Intel’s much-anticipated Haswell microprocessor has arrived, and desktop users are a little underwhelmed—rightfully so. In most games and applications, the Core i7-4770K is only slightly faster and more power-efficient than its Ivy Bridge predecessor. When overclocked, Haswell appears to require beefier cooling than Ivy. And socketed desktop users don’t get a taste of more exotic configurations with integrated Iris graphics and embedded DRAM.
While desktop Haswell derivatives might not seem like dramatic upgrades over their Ivy peers, the accompanying motherboards have taken a larger step forward. The enthusiast-oriented Z87 Express platform has more native connectivity than the old Z77: three times the number of 6Gbps SATA ports and 50% more USB 3.0 connections. Motherboard makers continue to add new features, too, and some have completely overhauled their firmware and software interfaces.
Maybe I’m a bit biased, being TR’s resident mobo monkey, but the current crop of Z87 boards might just be the most exciting thing about Haswell on the desktop. We’ve already seen Asus’ Z87-PRO in action. Today, it’s Gigabyte’s turn under the microscope. Say hello to the Z87X-UD3H.
Oh, black and blue, where would enthusiast-oriented motherboards be without you? At least Gigabyte isn’t using the same tired old color scheme across its entire lineup. Black circuit boards permeate the firm’s 8-series family, but the accent colors differ based on the model. The PCB is also a departure from the norm; it has a matte coating rather than the glossy finish found on blacked-out boards from Asus and MSI.
The stealthy look is aided by black Chemi-Con capacitors rated to run for 10,000 hours at 105°C. These swanky caps are part of an Ultra Durable 5 Plus package that also includes high-end digital power circuitry from International Rectifier. Intel added on-die voltage regulation to Haswell, so I’m not sure if exotic motherboard power circuitry is really necessary. Surely the accompanying heatsinks are overkill.
Chunky heatsinks wall off two sides of the socket. The fin-less hunks of metal have relatively little surface area for heat transfer, though. Modern motherboard heatsinks seem to be more about form than function. They’ve become branding billboards, and the UD3H’s massive chipset cooler is a perfect example.
At least the chipset heatsink stays out of the way. The ones around the socket crowd some of the screw holes used by CPU coolers. Depending on your cooler’s retention bracket, the close proximity of the VRM heatsinks might make installation a little more difficult.
Provided you can wrestle their mounting hardware into place, larger CPU coolers have room to breathe on the UD3H. The images below highlight the distances between the socket and potential sources of clearance conflict.
Even at their peak, the VRM heatsinks are barely more than an inch tall. Pay attention to the distance between the socket and the DIMM slots. Most Intel boards put the memory slots in a similar position, and taller modules can easily interfere with aftermarket CPU coolers.
Despite having lots of surface area, the chipset heatsink is short enough to provide ample clearance for longer expansion cards. If you’re going to be running two graphics cards, Gigabyte recommends supplying a little extra juice to the PCIe x16 slots via that SATA-style power connector on the edge of the board.
The first two PCI Express x16 slots, which are to the right in the picture above, are linked directly to the CPU and can be arranged in an x16/x0 or x8/x8 config. The other x16 slot stems from the Z87 platform hub and offers four lanes of bandwidth. Those lanes are shared with the second and third PCIe x1 slots from the right. Only the x16 slots connected to the CPU offer gen-three connectivity; all lanes connected to the Z87 are limited to PCIe 2.0.
All eight of the Serial ATA ports support 6Gbps speeds. The black ports are tied to the Z87, which can arrange drives in RAID arrays and accelerate hard-drive performance with SSD caching. Initially, enabling hot-plugging kicked the Intel ports down to 3Gbps speeds on our board. An updated beta firmware (version F6G) resolved the issue, but that revision isn’t available to the general public. We’re told the update should be available online by the end of the week.
The grey SATA ports on the left come from an auxiliary Marvell controller. That chip can feed the two internal ports or a pair of eSATA connectors in the rear port cluster.
Six USB 3.0 ports dot the cluster, and they’re backed by internal headers for four more. Most of the SuperSpeed connectivity is fed by a pair of four-port Renesas hubs that split the USB 3.0 ports coming off the Z87 platform. The only ports with a direct line to the chipset are the two associated with the primary internal header.
A Realtek codec feeds the audio ports, including the digital S/PDIF output. Sadly, the board doesn’t support real-time encoding for multichannel digital output. Digital surround sound is limited to content with pre-encoded tracks, such as movies. Gamers who want surround sound will need to use the analog outputs. If you can get by with pseudo surround sound, the Realtek drivers support virtualization for stereo playback.
In addition to the Realtek chip, the UD3H sports a Texas Instruments amplifier capable of driving headphones up to 600 O. The amp is soldered on, so the arrangement isn’t as slick as the socketed OPAMPs available on some of Gigabyte’s higher-end Haswell boards.
Gigabyte kicks in a handful of other onboard goodies. Clustered with the usual mix of power, reset, and CMOS-clearing buttons are switches that toggle between the primary and backup firmware chips. The edge of the board is also lined with voltage probing points for obsessive overclockers. POST code display? Check.
The only separate accessory of note is the I/O shield, whose padded internal surface is much nicer than the usual array of sharp, pokey bits of metal. This shield will not only save your fingers from being sliced, but also prevent metal tabs from getting caught up in the ports as you slide the mobo into your system.
Motherboard firmware in HD
The Z87X-UD3H has all-new firmware. The 3D BIOS interface from the previous generation has been scrapped in favor of a radical redesign. Apparently, the last interface was made to satisfy the demands of Gigabyte’s marketing department. This time around, the engineers were given the freedom to do what they wanted. What they’ve created is quite good.
The first thing you’ll notice about the interface is the resolution. The full UI is drawn at 1080p, and it looks very crisp when displayed on a modern LCD. As a fallback, Gigabyte provides a low-resolution mode that ditches the hardware monitoring panes surrounding the central menu area. Users can boot into either resolution by default and switch between them with the touch of a button.
Like the Windows desktop, the firmware has a background wallpaper. You can choose between two pre-loaded options or supply one of your own. Given the number of hours Gigabyte’s engineers must spend looking at firmware screens, I can understand their desire to offer background customization.
Customization is a big part of the firmware, which has a home screen that can be tweaked to suit personal preferences. The home screen has six tabbed windows that can be filled with frequently used options from anywhere in the firmware. Each window can store multiple settings, so you can cover a lot of ground with the custom menus. As if that weren’t enough, a separate favorites panel can be filled with shortcuts to individual firmware settings. You can select which window is displayed when the firmware first loads, too.
The UI is effectively built around two tiers of tabs. Switching between them is easy using keyboard shortcuts: shift + direction controls the top row, while ctrl + direction handles the bottom one. These combos make keyboard navigation particularly speedy, and the interface is responsive enough to keep up. Mouse input is a little jumpy, though. Lag isn’t an issue, but the sensitivity feels too high.
Overall, the keyboard and mouse are both well-served by the interface. Users can manipulate settings by dragging sliders, scrolling through lists of options, and keying in values directly.
Like all too many motherboards we’ve encountered recently, the Z87X-UD3H sneakily increases the CPU’s Turbo multipliers when the memory speed is defined manually. Even bumping up the DRAM clock to 1600MHz, the maximum speed officially supported by Haswell desktop CPUs, causes the single-core Turbo multiplier to be applied to all-core loads. On the Core i7-4770K, that means a clock speed jump of 200MHz when the CPU is fully loaded.
200MHz isn’t a whole lot, but the principle is problematic. Firmware settings should be independent of each other; modifying one should never cause a completely unrelated setting to change. Also, motherboard firmware should never overclock anything without the user’s explicit consent. Even an extra 200MHz with all-core loads could technically void the CPU warranty.
Rather than indicating that the CPU has been overclocked, the firmware makes it look like the proper Turbo multipliers are being used. The screenshot above is from a config that uses a 39X multiplier regardless of how many cores are active. Turbo limits must be set manually to ensure the proper multipliers are used in conjunction with manual memory clocks.
We encountered similar behavior on Gigabyte’s Z77N-WiFi earlier this year, and other motherboard makers have employed the same tactic. Gigabyte product manager Jackson Hsu told us his company is simply trying to compete with rivals like Asus, which we’ve caught doing this kind of thing for years. Every bit of performance counts in a competitive market, even if it’s acquired by iffy means. Since Asus has pledged to cease clandestine Turbo boosting with its 8-series mobos, we hope Gigabyte will follow suit.
Fan controls are another pet issue of mine, and the firmware does an OK job on that front. Temperature-based speed control is supported for the CPU fan in addition to three system fans. SYS2 and SYS3 are lumped together, though, and you can only define the slope of the fan profile.
Much better fan controls are available via Windows software, but first, a blast from the past:
Hate the fancy new interface? Hitting F2 switches to a classic UI that retains the menu layout of Gigabyte’s old-school BIOSes. This mode can be set as the default, as well.
All-new tweaking software
The last few generations of Gigabyte motherboards have come with a disjointed collection of Windows tuning utilities. Like the old firmware, all that’s been scrapped for Haswell. Gigabyte has been working on its new EasyTune software since last summer, and the result is a bold departure from the company’s previous tweaking applications.
EasyTune now consolidates system monitoring, overclocking, power tuning, and fan controls in a single app. The interface was modeled after Windows 8 tiles, which probably isn’t something Gigabyte should have admitted. Despite being great for tablets, Modern UI tiles are one of Win8’s most reviled features on the desktop. They feel a little clumsy in EasyTune, too, mostly because they’re freaking huge. The default (and apparently smallest) window size is 1600×900, which makes the tiles massive compared to the average mouse pointer.
This isn’t the first time we’ve seen a Gigabyte tuning utility with oversized UI elements. The old TouchBIOS app had a jumbo-sized interface, too. Apparently, someone in Gigabyte’s software department thinks PC enthusiasts are using touchscreens.
Fortunately, tiles don’t dominate the entire interface. The advanced control panels have reasonably sized widgets, and they squeeze in a lot of settings as a result. All the big-ticket overclocking options are available except for the base clock strap. There’s a decent array of power options, as well. Individual settings have pull-down menus in addition to sliders, though there’s no way to enter values directly with the keyboard.
If you don’t like the color scheme, the UI accents can be changed from blue to green or orange. More color options are coming, too, perhaps to ensure users can match all those different highlights on Gigabyte’s motherboards.
The Smart Fan section of EasyTune presents pre-baked profiles in addition to manual tuning controls. There’s also a calibration routine that tests the RPM range of each fan connected to the board.
As in the firmware, the CPU, SYS1, and SYS2&3 fans are controlled separately. The interface lets you drag around five points on each spinner’s speed profile. Since the duty cycle scale goes all the way down to 1%, fans can effectively be silenced when temperatures slip below a given threshold.
While tuning your cooling config, you’ll want to keep an eye on system temperatures. EasyTune can do that with its integrated monitoring panel, which also tracks voltages and fan speeds in real time. I’m not a huge fan of how the monitoring window looks, especially since it suffers from the same oversizing that taints some other EasyTune elements.
EasyTune’s auto-overclocker could also use some work. The automated tuner tests for stability as it ramps up clock speeds and voltages, but it reports higher frequencies than other utilities. CPU-Z and Intel’s Turbo monitor both showed our CPU running 200MHz slower than the auto routine claimed during its iterative tuning process. In our first run, the auto-tuner settled on a supposed 4.9GHz final clock speed, but the CPU was actually running at only 3.9GHz.
Only manual tweaking delivered decent overclocking results on the Z87X-UD3H. The auto-tuner proved unreliable, and even EasyTune’s pre-defined profiles were problematic. Our system wouldn’t post with the medium and extreme presets, which should’ve taken the chip to 4.5 and 4.7GHz, respectively. We had a little more luck with the firmware’s “CPU upgrade” preset for 4.5GHz, which was at least stable in Windows. However, that setting pumped almost 1.45V through the CPU. Throttling kicked in as soon as we fired up our stress test.
Our experience overclocking Haswell tells us the chip is particularly sensitive to higher voltages. Serious cooling is required beyond about 1.3V, so it’s a little surprising that the preset overvolted the CPU so much. Even with a dual-fan Corsair H80 water cooler, our 4770K tends to throttle at voltages higher than 1.35V.
When turning up the clock manually, we ended up at 4.5GHz, which required 1.25V to maintain stability under load. Booting the board at 4.6GHz wasn’t a problem, but it took 1.35V to keep BSODs at bay. While we were able to get that speed stable under load, CPU temperatures spiked over 90°C and invoked throttling. Backing off on the voltage allowed the CPU to maintain 4.6GHz for a few minutes, though our load test soon brought about a dreaded blue screen.
On the Asus Z87-PRO, we took the very same CPU up to 4.7GHz using identical hardware. The CPU was running on 1.35V in that config, but temperatures stayed under 90°C. Throttling wasn’t a problem.
Based on what we’ve heard from mobo makers, there may be more overclocking variance with Haswell than there was with Ivy Bridge.
Comparative performance testing is a big part of what we do here at TR. However, it’s been a while since motherboards have had a substantial impact on system performance. These days, application and gaming performance tends to be bottlenecked by the CPU and GPU. Adding an SSD can improve overall system responsiveness, and faster memory can provide a boost in some scenarios, but that’s about it. Thanks to increased platform integration, even peripheral performance is largely consistent from one mobo model to the next.
There are rare exceptions to this general rule, and the only way to find them is to test motherboards exhaustively. We did so with the Z87X-UD3H and comparable Haswell boards from Asus and MSI. We also threw an Ivy Bridge-based Z77 board into the mix. We didn’t encounter any big surprises, though. The highlights of our findings are summarized below.
Despite using the same Corsair DIMMs, 1600MHz DRAM frequency, and 9-9-9-24-1T timings as the other boards, the Gigabyte is a little behind in our memory bandwidth test. That bandwidth handicap doesn’t slow the Z87X-UD3H in our application tests, though. The UD3H is consistently within a few percent of the Z87 competition.
At least as far as percentages go, there’s more of a spread in our boot test. With all its fast-boot options enabled, the Z87X-UD3H takes longer than its rivals to reach the Windows 8 Start screen. Admittedly, the differences are only a few seconds. That’s not a lot of time, and the delay has some benefit. Unlike with the other boards, we had no problem getting into the UD3H’s firmware by hitting the delete key during the POST process.
The Z87X-UD3H’s peripheral performance largely equals that of its peers. However, its Gigabit Ethernet implementation struggled in our network file copy test.
RoboBench pushes files with eight simultaneous threads, and the UD3H has no problem keeping up when those files are large movies. But the Gigabyte board is notably slower when copying our mixed set, which includes a couple of movies in addition to smaller MP3s, RAW images, Excel spreadsheets, and miscellaneous files associated with TR web content.
The Z87X-UD3H uses an Intel NIC, so we were surprised by the poor showing. The board’s slow mixed-set copy speeds persisted when we ran our three-loop RoboBench script a second time, though.
There’s usually some variance in power consumption between different mobos, but may be less with this generation due to Haswell’s on-chip voltage regulation. We measured power draw at the wall socket with our test system at idle, playing a 1080p YouTube video, and under a full load combining Cinebench rendering with the Unigine Valley demo.
It’s hard to get worked up about relatively small differences in power consumption. These deltas won’t amount to much on your power bill. For what it’s worth, the Z87X-UD3H is about as power-efficient as the Z87-PRO.
That’s it for our performance highlights. If you’ve seen enough test results, feel free to skip ahead to the conclusion for our final thoughts on the UD3H. Otherwise, flip to the next page for the full motherboard specs, details on our system configuration, and all of our benchmark data.
Most of the essentials have been covered already, but here’s a complete list of the Z87X-UD3H’s hardware specifications and vital firmware options.
|Platform||Intel Z87 Express, socket LGA1150|
|DIMM slots||4 DDR3, 32GB max|
|Expansion slots||2 PCIe 3.0 x16 (x16/x0, x8/x8) via CPU
1 PCIe x16 (shared with x1s, max x4) via Z87
3 PCIe x1 via Z87
1 PCI via PCIe bridge
|Storage I/O||6 SATA RAID 6Gbps via Z87
2 SATA 6Gbps via Marvell 88SS9172
|Audio||8-channel HD via Realtek ALC898|
1 PS/2 keyboard/mouse
6 USB 3.0 via Renesas uPD720210
2 USB 3.0 internal headers viauPD720210
6 USB 2.0 via internal headers via Z87
2 eSASTA via Marvell 88SS9172
1 Gigabit Ethernet via Intel I217-V
1 analog front out
1 analog center out
1 analog rear out
1 analog line in
1 analog mic in
1 digital S/PDIF output
|Overclocking||Per-core Turbo multiplier: 8-80X
Uncore multiplier: 80-80X
Host/PCIe clock: 80-133MHz
DRAM clock: 800-2133MHz
CPU gear ratio: 1.0, 1.25, 1.66, 2.5
CPU voltage: 0.5-1.8V
CPU graphics voltage: 0.5-1.7V
CPU ring voltage: 0.8-1.8V
CPU input override: 1.0-2.905V
System Agent offset voltage: -0.3 – 0.4V
Digital I/O offset voltage: -0.3 – 0.4V
PCH core voltage: 0.65-1.3V
PCH IO voltage: 1.05-1.9V
DRAM voltage: 1.15-2.1V
|Fan control||All: predefined silent, normal profiles
CPU, SYS1, SYS2&3 slope PWM: 0.75-2.5
There’s a reason I put that skip-ahead link on the previous page. This ginormous table kind of breaks up the flow unless you dig such nerdy details. If you do, you’ll probably appreciate a gratuitous system shot. Thanks for not skipping to the conclusion!
Our testing methods
We used the following system configurations for testing.
|Processor||Intel Core i7-3700K 3.5GHz||Intel Core i7-4770K 3.5GHz|
|CPU cooler||Corsair H80|
|Motherboard||Asus P8Z77-V||Asus Z87-PRO||Gigabyte Z87X-UD3H||MSI Z87-GD65 Gaming|
|Platform hub||Intel Z77 Express||Intel Z87 Express||Intel Z87 Express||Intel Z87 Express|
|Chipset drivers|| Chipset: 184.108.40.2067
| Chipset: 220.127.116.117
| Chipset: 18.104.22.1687
| Chipset: 22.214.171.1247
|Audio||Realtek ALC892||Realtek ALC898||Realtek ALC892||Realtek ALC892|
|Memory size||16GB (2 DIMMs)|
|Memory type||Corsair Vengeance Pro Series DDR3 SDRAM at 1600MHz|
|Graphics||Asus GeForce GTX 680 DirectCU II with 320.18 drivers|
|Hard drive||Corsair Force Series GT 120GB
Samsung 830 Series 256GB
OCZ RevoDrive 3 X2 240GB
|Power Supply||Corsair AX850 850W|
|OS||Microsoft Windows 8 Enterprise x64|
Thanks to Intel, Corsair, Samsung, OCZ, and Asus for providing the hardware used in our test systems. We should also thank the motherboard makers for providing their products for review.
We used the following versions of our test applications:
- 7-Zip 9.20 64-bit
- TrueCrypt 7.1a
- Chrome 27.0.1453.94
- x264 r2334
- DiRT Showdown demo
- CrystalDiskMark 3.0.2f
- FRAPS 3.5.9
- TR RoboBench 0.1
- RightMark Audio Analyzer 6.2.5
- Cinebench 11.529
- Unigine Valley 1.0
Some further notes on our testing methods:
- All testing was conducted with motherboard power-saving options enabled. These features can sometimes lead to slightly slower performance, particularly in peripheral tests that don’t cause the CPU to kick into high gear. We’d rather get a sense of motherboard performance with real-world configurations, though; we’re not as interested in comparing contrived setups with popular features disabled.
- DiRT Showdown was tested with ultra detail settings, 4X MSAA, and a 1920×1200 display resolution. We used Fraps to log a 60-second snippet of gameplay from the demo’s first race. To offset the fact that our gameplay sequence can’t be repeated exactly, we ran this test five times on each system.
- Power consumption was measured at the wall socket for the complete system, sans monitor and speakers, using a Watts Up Pro power meter. Our video playback load used this 1080p YouTube trailer for the movie Looper. The full-load test combined AIDA64’s CPU stress test with the Unigine Valley DirectX 11 demo running in a 1280×720 window.
- The Force GT 120GB SSD was used as the system drive for all tests. The Samsung 830 Series 256GB was connected as secondary storage to test Serial ATA and USB performance, the latter through a UASP-compatible Thermaltake BlacX 5G docking station. With RoboBench, we used the Samsung SSD as the source drive and the OCZ RevoDrive 3 X2 240GB as the destination for the read speed tests. Those roles were reversed for RoboBench’s write speed tests.
The Samsung/OCZ tag team also powered our Ethernet transfer tests. The RevoDrive served as the source and destination on the host system, while the 830 Series SSD performed those duties on the remote machine. That remote rig was based on an Asus P8P67 Deluxe motherboard with an Intel 82579 Gigabit Ethernet controller. The two systems were connected via a single Cat 6 Ethernet cable.
The Samsung and OCZ SSDs were secure-erased before each test that involved them. The Corsair drive was also wiped before we loaded our system image.
- Analog audio signal quality was tested using RMAA’s “loopback” test, which pipes front-channel output through the board’s line input. We tested with the boards idling and with a combined load consisting of Cinebench’s multithreaded rendering test, the Unigine Valley demo, and a CrystalDiskMark 4KB random I/O test running on the Samsung SSD attached via USB 3.0.
The tests and methods we employ are usually publicly available and reproducible. All tests were run at least three times, and we reported the median of those results. If you have questions about our methods, hit our forums to talk with us about them.
Since all our systems used Corsair DIMMs running at 1600MHz with identical timings, don’t expect meaningful differences in memory bandwidth.
We tested the latest Kraken release, version 1.1, in Chrome 27.
TrueCrypt disk encryption
TrueCrypt’s AES algorithm benefits from acceleration via Intel’s AES-NI instructions, which are supported by our Ivy Bridge and Haswell CPUs. We’ve also included results for another algorithm, Twofish, that isn’t accelerated via dedicated instructions.
7-Zip file compression and decompression
The figures below were extracted from 7-Zip’s built-in benchmark.
x264 video encoding
We’ve devised a new x264 test, which involves one of the latest builds of the encoder with AVX2 support. To test, we encoded a one-minute, 1080p .m2ts video using the following options:
–profile high –preset medium –crf 18 –video-filter resize:1280,720 –force-cfr
The source video was obtained from a repository of stock videos on this website. We used the Samsung Earth from Above clip.
We busted out our Inside the second methods to testing gaming performance. While we aren’t showing all of our fancy latency graphs, we have included results for FPS and the 99th percentile frame time.
Here, we measured the boot time after a full system shutdown. We used a stopwatch to time each test and stopped the clock when “Start” first appeared on the Windows 8 Start screen. Each board was tested with and without its fast-boot options enabled, and we took full advantage of Win8-specific features when in fast-boot mode.
The processor, graphics card, and storage are far more important to overall system performance than the motherboard—at least when the CPU is running at stock speeds. We configured the Z87X-UD3H’s firmware with the correct Turbo multipliers to ensure that the board didn’t gain an unfair advantage in our tests.
Serial ATA performance
TR developer extraordinaire Bruno “morphine” Ferreira created RoboBench, a scripted file copy benchmark that relies on Windows’ built-in robocopy command to execute eight parallel file transfer threads. The movie file set contains eight similarly sized files totaling 5.6GB, while the mixed set has a diverse collection of 14,000 files that adds up to 10.6GB.
Friends don’t let friends use third-party SATA controllers. The Z87X-UD3H’s Marvell chip is slower than not only the native Z87 implementation, but also the auxiliary ASMedia controller used by some of the other boards. Thankfully, the native ports offer competitive SATA performance overall.
For the most part, the USB scores are all very close. The Z87X-UD3H bounces around the standings, as do its rivals. System power management routines can affect performance in these kinds of peripheral tests, and there tends to be more variability from run to run. I wouldn’t be concerned about the relatively small differences in these results.
PCI Express performance
Most of our PCIe performance tests show the Z87X-UD3H on equal footing with its peers. However, the Gigabyte board is on the wrong side of an outlier in the random read test.
Despite struggling in our mixed-set RoboBench tests, the UD3H’s GigE adapter performs admirably elsewhere. I wonder if the lower CPU utilization measured by NTttcp has something to do the lower performance in RoboBench. Hmmm.
Analog audio signal quality
RightMark Audio Analyzer grades analog signal quality on a scale between “very poor” and “excellent.” We’ve translated those values to a numerical scale that starts at low of one and peaks at six. Higher values are better.
Our first set of results was gathered with the systems idling (apart from the RMAA app, of course). The second batch is based on tests conducted with the system under a combined CPU, GPU, and USB load.
|RightMark Audio Analyzer audio quality at idle: 24-bit/192kHz|
|Frequency response||Noise level||Dynamic range||THD||THD + Noise||IMD + Noise||Stereo Crosstalk||IMD at 10kHz||Overall score|
|MSI Z87-GD65 Gaming||6||4||4||5||3||4||5||5||4|
|RightMark Audio Analyzer audio quality under load: 24-bit/192kHz|
|Frequency response||Noise level||Dynamic range||THD||THD + Noise||IMD + Noise||Stereo Crosstalk||IMD at 10kHz||Overall score|
|MSI Z87-GD65 Gaming||6||4||4||5||3||4||5||5||4|
According to RMAA, the Z87X-UD3H’s integrated audio is roughly equal to that of its peers, at least as far as analog output quality is concerned. The scores for all the boards are very close, and there’s no apparent degradation in signal quality when we fire up our full system load.
Seemingly every new Intel desktop platform brings aboard functionality formerly handled by motherboards, making it that much more difficult for manufacturers to differentiate their wares. Gigabyte has found new ways to spice up its 8-series products, though.
The high-def firmware is the biggest single upgrade. Its 1080p interface looks fantastic—better than anything else I’ve seen in motherboard firmware. The windowed UI is also easy to navigate and customize. Too bad it jacks with proper Turbo behavior and offers only basic manual fan speed controls. The former is an easy fix, and Gigabyte should be able to improve upon the latter within the swanky framework it’s established.
Gigabyte’s EasyTune software has also been overhauled, albeit with mixed results. The revamped tweaking software is a nice upgrade overall: overclocking options abound, and the fan controls are particularly good. But EasyTune also has a massive desktop footprint that’s too often filled with oversized tiles. The auto-overclocking routine is a little flaky, too.
Next to its all-new firmware and software, the Z87X-UD3H’s hardware represents more modest evolution. The additional native 6Gbps SATA and USB 3.0 ports are welcome, and you shouldn’t run out of either. The headphone amp is also a neat touch. Gigabyte even pulls off a black-and-blue aesthetic without looking like a copycat.
Overall, the Z87X-UD3H has competitive features, performance, and power efficiency. At $170 online, it delivers a lot of goodness for a reasonable price. The higher CPU temperatures we encountered are a bit of a concern, but Haswell seems to be a little finicky when overclocked, and it’s too early to tell whether the UD3H represents an outlier or the norm for our CPU. More testing is needed, and we have additional boards lined up.
As we turn our attention to other Haswell boards on the market, I hope Gigabyte can iron out the handful of little kinks we discovered in the Z87X-UD3H. Quirks aside, this is a nicely balanced enthusiast offering with firmware that feels a lot more like a new generation than the Haswell desktop CPUs do.