MSI’s Z87-GD65 Gaming motherboard reviewed

Despite the broader downturn in the industry, the PC gaming market grew 8% last year. No wonder we’ve seen an uptick in the number of gaming-oriented motherboards. These boards are similar to typical enthusiast fare, but they’re usually endowed with more premium peripherals and electrical components—and priced higher to match.

MSI’s Z87-GD65 Gaming is different. While the board has fancy components, a Killer NIC, and Creative-enhanced integrated audio, its $180 asking price is much lower than that of typical ATX gaming boards, which live well north of $200. Spending a little less on the motherboard makes perfect sense for gaming, since your dollars are usually better directed toward a more powerful graphics card.

There’s no shortage of competition in the middle of the enthusiast market, so the Z87-GD65 Gaming needs more than just conceptual soundness to rise above the fray. Fortunately, it’s also equipped with a bunch of firmware upgrades and all-new tweaking software. Both will have to deliver if MSI hopes to keep up with the innovations available in the Haswell-ready Z87 boards we’ve already reviewed: the Asus Z87-PRO and Gigabyte Z87X-UD3H. Let’s see how the MSI Z87-GD65 Gaming stacks up.

Were it not for the Far Cry offshoot’s clashing neon color scheme, I’d be tempted to dub the GD65’s aesthetic Blood Dragon. The red tone is particularly rich, and it makes the dragon graphic really stand out on the otherwise blacked-out landscape. The hue also harkens back to the bright red boards of MSI’s past, thankfully without the accompanying rainbow of multicolored slots and ports.

Black capacitors make the circuit board look particularly stealthy. They’re part of a collection of “Military Class 4” components that includes higher-efficiency capacitors and ferrite-core chokes. Anything less would be uncivilized on a modern enthusiast gaming board.

In this price range, so would having fewer than three PCI Express x16 slots. All three of the GD65’s x16 slots are connected to the CPU. 16 lanes of PCIe 3.0 bandwidth can be dedicated to the top x16 slot, split evenly between the top two slots, or divvied up between all three slots in an x8/x4/x4 setup. Most mobos link their third x16 slot to the chipset, where it has to share bandwidth with the PCIe x1 slots. The only PCIe sharing here occurs between the x16 slots.

MSI squeezes an mSATA connector between the slot stack and CPU socket. The mini SSD slot is linked the Z87’s sixth Serial ATA port, which means connected drives can participate in RAID arrays or SRT caching configs.

The mSATA slot is slung low enough to avoid bumping into longer expansion cards or oversized CPU coolers. Clearance shouldn’t be an issue for the VRM heatsinks, either. Their matching dragon-head designs are sloped, giving the socket some breathing room. This shape also makes it easier for my clumsy fingers to get at the CPU mounting hole that lies between them.

We can’t check every hardware combination for compatibility, but we can measure the distance between the CPU socket and various landmarks.

Apart from the VRM heatsinks, which are relatively short, the socket area is wide open. Note the proximity of the DIMM slots, though. Most Z87 boards have similarly-sized gaps between their sockets and memory, and problems often arise when pairing oversized CPU coolers with taller DRAM modules.

One might think the monstrous chipset heatsink would present clearance problems, as well, but it’s short enough to avoid interfering with longer expansion cards. The SATA ports are edge-mounted to keep cabling out of the way, too.

In the picture above, the first three pairs of SATA ports to the right are connected to the Z87 platform hub. The last pair also offers 6Gbps connectivity, but through an auxiliary ASMedia chip. Drives connected to that controller can’t participate in Intel-managed RAID arrays or caching schemes. Spoiler alert: the auxiliary ports are also slower than the Z87’s native ones.

Although comparable boards from Asus and Gigabyte split the Z87’s USB 3.0 connectivity using hub chips, the Z87-GD65 Gaming connects all of its SuperSpeed ports exclusively to the chipset. Four ports occupy the rear cluster, and two more are available via an internal header.

In addition to USB 3.0, the cluster houses a largely standard array of I/O ports. The clear CMOS button is a nice touch, and so is the addition of a coaxial S/PDIF output. More on the integrated audio in a moment. First, we have to address that bright red Ethernet port.

The port connects to a Killer E2205 networking controller, which is something of a novelty as far as integrating networking goes. In addition to supporting Gigabit speeds, the chip comes with complementary software that automatically prioritizes packets associated with games and streaming media. That’s a nice feature if you want to log a quick Battlefield session without pausing your BitTorrent download, but it doesn’t let your PC’s packets jump ahead of those associated with other systems on your local network—or on the Internet at large.

MSI has spruced up the integrated audio in several ways. The Realtek codec is the latest ALC1150 revision, which has a higher signal-to-noise ratio than the old chip but doesn’t add any features of note. Shielding protects the codec from electromagnetic interference, and the signal traces are isolated from other circuitry to reduce interference. Don’t ask me why the Audio Boost logo is backlit by a pink LED, though.

At least there’s some actual boost involved. The Z87-GD65 Gaming uses a Texas Instruments OPA1652 amplifier to drive headphones up to 600 Ω. Creative’s Sound Blaster Cinema software is layered on top, providing surround-sound virtualization, smart volume management, and a handful of routines that massage audio output to enhance dialog and other types of sounds. Too bad neither the Creative software nor the Realtek drivers support real-time encoding for multichannel digital audio. As a result, surround-sound gaming audio is limited to analog output.

The picture above provides a glimpse of some other onboard goodies: the Go2BIOS button boots the board right into the firmware, the POST code display provides diagnostic info, and the OC Genie button handles automatic overclocking. Hardcore overclockers also have access to voltage probing points in the opposite corner of the board.

You don’t have to be a seasoned tweaker to appreciate the port blocks included in the box. The blocks are a godsend when wiring front-panel connectors, regardless of whether it’s your first build or your fiftieth.

MSI throws in a couple of cheesy accessories—a sticker and a door-hanger—but neither adds substantial value to the overall package. More interesting things are afoot in the firmware.

Firmware tweaking

We’ve seen more innovation in motherboard firmware since the industry transitioned from the old BIOS standard to the new UEFI. MSI has cooked up a few UEFI enhancements for its 8-series boards, but it’s only made minor alterations to the main interface.

The UI combines BIOS-style menus with a smattering of mouse-centric elements, including a drag-and-drop boot sequence in the status window along the top. While the resolution is relatively low, the presentation is clean and crisp overall.

Thanks to snappy transitions and smooth mouse tracking, the firmware feels very responsive. MSI also speeds the tweaking process by allowing important variables to be keyed in directly. You won’t find exotic-looking sliders or drop-down menus among the overclocking options, though. MSI has reserved its major UI tweaks for the all-new fan control interface.

The Hardware Monitor consolidates fan control and real-time system monitoring in a single window. Navigating it with the keyboard is a little awkward, but the tabbed interface and draggable sliders are a good fit for mouse input.

This is the first time we’ve seen fan profiles actually illustrated in motherboard firmware. Although the points on the graph can’t be dragged with the mouse, they do reflect changes made with the sliders below. The graph also tracks temperatures and fan speeds, making me wish there were a way to apply an arbitrary CPU load from within the firmware. It would be nice to dial-in fan profiles without having to boot into the OS to check temperatures and noise levels under load.

In addition to looking good, the fan controls are quite robust. Minimum and maximum temperatures and fan speeds can be set for each of the two CPU headers and for three system headers. Static speed control is also available for the system headers. The minimum speeds are locked at 50% for system fans and 12.5% for CPU fans.

MSI has also added a Board Explorer feature that provides information on connected components. Clicking highlighted regions brings up information on the attached hardware. There are also pop-up windows that provide a better angle on the rear I/O cluster and internal SATA ports.

In another nice touch, the firmware’s OC Profile manager has been upgraded to show how a profile’s key CPU, DRAM, and IGP settings compare to current values. This comparison view is only available when loading profiles from USB storage, though. You’re flying blind when flipping through the six profile slots built into the firmware itself.

The practice of sneakily increasing Turbo multipliers seems to have permeated the industry. These days, it’s common for mobo makers to silently extend the fastest single-core Turbo multiplier to all-core loads, resulting in an effective 200MHz overclock when chips like the Core i7-4770K are fully loaded. This boost is usually enabled only when users define the DRAM frequency manually. However, beta versions of the Z87-GD65 Gaming’s firmware increase Turbo multipliers automatically; their default configuration technically voids the CPU warranty.

Obviously, motherboard firmware shouldn’t overclock the processor without the user’s explicit consent. MSI doesn’t do anything shady in the so-called MP firmware loaded onto retail boards and made available through its website. These official releases follow the correct Turbo multipliers even when the memory speed has been tweaked manually.

Adding to the confusion, the Enhanced Turbo feature responsible for boosting multipliers defaults to “auto” in both beta and MP versions of the firmware. There’s no way to tell whether the auto setting is juicing clock speeds without verifying frequencies manually on a loaded system. Thankfully, the Enhanced Turbo feature can at least be disabled.

Since MSI’s production firmware honors the correct Turbo multipliers, typical users should be safe from inadvertent overclocking. The misbehaving betas aren’t even available on the company’s website. Instead, they came to us directly from MSI along with a suggestion that we use the latest beta for testing. Perhaps I’m cynical, but that sure smells like an attempt to inflate benchmark scores artificially in online reviews.

Command Center software

MSI has all-new Windows-based tweaking software for its 8-series motherboards. Command Center replaces the old Control Center utility, and like its predecessor, the app gets off to a slow start. According to my stopwatch, Command Center takes 11 seconds to load. Tweaking utilities typically spend a few seconds polling system variables upon startup, but MSI’s software seem to take longer than everyone else’s. You can practically reboot into the firmware in less time than it takes to load the Windows utility.

Once open, the Command Center interface feels a little awkward. While the transitions and animations are smooth, the organization is somewhat disjointed, and a lot of the options are buried in sub-menus accessible only through the Advanced button in the bottom left corner of the screen.

Making changes should be easy for newbies and seasoned enthusiasts alike, though. Sliders abound, and values can be entered directly with the keyboard.

Command Center has enough overclocking options to satisfy most users. The fan controls are also extensive. As in the firmware, all five fan headers can be controlled individually. MSI has also added a tuning mechanism that determines the actual speed range of each connected fan.

CPU and system fans are accessed via different windows, but the functionality is similar for both camps. Users can click and drag up to two points along each fan’s profile.

The latest tweaking apps from Asus and Gigabyte offer many more manipulation points for each profile, making Command Center’s fan controls feel a little limiting. Those competing solutions also allow fans to be spun down completely—a stark contrast with Command Center, which enforces a 20% minimum speed for CPU fans and a 50% baseline for system fans.

Monitoring system variables is essential when fine-tuning a system’s cooling configuration, and Command Center’s history function makes it easy. Voltages, temperatures, and fan speeds can be tracked and saved with a simple interface. If only the historical records could be exported to a CSV or other common file format for further analysis.

In a bit of a surprise, MSI has integrated Dataram’s RAMDisk software into Command Center. Only a few clicks are required to set aside a portion of system memory as storage for one’s page file, browser caches, temporary files, and anything else. DRAM storage is volatile, so you’ll lose the contents of the drive if the power is cut. Fortunately, an automated routine backs up the contents of the RAM drive at regular intervals. The latest backup can be restored automatically after a reboot.

As much as I appreciate turning wicked-fast DRAM into system storage, SSDs make it awfully difficult to get excited about RAM disks. Flash-based drives offer practically instantaneous access times and cost a lot less per gig. They’re also available in much higher capacities, and they retain data when powered down. Our experience with Control Center’s RAMDisk component wasn’t entirely trouble-free, either. Allocating 4GB of our test system’s 16GB memory footprint produced a blue-screen error after we rebooted. We tried again with a smaller 1GB RAM drive and encountered no problems.

If you don’t want to tweak system settings with Command Center, the Z87-GD65 Gaming provides an alternative: an MSI-branded version of Intel’s Extreme Tuning Utility. The software has loads of overclocking options but is devoid of fan controls. With Intel getting out of the motherboard business, this may be the last time we see the XTU in action.

Overclocking

The Z87-GD65 Gaming supports overclocking via manual tuning and via its automatic OC Genie mechanism. OC Genie has two gears: the first took our Core i7-4770K to 4GHz on 1.1V, while the second pushed it to 4.2GHz on 1.2V. Both options cranked our Corsair Vengeance Pro memory up to its maximum rated speed of 2400MHz.

We’ve had the same Haswell CPU and Corsair H80 water cooler stable at 4.7GHz, so even OC Genie’s second gear is fairly conservative. The auto-overclocking routine relies on pre-defined profiles and can’t adapt to the individual characteristics of each CPU. At least the OC Genie configurations were stable, though. We’ve used more aggressive auto-tuning routines that produced configs that caused blue screens or throttling under load. It makes sense for auto-tuning mechanisms to do too little rather than too much.

Manual tuning remains the best way to get the most out of your CPU, so we ventured into the firmware to fiddle with the multiplier directly. We started at 4.2GHz and decided to let the motherboard choose the CPU voltage automatically. It didn’t choose wisely; the system booted but blue-screened after a few minutes under load. Feeding the CPU 1.25V in the firmware banished the BSOD, and we were soon in pursuit of higher frequencies.

Along the way to our 4.6GHz final speed, we discovered that CPU-Z’s voltage readings are completely bogus for Z87-GD65 Gaming. The voltages displayed by Command Center more closely matched what we set in the firmware, although they were consistently about 25 millivolts higher.

Haswell overclocking seems to be largely dependent on voltage. Too little, and the chip becomes unstable under load. Too much, and the CPU temperature climbs high enough to invoke throttling. On the Z87-GD65 Gaming, we hit 4.6GHz with the CPU voltage set to 1.3V. The CPU temperature spiked to 85°C, but we didn’t detect any throttling. The next step up the multiplier ladder required 1.375V to avoid BSODs under load—enough of an increase to push the CPU temperature past 90°C and provoke throttling.

We managed a throttle-free 4.7GHz with identical hardware on Asus’ Z87-PRO and reached 4.5GHz on Gigabyte’s Z87X-UD3H. Based on those results, the Z87-GD65 Gaming appears to have competitive overclocking chops.

Performance highlights

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 Z87-GD65 Gaming and comparable Haswell boards from Asus and Gigabyte. We also threw an Ivy Bridge-based Z77 board into the mix. We didn’t encounter any big surprises, but we do have some highlights to share.

The first one is a bit of a lowlight for the Z87-GD65 Gaming. Despite its branding, the board scored a little lower than its Z87 peers in our gaming test. It consistently produced slightly lower FPS averages and 99th percentile frame times. While the differences were consistent across a second batch of test runs, they were also far too slim to notice while playing.

Narrow gaps dominate our performance results, as you can see in the following application benchmarks.

Unlike games, which are primarily bound by GPU horsepower, applications tend to be bottlenecked by the CPU. No wonder these tests show larger deltas between the Ivy Bridge and Haswell platforms than between the individual Z87 boards.

We typically see more drama when measuring system boot times. Boot performance varies more from board to board, and each one has options to speed up the process.

Enabling all of the Z87-GD65 Gaming’s fast boot options shaves almost seven seconds off the boot time, allowing the board to get into Windows 2.5 seconds quicker than its closest rival. Every second counts, but not necessarily for very much. Cold booting is something few folks do more than once a day.

While the GD65’s boot time is still notably quick, its peripheral performance is almost entirely unremarkable. That’s not necessarily a bad thing; all Z87 boards should have comparable peripheral performance thanks to their shared chipset. We did, however, run into one exception in our networking tests.

Even though we disabled the Killer NIC’s real-time monitoring capability, the Z87-GD65 Gaming still exhibited higher CPU utilization in NTttcp. The Intel GigE controllers used by the other boards appear to be less demanding.

At least the Killer NIC kept up in the speed department. The Z87-GD65’s throughput matched that of the competition in not only NTttcp, but also our network file copy tests.

Power consumption

There’s usually some variance in power consumption between different mobos, but there 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.

The Z87-GD65 Gaming draws more power than its peers at idle and under load. The margins widen when the system is stressed, hinting that the GD65’s power circuitry may be less efficient than what’s employed by the Asus and Gigabyte alternatives. That said, even the 9W difference under load won’t amount to much on your utility bill.

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 GD65. Otherwise, flip to the next page for the full motherboard specs, details on our system configuration, and all of our benchmark data.

Detailed specifications

Most of the essentials have been covered already, but here’s a complete list of the Z87-GD65 Gaming’s hardware specifications and vital firmware options.

Platform Intel Z87 Express, socket LGA1150
DIMM slots 4 DDR3, 32GB max
Expansion slots 3 PCIe 3.0 x16 (x16/x0/x0, x8/x8/x0, x8/x4/x4) via CPU

4 PCIe x1 via Z87

Storage I/O 6 SATA RAID 6Gbps via Z87

2 SATA 6Gbps via ASMedia AS1061

Audio 8-channel HD via Realtek ALC1150
Wireless NA
Ports 1 DVI

1 HDMI

1 VGA

1 PS/2 keyboard/mouse

4 USB 3.0 via Z87
2 USB 2.0 via Z87

2 USB 3.0 internal headers via Z87

6 USB 2.0 internal headers via Z87
1 Gigabit Ethernet via Qualcomm Killer E2205

1 analog front out

1 analog center out

1 analog rear out
1 analog surround out
1 analog line in

1 analog mic in

1 TOS-Link digital S/PDIF output

1 coaxial digital S/PDIF output

Overclocking Per-core Turbo multiplier: 8-80X

Ring ratio: 8-80X
Base clock: 90-300MHz
DRAM clock: 800-3200MHz

DRAM reference: 200, 266MHz

CPU strap ratio: 1.0, 1.25, 1.66, 2.5

CPU core voltage: 0.8-2.075V

CPU core offset voltage: 0-0.99V

CPU ring voltage: 0.8-2.1V

CPU ring offset voltage: 0-0.99V

System Agent offset voltage: +/- 0-0.99V
Analog I/O offset voltage: +/- 0-0.99V
Digital I/O offset voltage: +/- 0-0.99V
PCH 1.05 voltage: 1.05-1.68V

PCH 1.5 voltage: 1.5-2.13V

DRAM voltage: 0.45-1.995V

Fan control CPU1,2: Min/max temperature, min/max speed (min 12.5%)

SYS1-3: Min/max temperature, min/max speed (min 50%)

SYS1-3: Manual speed 50-100%

Just look at all the scintillating detail you would have missed by skipping to the conclusion. You also would have missed this money shot of our test hardware:

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
Bios revision 1908 1007 F6B 1.3B2
Platform hub Intel Z77 Express Intel Z87 Express Intel Z87 Express Intel Z87 Express
Chipset drivers Chipset: 9.4.0.1017

RST: 12.5.0.1066

Chipset: 9.4.0.1017

RST: 12.5.0.1066

Chipset: 9.4.0.1017

RST: 12.5.0.1066

Chipset: 9.4.0.1017

RST: 12.5.0.1066

Audio Realtek ALC892 Realtek ALC898 Realtek ALC892 Realtek ALC892
Memory size 16GB (2 DIMMs)
Memory type Corsair Vengeance Pro Series DDR3 SDRAM at 1600MHz
Memory timings 9-9-9-24-1T
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:

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.

Memory bandwidth

Since all our systems used Corsair DIMMs running at 1600MHz with identical timings, don’t expect meaningful differences in memory bandwidth.

Productivity

Kraken JavaScript performance

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.

Video encoding

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.

Gaming

DiRT Showdown

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.

Boot 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.

For the most part, the Z87-GD65 Gaming is within a few percent of the other contenders in our application tests. There’s a bigger difference between the Z77 and Z87 boards than there is within the 8-series crowd.

Serial ATA performance

CrystalDiskMark

TR RoboBench

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.

If you’re connecting SATA drives to the Z87-GD65 Gaming, use the ports stemming from the chipset. They’re faster than the ones connected to the auxiliary controller. At least the ASMedia solution hits higher speeds than the Marvell chip used on the Z87X-UD3H.

USB performance

CrystalDiskMark

TR RoboBench

The USB 3.0 controller integrated into the Z87 chipset is very fast, and MSI doesn’t do anything to screw it up. Again, the board’s peripheral performance largely matches that of its peers.

PCI Express performance

CrystalDiskMark

For this test, we connected our PCIe SSD directly to the PCI Express lanes branching off the CPU. Since we used the same CPU on each board, it’s no surprise that the results are so similar overall.

Ethernet performance

NTttcp

TR RoboBench

Apart from its higher CPU utilization in NTttcp, the Z87-GD65 Gaming doesn’t distinguish itself in our networking tests.

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
Asus P8Z77-V 6 4 4 5 4 5 6 5 5
Asus Z87-PRO 6 5 5 5 4 4 5 5 5
Gigabyte Z87X-UD3H 6 4 4 5 3 5 5 5 5
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
Asus P8Z77-V 6 4 4 5 4 5 6 5 5
Asus Z87-PRO 6 5 5 5 4 4 5 5 5
Gigabyte Z87X-UD3H 6 4 4 5 3 5 5 5 5
MSI Z87-GD65 Gaming 6 4 4 5 3 4 5 5 4

Despite its Creative-infused integrated audio, the Z87-GD65 Gaming scores slightly lower than the others in our RMAA signal quality tests. Output quality doesn’t degrade when we apply a system-wide load, which would be more notable if the other boards weren’t able to maintain higher scores under similar conditions.

For onboard audio, features like surround-sound virtualization are arguably more important than minute differences in signal quality. If you’re really concerned with high-fidelity analog output, you’re much better off with a discrete sound card.

Conclusions

Let’s get one thing straight: gaming motherboards are no better at running games than typical enthusiast boards. There’s little that modern mobos can do to affect gaming performance, which is almost entirely bound by the system’s CPU and graphics card. Gaming motherboards can, however, offer features that might appeal to people who play games. That’s just what the Z87-GD65 Gaming attempts to do.

The integrated audio is bolstered by headphone amplification and surround-sound virtualization, both of which are well-suited to the stereo headsets that seem to be popular with gamers these days. Some of those folks may also appreciate the Killer NIC. While it won’t deliver substantially lower ping times, it can be useful for managing client-side network traffic. Lucid’s Virtu MVP 2.0 comes in the box, too, although a recent Intel driver update breaks its monopoly on allowing QuickSync video transcoding to work alongside discrete graphics. That combo is now supported natively by Intel’s Win8 graphics driver, though Windows 7 users still need Lucid’s GPU virtualization mojo.

MSI adds software of its own with Command Center. The new utility offers a decent array of tuning options but has a few rough edges. Command Center is still a work in progress. MSI expects to add a remote server capability that will enable smartphone-based tweaking and hardware monitoring. Android users can look forward to that functionality in late September, while iOS devotees will have to wait a little longer. As gimmicky as smartphone-based overclocking undoubtedly is, there’s real value in remote monitoring.

There’s nothing gimmicky about the firmware’s overhauled fan controls, though. The updated interface stands out as one of my favorite innovations of this first batch of Haswell boards. The main firmware GUI isn’t as flashy, but it’s easy to navigate and peppered with thoughtful features. MSI needs to stop messing with Turbo multipliers, though; restricting the practice to beta firmware doesn’t excuse the sneaky behavior.

Versus the other Haswell boards we’ve tested, the Z87-GD65 Gaming looks like a particularly good value. A lot of that is down to the $180 price tag, which is very affordable given the feature payload. The GD65 is a competitive all-around performer, too, and it appears to have solid overclocking potential. Add up all those attributes, and you have a good mid-range candidate for gamers, enthusiasts, and the substantial crossover between those two groups.

Comments closed
    • LoneWolf15
    • 6 years ago

    I’d rather have the previously reviewed Gigabyte board. Intel NIC, and what looks to be a better thought out UEFI.

      • mno
      • 6 years ago

      Speaking of Intel NICs, I really want to see a TR review of the ASRock Z87 Extreme6, since it has dual Intel NICs and is cheaper to boot (on Newegg especially with the current combo deal with CPU).

    • Bensam123
    • 6 years ago

    This may seem like a bit of overkill, but have you guys considered doing a top down shot of the board with a overlay that points out where each individual specific is on the board? Sorta like what Newegg does with some of it’s motherboards, only TR style. I think they’re quite informative. Also sorta representative in some of the newer bioses too with Board Explorer and such… I think it would add a bit to the article instead of just a chart.

    Curiously what LLC (Load Line Calibration) setting do you guys use while OCing? I’ve noticed you’ve never mentioned this, yet it’s an extremely important aspect of modern overclocking. It is definitely needed for stability at higher clock speeds and with finicky chips, it of course also has a downside to it (more stress on the CPU)… Do you guys just not use it or is it always set for extreme so you never mention it? Also it’s never mentioned if auto-OCing schemes use this or turn it up…

    It’s a fairly standard feature most OCers use…

    It is sorta funny that despite it being a ‘Killer Nic’ it’s never reached 10GigE speeds… You think that’s one of the first things they would’ve tackled to one up anyone that dares compare themselves…

      • ColeLT1
      • 6 years ago

      LLC has pretty much been made useless by the onchip VRMs of Haswell, and partially with adaptive voltage mode. LLC was there to offset voltage droop, when putting a high load on a chip the voltage would drop, and LLC would counteract that. With haswell, there is no voltage droop, it stays nice and flat. You can add LLC but it will just drive up your voltage under load, which is better done using adaptive mode (extra voltage only added once you go over stock turbo speeds).

        • Bensam123
        • 6 years ago

        I don’t think you understand how and why voltage drop exists… LLC doesn’t exist because motherboards produce erratic voltage.

        I would also like to see some sources backing up your ‘no vdroop’ on haswell.

        There is no single good definition online and you have to instead pull from multiple sources. You can start here:

        [url<]http://www.overclock.net/t/1313466/cpu-load-line-calibration-what-is-it[/url<] Even Anands definition isn't completely correct and needs to be updated: [url<]http://www.anandtech.com/show/2404/5[/url<] I guarantee you Haswell still has vdroop as it's part of the design of modern CPUs (not the motherboard or power delivery) in order to relieve stress from processors (and in turn increases their longevity), when they switch usage load levels (hence what load line 'calibrates'). It attempts to compensate for the vdroop, by applying more voltage when it takes place, not just increasing overall voltage (otherwise it would be no different from a voltage increase). LLC also changes the variance you would see in voltage (part of why Geoff was witnessing the voltage being a little bit out of spec from what he set).

          • ColeLT1
          • 6 years ago

          On sandy/ivy the voltage droops along with the core load, which is fine if you run it in stock form with a slower turbo bin for each core load, but most boards lock in all the turbos to the highest bin once you touch the ram speeds, or enable XMP. Fire off prime and do full threads, you will get a voltage, say 1.35v, but run prime with a single thread, and you get 1.39v. The higher the LLC the closer those 2 numbers will be together, you also have to drop your Voffset as you raise LLC because it will raise your overall voltage.

          On haswell, 1c, 2c, 3c, and 4c loads give me the same voltage, with LLC off.

          Edit: Here is what toms says, but I am just going by my experience with what I said above:
          [url<]http://www.tomshardware.com/reviews/core-i7-4770k-haswell-review,3521-10.html[/url<] Load-line calibration has very little, if any, effect with Haswell. There’s simply less flexibility to coax a killer overclock out of one of these CPUs using levers that were popular previously. Although the most enthusiast-friendly motherboards still include pages of options, a number of them simply don’t do anything beneficial. Flatly, it sounds like engineers and enthusiasts alike are still trying to figure out what the more obscure settings actually do.

            • Bensam123
            • 6 years ago

            LLC doesn’t just raise the voltage dude. Processors lock under high OCs due to the vdroop causing the processor to stall when it switches frequencies (up or down). It’s not dependent upon how fast it switches up or down though.

            I’ve never had to enable LLC when a motherboard is using core optimizations or anything that locks turbo multipliers to their highest state. It’s always when you start OCing the thing. The more you OC the higher you usually have to turn up LLC or you have to apply a lot more voltage to make up for it.

            Vdroop isn’t the same as a PSU being unable to supply the required voltage.

            You can’t monitor the effects of LLC without using some sort of specialized equipment with high speed measurement directly connected to the voltage regulators (I suppose maybe specialized software with haswell because it’s all on board, assuming there is a way to pull voltage every 1-5ms). You’re trying to look for something that lasts a fraction of a second to compensate for drops in power output. If your voltage for your processor is simply higher while under load or idle with LLC on that’s simply a side effect. Just looking at CPUZ wont tell you what LLC is doing, you’ll simply see side effects of it.

            This does raise a interesting point of how Haswell actually implements LLC… If the setting still exists it must be the motherboard telling the processor VRMs what to do (since you can still raise voltage and what not).

            Check out the two links I posted.

            I honestly wouldn’t trust anything out of Toms. It sounds like they don’t even know what LLC does.

    • ronch
    • 6 years ago

    I wish board makers would give their AMD boards this much attention to detail. All the most awesome boards are for the Intel platform. The FX may not reach up to Core i7 levels but there are still folks who want a sexy AMD machine.

      • DeadOfKnight
      • 6 years ago

      There’s no room for any more awesome in an AMD motherboard.

      • Farting Bob
      • 6 years ago

      Much less though. The market for higher end AMD boards is very small in comparison and if you can only sell 20% as many you arent going to give it all ther bells and whistles. AMD boards are generally cheaper at the low end and that is there main strength.

      • Deanjo
      • 6 years ago

      Really would need a new chipset for anything remotely exciting in AM3 land. Everything out there is pretty much the same thing as they were 3 years ago.

        • LoneWolf15
        • 6 years ago

        I don’t buy AMD, not because of CPUs –but because of chipsets. You’re absolutely right.

        Intel’s chipsets are so rock-solid…it’s a bigger reason to go Intel than the CPUs.

    • destroy.all.monsters
    • 6 years ago

    I have to assume that since it’s a Realtek chipset they’re using Creative’s MB-2 software over it and it isn’t going to improve (or necessarily weaken) the quality of the crab hardware underneath. The current KillerNIC is just an Atheros chipset and isn’t related whatsoever to the old MIPS based KillerNIC tech.

    So what you have here is a paint and tape “sports car” that’s aimed directly at the Asrocks (and to a lesser degree ASUS) by hitting the same tick boxes.

    Quality audio doesn’t come from the crab, and quality networking is generally from intel at this point.

    The design is cool – and I like the dragon graphic and the push buttons as well as the attention paid to the UEFI menus but it seems underwhelming overall.

      • ronch
      • 6 years ago

      When I heard that board makers are going to focus more on their audio implementations, I didn’t expect them to use this sort of gimmick where they use an ordinary audio codec with some crazy Creative software gimmick. I’d rather they actually put a Sound Core 3D in there, but it probably would end up making the board quite a bit more expensive.

      I guess Creative had to stoop to such low levels, that is, merely providing software to add some effects on top of a Realtek codec, to reclaim some of the market share and mindshare they’ve lost ever since onboard audio took off.

        • Wall Street
        • 6 years ago

        But they are improving the sound. You may not know what a TI op amp is, but the fact that it is driving the headphones should make headphone listening much better (if they made a nice clean power circuit for it). If they use op amps to buffer the other analog outputs I would be excited because we would be doing away with capacitor coupled outputs. But you probably don’t know what those are, because you fell prey to Sound Core 3D.

        Realtek is not synonymous with bad and Creative certainly isn’t synonymous with good.

          • ronch
          • 6 years ago

          No need to be a bit rude, dude. But yeah, I didn’t mention the op amps because I never use headphones, and I do read up on PC Audio so I’m quite familiar with the common terms. Realtek isn’t really bad at all — they just don’t have a reputation for discrete-class sound quality. Creative doesn’t automatically mean good quality either, but their discrete sound cards sound good enough, barring software issues, and their Sound Core 3D would probably sound better than any Realtek codec in such a limited space as the small corner of a motherboard where the sound circuits are normally found. If not, then at least the Sound Blaster 3D moniker will give MSI’s product additional ‘wow factor’, and isn’t that what MSI is aiming for?

          Anyway, my point is a matter of preference. I’d rather use the ALC1150 alone than use a piece of software that merely tries to make it look like a real sound blaster product, which it isn’t. Even Realtek has those fancy environmental effects found on their control panel app, but they don’t sound nearly as good as a Sound Blaster card.

          • UberGerbil
          • 6 years ago

          Downvoting you for condescension, upvoting ronch for classy response to same.

      • Bensam123
      • 6 years ago

      Yeah, I’ve learned to stay away from MSI. They have very attractive specs, design, price, and ‘quality’, but they have lots of reliability issues. So it’s all fluff and show.

        • Visigoth
        • 6 years ago

        I second this. I’ve had all sorts of problems with MSI motherboards, even my friends’ computers that I have to diagnose. Once I remove them and install another motherboard, all problems seem to magically go away. I don’t know if maybe they don’t do a thorough 24/7 stability test, or maybe their PCB is of a cheaper/lower quality, but over the years they keep releasing lower-grade motherboards compared to Asrock, Asus and Gigabyte.

          • Bensam123
          • 6 years ago

          I think they’re just designed to be cheap boards that are marketed well and unfortunately it’s really hard for hardware websites to test the ‘quality’ of something. People that know better don’t buy them.

          Scott’s latest A10 OCing results are a pretty good indicator of what you’re getting, same with Geoffs here…

          • ronch
          • 6 years ago

          I’ve encountered some HyperTransport-related issues with my old (circa 2007) MSI K9N-Neo-F, but apart from that, my last two MSI boards have been good enough. My current 990FXA-GD65 doesn’t have the finest UEFI BIOS around but I think it’s just fine. I hope they don’t fall down to the level of ECS though.

        • Stickmansam
        • 6 years ago

        I’m sort of caught in a bind since I live within a 20 min drive to a MSI service centre so even if my MSI stuff does have issues, I don’t have to ship it or anything since I just drop it off there. As long as its in warranty 😛

        Personally I prefer to use Gigabyte or ASUS but they’re usually a bit more expensive than MSI and MSI reliability has been good for me except the time I bought a damaged box mb which failed soon but I deserved that XD

    • danny e.
    • 6 years ago

    lost me at “.. and Creative…”

    I started boycotting them in 1999/2000 sometime and do not plan on changing that anytime soon.

      • danny e.
      • 6 years ago

      seems like several of the new haswell boards are going with some creative junk for some reason. ah, they’re trying to make my life difficult

        • ronch
        • 6 years ago

        No problem. You can probably skip the Creative software and use Realtek’s standard audio control panel alone instead.

          • danny e.
          • 6 years ago

          sweet.

      • ronch
      • 6 years ago

      This board actually uses a Realtek ALC1150, which is pretty good as far as audio codecs go. What I find interestly and a bit questionable is MSI’s gimmick to use Creative’s Sound Blaster Cinema software, which [u<]purportedly[/u<] adds audio effects. I guess it's similar to the way they use software with their own audio controllers to create EAX and stuff like that, except with this one they use your CPU for the job instead of their own chips. Having some audio gimmicks may be nice but I honestly don't see how this will actually make sound cleaner and fuller than a proper dedicated sound card. Should be fun to play with though.

    • phez
    • 6 years ago

    Definitely give the ROG boards a run for their money.

    • ronch
    • 6 years ago

    This would actually be my first choice if I were out for a Z87-based mobo. And enough with those fancy colors that Gigabyte and Asus use… this one looks seriously awesome.

Pin It on Pinterest

Share This