MSI X99A Gaming Pro Carbon Broadwell-E Motherboard

MSI’s X99 refresh is here in time for Intel’s new Broadwell-E processors! We put the new processor through its paces and our MSI X99A Gaming Pro Carbon was the lucky winner that we chose to torture in order to determine it’s "Broadwell-E readiness." Our time with the X99A Gaming Pro Carbon was filled with many surprises.

continued...

Subsystem Testing

NOTE: For all Subsystem Testing, an Intel Core i7 6950X engineering sample (3.0GHz / 3.5GHz Turbo) and 4x 8GB Corsair Vengeance LPX (2666MHz DDR4 16-18-18-36-2T@1.35v) memory modules running at DDR4 2133MHz speeds (stock testing, up to 2666MHz overclocked) were used. The CPU was cooled with a Koolance Exos 2.5 and CPU-370 water block. A Diamond Multimedia Radeon HD 7970 GHz Edition reference card running Catalyst 15.8 drivers was used for discreet video. Windows 10 Pro (x64) was used straight from the .ISO with no updates. A customized PowerShell script was run to disable telemetry and other non-essential services.

Sound Hardware

The X99A Gaming Pro Carbon uses the Realtek ALC1150 7.1 channel HD audio CODEC. MSI’s Audio Boost 3 implementation features audio enhancement by Nahimic for 7.1 channel emulation through stereo speakers. The audio solution also features PCB circuitry isolation, dual 600ohm headphone amplifiers, EMI shielding, front and rear amplified audio ports, and lastly Chemi-Con dedicated audio capacitors.

The following specifications were taken from the manufacturer’s website:

• Realtek® ALC1150 Codec

- 7.1-Channel High Definition Audio

- Supports S/PDIF output

Audio – Subjective Listening

For subjective listening you want to listen to something that covers a range of sound types. For this portion of the review I went with Five Finger Death Punch’s American Capitalist CD.

CD audio was excellent, just as I had anticipated.

Gaming Audio Quality

This is a new test, but I have been doing this for some time on a few of our motherboards to get a sense of how well these onboard audio solutions behave in regard to playing actual games. Due to the ease in which I can simply copy the game over the network without installing it, I have selected EA/BioWare’s Knights of the Old Republic MMO. No matter how you feel about this game, one area it does well in is audio. The iconic Star Wars music is part of the game’s sound track and its various themes can be heard in the game world. This is music that many people grew up with having watched the movies over and over again, and I think allows for ease of comparison on different platforms and sources.

The audio playback was quite excellent. It’s difficult to compare it to audio solutions I haven’t worked with in a while, but the X99A Gaming Pro Carbon’s audio compares favorably to any Realtek ALC1150 based solution I’ve ever worked with while still coming up short against the Creative Labs or ESS based solutions I’ve tested. MSI deserves a lot of credit for taking the ALC1150 and pushing it as far as they have in terms of sound quality. I think ASUS handles it better on select ROG models because the OP-AMP they use is essentially user programmable, but this one is a close second. This one also lacks the auto-switching impedance the ASUS models have. Having said that, the price point of the X99A Gaming Pro Carbon is a little less than that of the highest end ROG models.

Audio – Microphone Port Testing

The onboard audio MIC-IN port was tested using a Logitech Internet Chat Headset. Spoken words were recorded from the Windows Sound Recorder found under the Accessories folder in the start menu within Windows. The recording was using the highest quality settings available in the control panel for the audio device being used to record.

With the microphone boost option disabled the recording sample was almost crystal clear although the audio levels suffered. With the microphone boost option enabled, the sample was much louder, but introduced a slight amount of distortion.

DPC Latency

Deferred procedure call latency or DPC testing is something that we’ve been asked about and this is the first article we’ve done which integrates that type of testing. For those who may not know what DPC is I’ll explain. Deferred procedure calls are a function within Windows that allows higher priority tasks such as device drivers to defer lower priority tasks for execution at later times. It’s an interrupt and reassignment of sorts performed by the operating system.

DPC latency varies from board model to model and brand to brand. DPC issues show up in the form of audio dropouts and streaming video issues. Naturally this is something that the enthusiast would want to avoid. Fortunately, there is a nice tool for checking this which doesn’t even require and installer. I used the LatencyMon and let it run for 10 minutes to graph the results.

I thought it necessary to look at some systems which I wouldn’t have suspected of having any DPC latency issues around my house to get some baseline numbers for comparison. The utility graphs out the data nicely and tells you what your latency results mean in terms of the real world problems you might encounter with the current system configuration. I went with my own personal machine which uses the ASUS Rampage V Extreme’s onboard audio and Windows 10.

ASUS Rampage V Extreme

Article Image

Our baseline system has an absolute maximum interrupt to process latency of 225µs and the highest reported DPC routine execution time was 262μs. This means that you shouldn’t experience any drop out issues with audio or video on this system.

GIGABYTE X99A Gaming Pro Carbon

Article Image

In this test, the maximum interrupt to process latency was 56.31μs. That’s much less than that of our control system. The highest reported DPC routine execution time was a mere 247μs. Again this compares favorably to our control system.

Drive Performance

To test the capabilities of the on board USB 2.0 connections, we used an ACOMDATA HD060U2FE-72-USB 2.0/FireWire HDD connected first to the USB port. USB 3.0 functionality was tested using a Super Talent USB 3.0 SuperCrypt 32GB Flash Drive. SATA drive tests were performed using Western Digital Caviar Black WD1002FAEX hard drives on all SATA headers. The SATA drives were used for testing in RAID 0 16k block size configurations on all applicable controllers. Testing was also conducted using the same model SATA drives in a stand-alone SATA configuration on all applicable controllers. All drive benchmarks were done using the freely available CrystalDiskMark program, run with both 50MB and 100MB sized test sets.

Given that motherboards are now supporting UASP and various USB 3.0 boost methods on many models, we’ve updated our testing methodology to include a UASP test. The USB 3.0 implementation that some manufacturers are using does allow for a performance boost with non-UASP compliant hardware as you’ll see using what these companies call "turbo" mode. Granted the difference isn’t as pronounced as it is when enabling UASP on a device that supports it. The USB 3.0 Flash drive tests are essentially the same as these have always been since we started doing these tests, but with the added turbo mode test to showcase the feature in action. The USB 3.0 SSD UASP Enabled / Disabled tests are utilizing a Corsair Force GT 60GB SATA 3 SSD plugged into a Thermaltake BlacX 5G docking port which uses a USB 3.0 connection. This device was selected due to having UASP compatible firmware.

Storage Configuration

The storage configuration of the X99A Gaming Pro Carbon offers everything the X99 chipset has and then some. There are 10x SATA 6Gb/s ports, 1x SATA Express port, 1x NVMe compatible PCIe x4 M.2 port, and a U.2 port. With regard to USB support, there are 8x USB 2.0 ports provided by the chipset. The rest of the USB ports are all USB 3.1 compliant according to MSI’s specifications. For USB 3.1 there is an ASMedia 1142 controller and a VIA VL805 chipset in addition to the X99’s built in USB controller. The U.2 port shares bandwidth with the M.2 slot which makes perfect sense. I am not a fan of the vertical SATA Express port, and again MSI could have slide the ATX power and USB 3.0 front panel header to the right and given themselves a bit more room to make the SATA port arrangement more appealing.

Another interesting twist is that MSI has finally decided to do what ASUS and ASRock have been doing for some time and provided a utility for increasing USB 3.x speeds by going beyond the ancient BOT protocol. MSI claims a 30% performance increase which is a bold claim and unfortunately not true as you can see below. This generation we’ve seen less of an increase in enabling UASP and various turbo modes than we’ve seen in the past with earlier USB 3.0 controllers. Current controllers are testing roughly as high as the older ones did with UASP enabled. I’m seeing a 5-10% increase in performance at best on other motherboards.

50MB Test Set

Article Image

Article Image

In the 50MB sequential read test the AHCI controller was the fastest configuration tested. As I mentioned before we got very small gains enabling MSI’s "Fast" USB mode. The option was available on all the controllers, although the Intel controller gained the most from the fast mode, the ASMedia was far faster overall. USB 2.0 performance was precisely what I expected. In the 50MB sequential write test the AHCI standalone controller achieved a result of 128MB/s. This is roughly half what was achieved by the RAID 0 configuration which is again what we normally see in this test. The USB performance was slightly above what we normally see, but only by a very small margin.

100MB Test Set

Article Image

Article Image

In the 100MB sequential read test the RAID 0 controller configuration was the fastest tested. The USB performance was again very close in normal and fast modes. In the 100MB sequential write test we saw nearly double the performance of the AHCI standalone controller in RAID mode once again and consistent USB 3.1 performance. USB 2.0 performance was again precisely as we expected it to be.

Network Utilization Tests

LAN Speed Test software was used with Windows Task Manager to determine the performance levels of the onboard network interface. LAN Speed Test was used to measure bandwidth and transfer speeds, while Windows Task Manager monitored CPU utilization on the test system. For the testing, a 800MB file test was used with the default packet configuration for the application. The test was run three times with the middle result chosen. Results were captured for the low, medium and high transfer rates. The test was performed using a plenum rated category 5e crossover cable to bypass any traffic, routing or other transfer issues and possible packet loss or corruption that can be caused by a router/switch or hub. The cables were connected between two test machines, one using the onboard NIC(s) of the board being reviewed and the other is an Intel EXPI9400PT 10/100/1000Mbps PCI-Express Gigabit Ethernet adapter installed into a test machine using an Intel D5400XS motherboard.

Wireless network testing, if applicable was performed using an AdHoc connection between the review system and my own machine based on equipped with an Atheros AR9002WB-1NG wireless adapter which is integrated into my ASUS Rampage III Black Edition motherboard. The control panel used all the default settings save for the AdHoc 11N setting which was enabled on both systems.

There is a single PCI-Express 2.0 based Intel i218v Gigabit Ethernet controller integrated into the x99A Gaming Pro Carbon. It is capable of 10/100/1000Mbit speeds.

LAN1 (Intel i218v)

Article Image Article Image Article Image Article Image Article Image

In the write test, the average throughput of this controller was 42.58MB/s. In the read or download test, the throughput was naturally higher at 61.59MB/s. In the write or upload test, our maximum transfer rate was 53.78MB/s. In the download or read test our maximum transfer rate was slightly higher than our average at 62.36MB/s. Our minimum transfer speed in the write or upload test was 39.59MB/s. In contrast the read or download test showed a minimum throughput of 60.95MB/s which isn’t very far off our average or maximum transfer rate numbers. CPU usage was higher than I expected at 7% in the write test. The usage graph is fairly consistent with few spikes during the duration of the test. Observable spikes were almost too small to see. We saw the same 7% utilization in the read or download test. In this case however the usage was more consistent and the graph level during the test.

Benchmark Test Systems

The following system configurations were used for the Sandra memory benchmark graph, as well as all graphs listed under the Application and Gaming Benchmarks sections:

Article Image

SiSoft Sandra

Article Image

Note that all results above were obtained running the installed memory in dual-channel or quad-channel memory modes where applicable.

Note: We were unable to run the clocks of the 6950X at the same 4.5GHz clock speeds we normally run for everything else. As a result, we’ve got numbers of the 5960X at 4.5GHz and 4.3GHz for comparison. Also given that this is the first Broadwell-E based set of benchmarks we’ve run, we don’t have anything to directly compare it to aside from the older 5960X at both 4.5GHz and 4.3GHz.

The X99A Gaming Pro Carbon did a little better with the 5960X than it did the 6950X.

Article Image

In the CPU test, the 6950X was quite a bit different exceeding the performance of the 5960X by a wide margin due to the extra cores on the new Broadwell-E processor.

Hyper Pi

Article Image

In the Hiper Pi test we utilize 4 cores only, the 6950X shows to be a tad bit slower in terms of IPC than the previous Haswell-E processor.

wPrime

Article Image

In this single threaded test the 6950X once again outshines everything but the Skylake based systems but keep in mind those to have a 200MHz higher clock.