MSI’s X99S MPower motherboard reviewed

Intel’s X99 chipset and Haswell-E processor represent the most devastating combo in the desktop division. The chipset opens with a flurry of USB and SATA ports, plus a dash of M.2 and SATA Express. Then comes the thunder: a 22-nm CPU with quad DDR4 memory channels, up to eight Hyper-Threaded Haswell cores, and as many as 40 PCIe lanes. No other desktop duo comes close to that kind of firepower.

One’s options are somewhat limited on the CPU front. Intel sells just three versions of the chip, and only the top model has all eight cores enabled. There are, however, myriad motherboard choices. Over 40 X99 boards are listed at Newegg, each one with a slightly different personality.

We’ve already tested Asus’ X99 Deluxe and its pared-down X99-A cousin, plus Gigabyte’s rave-tastic X99-UD4, and we’re not done yet. Today, the spotlight turns on another contender for the heavyweight belt: MSI’s X99S MPower.

Priced at $284.99, the MPower sits smack in the middle of the X99 range. And it continues the trend toward increasingly blacked-out motifs. Only sparing splashes of canary yellow interrupt the stealthy theme, though white LEDs embedded in the underbelly emit a ground-effect glow when the board is powered on. MSI includes a firmware switch for full stealth mode, which extinguishes all the onboard lighting save for the POST code display.

Although the MPower sticks to the standard ATX footprint, its selection of power connectors is a little unusual. In addition to sporting the 24-pin primary and 8-pin auxiliary 12V connectors present on most X99 boards, the MPower has a second auxiliary 12V plug meant for four-pin connectors. As far as I can tell, the extra input is supposed to be used in addition to the eight-pin connector rather than instead of it—for moar overclocking, of course.

Overclocking requires beefy cooling, which is somewhat at odds with Haswell-E’s crowded socket. There isn’t much room for oversized coolers between the VRM heatsink, memory, and closest PCI Express slot. Since we can’t test compatibility with every combination of components, we’ve provided some clearance measurements:

Note the two sets of memory clearances. The slots closest to the socket only need to be populated with eight-DIMM configs, while the second row is used by the four-DIMM setups we’d expect most folks to be running.

The closest potential obstruction is actually the VRM heatsink, but it’s shorter than standard-height DIMMs, so it’s unlikely to interfere with aftermarket coolers. As with most modern motherboards, memory modules with tall heat spreaders represent the biggest threat.

While we’re discussing clearances, we should visit the MPower’s expansion slots, which are somewhat awkwardly spaced for CrossFire and SLI configs.

The first and second x16 slots from the left are used for dual-card setups. Their close proximity precludes triple-slot coolers, and more importantly, it leaves no breathing room between double-wide duos. Cards installed in the second slot will also heat up the area around the M.2 storage slot, which could trip the thermal throttling mechanism in some mini SSDs.

Triple-GPU setups tap into the third or fourth slot depending on the CPU. The fact that Intel offers versions of Haswell-E with either 40 or 28 PCIe lanes makes things a little complicated, so we’ve mapped out the possible lane assignments in the block diagram below. Click the buttons to toggle between the different CPU configurations.


The third x16 slot is only active with 28-lane CPUs, while the fourth is meant for 40-lane chips.

In an interesting twist, the M.2 slot can switch between four Gen3 lanes in the CPU and two Gen2 lanes in the chipset. The CPU link provides more bandwidth to SSDs that can take advantage, while the chipset connection should offer more robust boot support via Intel’s Rapid Storage Technology software. A firmware toggle dictates the source of the M.2 lanes.

As with most 9-series motherboards, the chipset’s M.2 link is shared with SATA Express, which leads us nicely to the storage section on the next page.

Storage, ports, and the little things

All the X99S MPower’s internal storage ports are consolidated in the bottom right corner. The SATA cluster faces the edge, which makes clean cable routing a little easier, while the orphaned SATAe jack sticks out like a new standard begging for attention. Ahem.

The SATA Express connector doubles as dual Serial ATA ports, bringing the 6Gbps total up to 10. Only six of those ports can be combined in RAID arrays via Intel’s drivers, though. The remaining ports operate in single-drive IDE or ACHI mode and can only be used in RAID arrays managed by third-party software. That limitation is baked into the chipset—MSI can’t do anything about it.

On the USB front, the MPower serves up internal headers for four SuperSpeed ports. Another eight ports are accessible in the rear cluster.

All the MPower’s internal USB 3.0 headers are fueled by the chipset, as are two ports at the rear. The rest are supplied by a two-port ASMedia controller and a four-port Via one. Each of those third-party chips is tied to a single Gen2 PCI Express lane, so I wouldn’t recommend plugging too many high-speed devices into the associated ports. Our tests indicate that the native ports are faster, anyway.

MSI doesn’t provide a guide for which ports go where, so we’ve whipped up a legend for the connections in the rear cluster:

PS/2 LAN Audio
USB2 USB3 USB3 USB3 USB3
USB2 USB3 USB3 USB3 USB3
1 Gigabit Ethernet via Intel I210-AT
8-channel audio via Realtek ALC1150 and amplifier
2 USB 2.0 via X99
4 USB 3.0 via Via VL805 controller
2 USB 3.0 via ASMedia ASM1042AE controller
2 USB 3.0 via X99

Intel’s Gigabit Ethernet controllers can be less finicky than other solutions, so kudos to MSI for making the right choice on the networking front.

The onboard audio is based on the latest Realtek codec, complete with the usual enhancements. MSI’s Audio Boost upgrades include Chemi-Con capacitors, dual amplifiers, isolated traces, and additional codec shielding. Surround-sound virtualization is available through Realtek’s drivers, but there’s no real-time DTS mojo for multi-channel digital output.

At least the analog output is passable. The sound quality is acceptable, and my ears didn’t detect any hissing or other interference even with the system pinned by a combined CPU, GPU, and storage torture test.

One element of the port cluster not shown in our legend is the CMOS reset button next to the black USB ports. It lights up when the onboard LEDs are enabled, making the switch easier to find when peering around the back of a case. That’s one of a handful of little features that deserve special attention.

The MPower has a cushioned I/O shield that matches the blacked-out theme, port blocks that simplify front-panel wiring, voltage probing points for hardcore overclockers, and buttons for a range of different functions, including the OC Genie auto-tuner. That FASTB1 button on the right is one of two direct-to-firmware shortcuts. Hitting it boots the board into the UEFI without requiring the user to mash the delete key. The same effect can be achieved by holding the power button for four seconds.

The cable labels are a little cheesier, I’ll admit, but they do make sense for a board with this many SATA ports. Too bad there are only enough labels to cover six drives rather than the 10 supported by the board.

Next, let’s check out the tweaking interfaces and see how well the MPower overclocks.

Firmware and software tweaking options

Apart from the color scheme, X99S MPower’s firmware and software are pretty much identical to what you get with MSI’s other 9-series motherboards, including the Z97 Gaming 7 we reviewed last year. That article dives deep into both interfaces, so we won’t rehash all the gory details here. There are, however, a few highlights to point out.

The firmware’s overclocking section has most of the controls one might expect. It’s easy to navigate and loaded with tweaking options, but it also has a few quirks. The processor’s C1E sleep states are disabled by default, probably to boost performance in synthetic storage tests. Also, there’s no direct control over the CPU strap, which is adjusted automatically based on the base clock.

A more troublesome omission is the lack of proper support for DDR4-2800 memory. Despite the fact that those modules are available online from multiple vendors, the MPower lacks the memory multiplier needed to hit 2800 MT/s with the CPU’s stock 100MHz base clock. Overclocking is required to run DDR4-2800 RAM.

Activating the XMP profile for our Corsair DDR4-2800 DIMMs pushed the base clock to 127.3MHz, but the system wouldn’t boot at that speed. Our CPU evidently isn’t comfortable with such a high base frequency—at least without additional tweaking. For what it’s worth, this isn’t the first X99 board we’ve encountered with kludgy DDR4-2800 support.

MSI’s firmware fan controls are good enough that I can almost forgive the missing memory multiplier. Intuitive graphical profiles offer temperature-based speed control for each onboard header. These profiles work with both three- and four-pin system fans, but they only support four-pin units for the CPU.

The range of supported fan speeds is a little more limited than I’d like. System fans can only slow to 50% of full speed, while CPU fans have a 12.5% minimum. It would be nice if the firmware allowed both fan types to spin down completely.

The software-based overclocking and fan controls are similar to what’s in the firmware. Navigating the Windows utility’s interface is a little awkward, but the software is quite powerful, with configurable alerts and hooks for remote control via a companion smartphone app. There’s even an integrated RAM disk utility for those looking to set benchmark records.

Overclocking

Automated overclocking is managed by MSI’s OC Genie intelligence, which applies pre-baked profiles based on the system’s CPU. In first gear, the auto-tuner took our Core i7-5960X up to 3.7GHz at 1.05V, just a smidgen above the stock voltage. Second gear was only slightly more aggressive, kicking the frequency up to 3.8GHz at 1.15V.

To MSI’s credit, the system was rock solid with both configs. CPU temperatures stayed under 50°C during our torture tests, in part thanks to the 280-mm liquid cooler strapped to the chip.

Our Haswell-E processor is capable of much higher speeds with this particular cooler, and so is the MPower. I pushed the CPU to 4.3GHz with multiplier tweaks alone. The firmware’s “auto” defaults weren’t sufficient at higher speeds, but bumping the CPU up to 1.3V allowed it to reach 4.5GHz without issue. BSOD errors reared their head at 4.6GHz, much like on the other X99 boards we’ve tested with this CPU-and-cooler combo.

Overclocking with the X99S MPower is pretty straightforward regardless of which approach you take. Uninitiated newbies can get their feet wet with little effort, while seasoned veterans have access to a wide range of advanced options. The firmware even has built-in configuration profiles for extreme overclocking with sub-zero cooling. All the cool kids are huffing liquid nitrogen these days.

Performance highlights

When equipped with the same components running at the same speeds, motherboards typically have little impact on system performance. The X99S MPower’s awkward DDR4-2800 support forced us into a not-quite-apples-to-apples comparison, though. We had to test the board with a slower DDR4-2666 memory speed to prevent the CPU from running out of spec. Gigabyte’s X99-UD4 has a similar issue with DDR4-2800 RAM, so we tested it with the same slower memory speed as the MPower.

Turns out the slower frequency doesn’t make much of a difference outside of targeted memory bandwidth tests like Stream.

Even in this test, the gaps between the MSI and Asus boards are relatively narrow. The MPower has a little more pep than the UD4 despite using the same memory frequency and timings.

Having a little less memory bandwidth on tap doesn’t hurt either board in our application and gaming tests:

Although there are small differences between the boards in each test, none of the contenders stands out as being consistently faster or slower than its peers. That said, the MPower has a distinct advantage in cold boot time, where it’s a few seconds faster than the closest contender:

The value of a few seconds for something as rare as cold booting seems a bit dubious to me, but it’s tough to find clear-cut performance differences between modern motherboards. I’ll take what I can get, especially when the MPower’s peripheral performance is on-par with that of other X99 boards.

Power consumption

Motherboards usually influence power consumption, so we measured total system power draw (sans monitor and speakers) at the wall socket over a five-minute idle period and under a load comprising Unigine’s Valley graphics benchmark and Cinebench’s multi-core CPU rendering test. The Asus boards were tested with and without their EPU power-saving measures enabled.

Although the MPower isn’t as efficient as the UD4, it consumes less than 10W more, which is a pittance in the context of a complete system. That delta is too small to have a dramatic impact on system temperatures and noise levels inside a typical desktop rig.

Detailed specifications

Here’s the X99S MPower’s full spec sheet in case we missed anything.

Platform Intel X99, socket LGA2011-v3
DIMM slots 8 DDR4, 64GB max
Expansion slots 4 PCIe 3.0 x16 via CPU

2 PCIe 2.0 x1 via X99

Storage I/O 1 SATA Express via X99 (shared with 2 SATA RAID 6Gbps)

1 M.2 2280 PCIe 3.0 x4 via CPU/PCIe 2.0 x2 via X99

6 SATA RAID 6Gbps via X99

4 SATA 6Gbps via X99

Audio 8-channel HD via Realtek ALC1150 with amplifier

Surround virtualization via Realtek drivers

Wireless NA
Ports 2 USB 3.0 via X99

2 USB 3.0 via ASMedia ASM1042AE

4 USB 3.0 via Via VL805

4 USB 3.0 via internal headers and X99

2 USB 2.0 via X99

4 USB 2.0 via internal headers and X99

1 Gigabit Ethernet via Intel I210-AT

1 analog front/headphone out (amplified)

4 configurable analog ports

1 digital S/PDIF out

Overclocking All/per-core Turbo multiplier: 12-80X

Ring ratio: 12-80X

Base clock: 90.9-300MHz

DRAM reference clock: 200, 266MHz

DRAM: 1333-3200MHz

CPU core voltage: 0.8-2.1V

CPU ring voltage: 0.8-2.1V

System agent voltage: 0.8-1.85V

CPU input voltage: 1.2-3.04V

CPU VCCIO voltage: 0.7-2.3V

DRAM A/B voltage: 0.6-2.0V

DRAM C/D voltage: 0.6-2.0V

DRAM A/B VPP voltage: 1.24-3.77V

DRAM C/D VPP voltage: 1.24-3.77V

DRAM A/B reference voltage: 0.12-1.235V

DRAM C/D reference voltage: 0.12-1.235V

PCH 1.05 voltage: 0.7-2.32V

PCH 1.5 voltage: 0.8-2.3V

Fan control 2 x CPU (PWM), 3 x SYS (DC and PWM):

Fixed speed or four-point manual profile

Our testing methods

If you’ve made it this far, you may be curious about what our test system looks like. Here it is:

We used the following configurations for testing.

Processor Intel Core i7-5960X
Motherboard Asus X99-A Asus X99 Deluxe Gigabyte X99-UD4 MSI X99S MPower
Firmware 1004 0501 F9a M30
Platform hub Intel X99 Intel X99 Intel X99 Intel X99
Chipset drivers Chipset: 10.0

RST: 13.1.0.1058

Chipset: 10.0

RST: 13.1.0.1058

Chipset: 10.0

RST: 13.1.0.1058

Chipset: 10.0

RST: 13.1.0.1058

Audio Realtek ALC1150 Realtek ALC1150 Realtek ALC1150 Realtek ALC1150
Memory size 16GB (4 DIMMs)
Memory type Corsair Vengeance LPX DDR4-2800 SDRAM
Memory config 2800MHz @ 16-18-18-36-2T 2800MHz @ 16-18-18-36-2T 2666MHz @ 16-18-18-36-2T 2666MHz @ 16-18-18-36-2T
Graphics Asus GeForce GTX 680 DirectCU II with 340.52 drivers
Storage Corsair Force Series GT 120GB

Samsung 840 Pro 256GB

Power supply Corsair AX850 850W
OS Microsoft Windows 8.1 Pro x64

Thanks to Asus, Cooler Master, Corsair, Intel, and Samsung for providing the hardware used in our test systems. And thanks to the motherboard makers for providing those.

We used the following versions of our test applications:

Some further notes on our test 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. The full-load test combined Cinebench’s multithreaded CPU rendering test with the Unigine Valley DirectX 11 demo running in a 1280×720 window. We reported the peak power consumption during the Cinebench run. Our idle measurement represents the low over a five-minute period sitting at the Windows desktop.
  • The Force GT 120GB SSD was used as the system drive for all tests. The Samsung 850 Series 512GB was connected as secondary storage to test Serial ATA and USB performance, the latter through a USAP-compatible Thermaltake BlacX 5G docking station. The Samsung SSD was secure-erased before each test that involved it. The Corsair drive was also wiped before we loaded our system image.
  • Ethernet performance was tested using a remote rig based on an Asus P8P67 Deluxe motherboard with an Intel 82579 Gigabit Ethernet controller. A single Cat 6 Ethernet cable connected that system to each motherboard.
  • Analog audio signal quality was tested using RMAA’s “loopback” test, which pipes front-channel output through the board’s line input. We tested while the system was loaded with Cinebench’s multithreaded rendering test, the Unigine Valley benchmark, and a CrystalDiskMark 4KB random I/O test running on the Samsung SSD attached via USB 3.0.

The tests and methods we employed are publicly available and reproducible. All tests except power consumption were run at least three times. Unless otherwise indicated, we reported the median result for each test. If you have questions about our methods, hit our forums to talk with us about them.

Conclusions

MSI’s X99S MPower is a solid foundation for a Haswell-E build. It takes full advantage of the platform’s capabilities while letting overclockers wring even more performance from the CPU.

Nailing the formula is tricky for a board this complex, and a couple of issues do stand out. The lack of proper multiplier support for DDR4-2800 memory is frustrating, especially since that speed grade isn’t exactly rare. Also, the slot spacing seems geared toward three-way setups at the expense of breathing room for more popular (and sensible) dual-card configs. Both issues will only affect certain configurations, so they’re hardly show-stoppers.

Otherwise, the MPower is great. It’s targeted specifically at the overclocking crowd, including the extreme subset that seeks benchmarking records with sub-zero cooling. MSI has integrated a handful of features specifically for folks on the ragged edge, but it hasn’t forgotten about more casual users. The integrated auto-tuner is conservative enough to give less experienced users a boost without requiring extreme cooling, and the manual controls make it easy to push the CPU’s limits within the bounds of conventional cooling.

Even at stock speeds, the MPower has a lot to offer. USB 3.0 connectivity abounds, and all the major storage standards are covered. The dual-mode M.2 slot is particularly appealing given the challenges associated with booting some mini SSDs—and the greater speeds possible with next-gen units. And don’t forget the graphical fan speed controls, slick boot-to-firmware options, and builder-friendly extras.

Although a couple of niggles keep the MPower from being our favorite X99 motherboard, they don’t spoil the overall package. This is a solid choice for overclockers and enthusiasts of all levels.

Comments closed
    • CrazyElf
    • 5 years ago

    The X-Power version of this board does seem to fix most of these issues, although on the downside, it costs ~$100 USD more. It’s probably worth it, imo. If you’re going to spend so much on a Haswell-E chip, it’s something that you might as well do.

    I like MSI boards, although like all manufacturers they have issues. Not all the fan headers are PWM for example and I don’t like Killer NICs (higher CPU use). But otherwise they make good boards and I have had good experiences with their tech support.

    My big wish is to see an MSI socket 2083 motherboard of Haswell-E, as Gigabyte and Asus now have such boards out.

    • Deanjo
    • 5 years ago

    I really wish I could trust MSI’s boards but I’ve just had waaaay to many issues with them (particularly their buggy BIOS’s) in the past. They have some nice boards feature wise but if I can’t depend on them working properly I might as well be throwing money into the trash.

      • Waco
      • 5 years ago

      I hate that they use Killer NICs more than anything else…

    • Zorb
    • 5 years ago

    That’s the nicest looking board i’ve ever seen from MSI…. keep it up!

      • Pez
      • 5 years ago

      She’s a looker alright. Now just needs some yellow watercooling loop goodness…

Pin It on Pinterest

Share This