core model’s higher IGP clock speeds seem to do wonders in the games we tested, but alas, the dual-core Ivy is quite a bit slower across the board—and that means games generally aren’t playable at the same settings. It’s a shame, really, and it rules out Ivy Bridge ultrabooks as compelling systems for on-the-road gamers—unless manfuacturers sneak in some discrete GPUs.
Our numbers have also given us a rough preview of how Intel’s 17W Ivy Bridge CPUs might compare to AMD’s upcoming 17W Trinity APUs. The AMD parts may have better graphics performance (though their GPUs will be clocked lower, too), but it’s looking like AMD has little chance of catching up on the CPU front—not when 17W Ivy Bridge matches or outruns the fastest 35W Trinity more often than not. AMD could end up having to offer its 17W parts at a substantial discount in order to be competitive. We’ll look at those when they come out.
For now, Ivy Bridge ultrabooks are the obvious choice for folks seeking solid performance in ultra-slim laptops. Users running previous-gen, Sandy Bridge ultrabooks probably don’t need to worry about upgrading, but I expect everyone else will at least want to take a look at the Ivy-based ultrabooks that come out over the next few weeks and months.