ASRock X79 Extreme9 LGA 2011 Motherboard Review

Our experiences around here with ASRock haven’t been the best. Though the last one we tested actually did really well. As a result of our great experiences with the Extreme4 we can only hope the ASRock X79 Extreme9 follows suit. The Extreme9 is ASRock's most expensive Sandy Bridge E motherboard, we hope it acts like it.

continued...

Subsystem Testing

NOTE: For all Subsystem Testing, an Intel Core i7 3930K (3.20GHz) and 4 x 4GB Corsair Vengeance (1600MHz DDR3 9-9-9-24-1T @1.50v) memory modules running at DDR3 1600MHz were used. The CPU was cooled with a home brew water cooling setup consisting of a Swiftech MCR320 triple radiator, 3x120mm fans, Swiftech MCP655-B pump, and a Koolance CPU-370 water block with LGA2011 mounting hardware.

Sound Hardware

With regard to audio, there usually isn’t much to talk about. However rather than integrating the normal Realtek ALC889 or ALC892, the ASRock X79 Extreme9 uses the Creative Labs "Core3D" quad core audio processor. It’s been combined with one of the board’s Broadcom NICs to create the ASRock "Game Blaster" card. This is a lot like what ASUS did with the Rampage III Black Edition and its combination Killer NIC E2100 and Xonar audio. Unlike that solution, the Game Blaster may have a nice processor, but it lacks the upgradable OPAMPs the ASUS setup had. Audio-wise it sounds good, but compared to that solution (which is now in my personal machine which I use for gaming) the Game Blaster sounds flat. In all honestly it sounds pretty much like the Realtek ALC892’s do. Normally I don’t test game audio in motherboard reviews instead opting for music, but in this instance I actually did play Star Wars The Old Republic on the test machine and found that the audio was sub-par compared to the Xonar bundled with the ASUS R3BE. Audio is of course a subjective thing so your mileage may vary.

The specifications for the Game Blaster were taken directly from ASRock’s website and are listed below.

- Supported by the bundled ASRock Game Blaster

- 7.1 CH HD Audio

- Creative Sound Core3D quad-core sound and voice processor

- Supports THX TruStudioآ™ PRO

- Supports CrystalVoice

- Supports EAX1.0 to EAX5.0

There isn’t much to say. The feature set is a bit more robust than what is typically offered on the Realtek ALC892, but the Realtek ALC892 actually supports a lot of features, but these features have to be licensed individually which is why the specifications vary wildly from one motherboard to another.

Audio آ– 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 as predicted.

Audio آ– 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.

The recording sample was generally free of distortion, and was more audible than the typical Realtek ALC892 is. So in this instance, the Game Blaster shines a little bit.

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 3.0 program, run with both 50MB and 100MB sized test sets.

There is actually quite a bit to talk about with regard to storage and the X79 Extreme9. As we’ve mentioned before, X79 like Z68 and P67 before that only natively supports two devices at SATA 6Gb/s speeds. The other four ports are only capable of SATA 3Gb/s speeds. As a result the X79 Extreme9 relies on third party controllers as does every other board out there. What’s interesting is how this board differs from the earlier reviewed ASRock X79 Extreme4. The Extreme9 actually uses a Texas Instruments USB 3.0 controller which supports the newer UAS protocols for increased speed (This is where ASRock’s marketing department came up with the "X-Fast USB 3.0" terminology.) and the Extreme4 uses the ASMedia 1042 controller. Both support UASP but the latter is a wholly owned subsidiary of ASUS. I find it odd that two boards in the same series source entirely different USB 3.0 controllers. Even more strange is that the Extreme9 takes a step backwards from the Extreme4 as the Extreme4 uses the newer and faster ASM1061 controller rather than the older Marvell 9172. Actually the ASRock Extreme9 has three added RAID controllers. Two of which are the Marvell 9172’s and the other is the Marvell 9220. The latter controller is actually among the fastest, if not the fastest third party controllers we’ve seen integrated into a motherboard. It may be this controller that lead ASRock to an all Marvell solution.

I suppose the Extreme4 may have been designed later, but it’s interesting to note that the higher end "flagship" board isn’t necessarily better in every respect. One would generally expect it to be. Another difference between the Extreme9 and the aforementioned ASRock X79 Extreme4 is that the Extreme4’s third party drive controller supported RAID modes technically, but RAID wasn’t enabled in BIOS. The reason for this is likely due to one of two reasons. Either it’s a marketing thing in which case if you want more RAID controllers, you need to move to a more expensive board, or it’s because there is insufficient option ROM space to run both controllers simultaneously. It’s difficult to know what ASRock’s reasoning is. The latter issue isn’t clear cut as you can work around this problem to a degree.

With regard to the ASRock X79 Extreme9, ASRock has in fact solved the problem with limited OROM space somewhat. This is the first board I’ve been able to get Intel and Marvell RAID controllers to both run in RAID mode at the same time, albeit with some caveats. ASRock provided a setting for priority between UEFI firmware and legacy ROM. Some other boards do too, but I’ve not had much success working with those. The difference lies in the fact that ASRock gave you the choice to make the third party controllers bootable or not. (Well not the 9220, which is not bootable in any case I can come up with.) With legacy firmware having priority, you have problems with more than one controller in RAID mode. With UEFI firmware having priority, the Intel controller will operate normally. It remains bootable. While enabled, you can in fact use the other controllers in RAID mode. All three can even be enabled at once, and that alone is sort of a big deal considering how many others fail here. However, those secondary RAID controllers can’t be configured as bootable while the Intel controller is in RAID mode. So that limitation is still somewhat in effect. The system will not load an OS like this. As a result you can’t configure an array on any Marvell controller outside an operating system. You have to load Windows on a drive or array attached to the Intel chipset and then configure the array with Marvell’s Storage Tool. From here you can configure your drive arrays normally. All 6 internal ports can be configured as RAID capable and both eSATA ports can be as well.

Confused yet? Well keep reading, you soon will be. The Marvell 9220 is a special case unto itself. There are no provisions in the UEFI to configure this as a bootable drive controller. As a result one may consider the specifications listed in the manual as misleading. You can run a drive array on this controller, but you can only add an array to and configure this array in the OS environment through the Marvell Storage Utility. In that regard it’s just like the 9172 controllers. The difference is the 9172’s can be used as bootable controllers outside the OS provided that the Intel controller is in legacy IDE mode or AHCI mode only.

I give props to ASRock for making all the controllers actually useful, but the conditions for use aren’t specifically spelled out and took some time to decipher. UEFI lets us work around some of the OROM limitations but until better firmware is available for the third party controllers, we’ll always have issues like this. Unfortunately, while I’m giving ASRock props, there is yet another issue. During our testing SATA ports 0 and 1 failed. These are the SATA 3Gb/s ports attached to the Intel chipset itself. They didn’t die outright, but there are extreme performance problems with these resulting in a lack of responsiveness as well as I/O timeouts and errors. I was able to work around these issues, but it’s worth noting from a QC perspective as this isn’t the first hardware failure we’ve seen from ASRock.

50MB Test Set

Article Image

Article Image

The 50MB sequential read test normally shows dominance by the Intel controller in virtually every category. That’s not what we saw here. The Marvell 9220 is actually the fastest controller rested here. USB 2.0 performance is excellent and USB 3.0 performance is the best I’ve seen to date. The other controllers all scored where they should have. Moving onto the sequential write test, we see the Intel controller resume its customary lead over the competition. As usual we see nearly perfect scaling with RAID 0. This graph also illustrates the massive improvement the Marvell 9220 has over the 9172. USB 2.0 performance is looking really good here but the USB 3.0 performance tanked badly. This is probably the lowest read score I’ve ever seen. You’ll see more of this in the subsequent graphs, but I’ll just go out on a limb and say while its reads are great, the write performance of the TI controller leave a lot to be desired. Everything else again ends up where it should.

100MB Test Set

Article Image

Article Image

In the 100MB sequential read tests, the Marvell controllers absolutely dominate the field. Normally the Intel controllers score a bit better here in regard to RAID 0 performance. I’m uncertain as to why their performance is a bit lack luster here. But again the Marvell 9220 looks really good here and the 9172 does better than usual. As we saw in the earlier test, the USB 3.0 controller shows us excellent read speeds. Everything else falls within expectations. The 100MB write tests basically follow the 50MB tests. Everything scales about the same with the Intel’s dominating all but the Marvell 9220. The 9172 comes up short as I’d normally expect. USB 3.0 performance tanks again when it comes to writes. For whatever reason the IEEE1394 controller also dives. I’m unsure why, but I’m not sure that anyone cares about that.

Network Utilization Tests

Hagel Technologies’ DU Meter software was used with Windows Task Manager to determine the performance levels of the onboard network interface. DU Meter was used to measure bandwidth and transfer speeds, while Windows Task Manager monitored CPU utilization on the test system. For the testing, a 750MB Archive file consisting of several compressed WMA/MP3 files was used for the large file transfer, and 750MB worth of MP3/WMA files were used ranging in sizes from 3 to 30MB was used for the small files transfer test. 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 . The control panel used all the default settings save for the AdHoc 11N setting which was enabled on both systems.

ASRock once again turned to Broadcom for its Gigabit Ethernet controllers. It integrated a single Broadcom BCM57781 on the X79 Extreme9. This same controller is integrated into the Game Blaster card also bundled with the board. Both are capable of 10/100/1000mbit speeds as well as adapter teaming. LAN 1 is the integrated BCM57781 while LAN 2 is the Game Blaster card’s BCM57781.

LAN1

Article Image Article Image Article Image Article Image

The small files download test showed an average transfer rate of 79.0MB/s and a maximum of 97.5MB/s. CPU usage was higher than I’d like to see but par for the course with the Broadcom NICs at 4%. In our small files upload test we see a large disparity between average and maximum transfer rates. Transferring an uninspiring average of 55.5MB/s and 111MB/s at its peak. CPU usage remained almost unchanged at 3%.

Article Image Article Image Article Image Article Image

The large files download test showed an average transfer rate of 91.8MB/s and a maximum of 101MB/s. CPU usage was slightly higher than I expected reaching 5%. The large files upload test averaged 88.0MB/s with a maximum of 117MB/s. CPU usage again reached 4%.

LAN2

Article Image Article Image Article Image Article Image

The small files download test for the second NIC reached an average of 87.8MB/s. The maximum transfer rate came in at 98.8MB/s. CPU usage was once again 4%. As with the first NIC, we saw a massive drop in the average speed of the small files upload test despite the high maximum speed. The transfer rates were 52MB/s and 110MB/s respectively. CPU usage did fall slightly to 3% just as it did with the test on the first NIC.

Article Image Article Image Article Image Article Image

The average transfer rate of the large files download test was 97.5MB/s with a maximum of 102MB/s. CPU usage climbed compared to most of the test and hit 5%. The large files upload test showed some of the best results though with an average of 89.4MB/s and a maximum of 122MB/s. CPU usage again remained a steady 4% as it did with most of the tests.

All in all the Broadcom NICs get the job done. I really haven’t anything to complain about.

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

Graphs are labeled as follows: Motherboard - CPU Clock - Memory Clock

SiSoft Sandra 2011

Article Image

Note that all results above were obtained running the installed memory in dual channel mode with two exceptions The ASUS Rampage III Black Edition test was run in triple-channel mode and the ASRock X79 Extreme9 was run in quad-channel mode.

Any X79 board is difficult to touch in regard to bandwidth as you can see by the graph. That being said, the Gigabyte board outshines our ASRock sample by a hair.

Article Image

And in this test the ASRock turns the tables on the Gigabyte board.

Hiper Pi

Article Image

Here is that reversal again.

wPrime

Article Image

And againآ…آ…..