We have tinkered around with a few of the ASUS workstation motherboards here at TweakTown. These have been boards that arguably could be setup as a desktop and run to get a very good combination of features, performance, and stability (with some overclocking sprinkled on top).
Now, while these have been lots of fun and have shown that the WS team can certainly hold their own against even the ROG desktop team, we have not really gotten deep into the ASUS WS offerings. Today that all gets to change as we tear the box off of a Dual Socket 1366 board from ASUS. The ASUS Z8NA-D6C is something; you get two Nehalem CPUs and six slots of RAM all stuffed into an ATX package.
You also get an ASPEED VGA controller with 8MB of its own memory (for server usage), two Intel 8475L GBe LAN chips, and much more for around $300 at NewEgg.com. Shall we see what it can do as a desktop/workstation?
The Box and What's Inside
Package and Contents
I like the clean and very exotic look of some of the WS boards. This one in particular has a great look to it. Black and glossy like a sports car with a graphic on the front to drive home that feeling. You get very little in the way of a hard-sell on the front of the box.
That does not mean it is not there; it is all hidden under the front flap. In an odd departure, the "features listing" is in a very large and "shouting" font. When I opened this up I felt like it was yelling at me. Once you get past the volume, this list does show off some of the reasons to pick up the Z8NA-D6C.
The back of the box breaks down the two types of Z8NA-D6 Boards you can get. We have the D6C, while the D6 is a much more powerful product with remote management and an LSI SAS RAID controller built in.
After viewing all of that, you find that the top is just a slip cover with an almost identical box underneath (without the front flap).
The D6C does not come with much in the way of accessories; just a few SATA Cables, the manual, driver DVD-ROM and an I/O plate.
To stuff everything needed for a full two 1366 sockets (complete with memory banks for each) and the required chipset into an ATX setup, ASUS had to make some interesting and difficult design choices. Some of these are very visible others are not so obvious.
One of the most visible choices is the socket layout and placement. The two 1366 sockets have been placed "head-to-head". The reasons for this are fairly simple. By placing them this way they can route the traces in an almost mirrored fashion. If you look at the board you can see those traces and how they extend in along a roughly 45 degree angle but in opposite directions.
You can also see something that I have talked about before. Notice how many of the traces are not straight, but appear to be squiggly? That is called trace tuning. It is when you need to make sure signals reach a particular point at the right time. By zig-zagging them you can slow down the speed that the signal reaches its destination.
Looking at another design choice that is growing more common in server boards, we see both the 24-pin and the 8-pin aux power connector at the top of the board.
This move does a couple of things; it shortens the signal paths to the power regulation for the two sockets and helps to keep cables out of the way of air flow.
Moving to the lower half of the board we find an odd assortment of slots. You have a single x16 physical slot that actually happens to be x16 electrical. The two x8 slots are not so forthcoming. They are both only x4 electrical. That sneaky little x1 slot is not that at all. It is a MIMO slot for the optional audio card. We did not get this card in our package so we ended up using the single PCI slot for our Xonar D2.
Here we have that little ASPEED GPU and its associated 8MB of memory. Above this we can see a jumper. That jumper allows you to completely turn this GPU off. For our testing this was left off as the ASPEED is really only meant to give you basic display functions and while handy for servers is of little use for most workstations.
What's missing from this picture? Well it is the SAS ports that would come from the LSI SAS RAID controller. That controller is not present on the D6C so instead we get some blank space here. Let's take a look at the ports that are visible. There are six in view, four are red and two are black. That actually means nothing. They all are controlled by the Intel ICH10R. The two black ones just let you know they are ports five and six; this can help you quickly ID which port is one when mounted in a case and the board level lettering is not visible.
More missing here; there is room and points for a floppy port, but it is not there. We do find a single board mounted USB port. This is good for software dongles like the kind you need for Lightwave 3D.
The ports are light, but they normally would be on a server board. You get two PS/2 ports two USB 2.0 ports, a legacy serial port, the VGA port and two GB LAN ports.
BIOS and Overclocking
The BIOS on the Z8NA-D6C is not like other BIOSes from ASUS (at least not the enthusiast BIOSes). It is meant for a server or workstation. These are systems that you want for stability more than clock speed. You are not going to find overclocking tools in here, but you are going to find the tools you need to get the Z8NA-D6C running at its best.
When you first enter the BIOS you are going to see the familiar AMI layout and a main page that looks like many other AMI based systems.
Even looking at the top portion of the Advanced CPU settings page there is not much that's different, at least not yet.
Once we move our way down to the lower half of the page we find all of the nice tweaks and extras that go with a real server or workstation board (and especially dual CPU boards). Two controls that are probably not familiar on this page are the DCU and DCA. These control how the cache on the CPUs operate. The DCU controls the mode of the cache (ECC or No ECC) and the DCA controls the amount of time the system delays (in bus clocks) from snoop to prefetch.
The Northbridge page also has some new items. You can tell the Z8NA to throttle the memory if it gets too hot, adjust the ECC mode, change the NUMA aware policy, and much more.
Under the server page you can setup the terminal access through that legacy serial port we showed you.
ASUS also gives you control over the USB ports and headers available on the Z8NA-D6C. Of course what ASUS motherboard would be complete without some nice fan controls? For our testing we left this set to generic mode.
There were no overclocking options in the BIOS and there was no utility provided (and there were none to be found online). Maybe later we can find a board (or ASUS will add a special BIOS) that will let us overclock a pair of these Xeons. For now, we will only present stock scores.
As all overclocking results are dependent on the hardware you use, your results may vary. Results of our overclocking tests are included in the performance section with the stock scores.
Important Editor Note: Our maximum overclocking result is the best result we managed in our limited time of testing the motherboard. Due to time constraints we weren't able to tweak the motherboard to the absolute maximum and find the highest possible FSB, as this could take days to find properly. We do however spend at least a few hours overclocking every motherboard to try and find the highest possible overclock in that time frame. You may or may not be able to overclock higher if you spend more time tweaking or as new BIOS updates are released. "Burn-in" time might also come into play if you believe in that.
Test System Setup and Comments
As you can imagine getting everything installed on the Z8NA-D6C was tricky. It was not exactly hard, but you did have to install things in the right order to avoid complications.
After all the hardware was hooked up we installed Windows 7 x64 Ultimate. This went very smoothly until we tried to install the drivers from the DVD. At this point all we saw were errors telling us that the installers did not support our OS.
Talk about frustrating; we ended up either using the drivers provided by Windows or grabbing a couple off of the Intel and ASUS websites. As we also mentioned, there was no audio codec included. During the setup of several of our regular tests we were given errors about this. We ended up having to use our PCI ASUS Xonar D2 to get around this small issue. Once we passed these two hurdles were we good to go.
Synthetic Tests - Part I
With any system you will want to see a combination of synthetic and real-world testing. Synthetics give you a static, easily repeatable testing method that can be compared across multiple platforms. For our synthetic tests we use Everest Ultimate, Sisoft Sandra, Futuremark's 3DMark Vantage and PCMark Vantage, Cinebench as well as HyperPi. Each of these covers a different aspect of performance or a different angle of a certain type of performance.
Memory is a big part of current system performance. In most systems, slow or flakey memory performance will impact almost every type of application you run. To test memory we use a combination of Sisoft Sandra, Everest and HyperPi 0.99.
Version and / or Patch Used: 2010c 1626
Developer Homepage: http://www.sisoftware.net
Product Homepage: http://www.sisoftware.net
Buy It Here
As you might have expected the memory bandwidth on the Z8NA-D6C is quite good. This is a must to keep all 16 threads loaded up with data.
Version and / or Patch Used: 5.30.1983
Developer Homepage: http://www.lavalys.com
Product Homepage: http://www.lavalys.com
Buy It Here
Everest Ultimate is a suite of tests and utilities that can be used for system diagnostics and testing. For our purposes here we use their memory bandwidth test and see what the theoretical performance is.
Stock Memory Performance
Everest gives us some more information here; we see that the caching on the Xeons is much faster than your average Core i7.
Version and / or Patch Used: 0.99
Developer Homepage: www.virgilioborges.com.br
Product Homepage: www.virgilioborges.com.br
Download It Here
HyperPi is a front end for SuperPi that allows for multiple concurrent instances of SuperPi to be run on each core recognized by the system. It is very dependent on CPU to memory to HDD speed. The faster the components, the faster it is able to figure out the number Pi to the selected length.
For our testing we use the 32M run. This means that each of the four physical and four logical cores for the i7 and the four physical cores of the i5 is trying to calculate the number Pi out to 32 million decimal places. Each "run" is a comparative to ensure accuracy and any stability or performance issues in the loop mentioned above will cause errors in calculation.
On the surface it looks like the Z8NA-D6C is slow (and it is in comparison). What we have to keep in mind is that there are 16 instances of HyperPi running; each trying to calculate the number Pi out to 32 million places. This is going to back up the memory on the system as well as the CPUs. Since we have 12GB of RAM that ends up breaking down to 1.5GB of RAM per CPU core or .75 GB per thread, this could bring down a few other scores as the memory has a hard time getting information to the CPUs.
Synthetic Tests - Part II
Disk Drive Controller
The system drive controller is an important part of system performance. In most modern boards your drive controller will run off of the PCI-e bus. The PCI-e bus performance can be affected by poor trace layout as well as many other design choices that show up on different boards.
For testing we use Sisoft's Sandra and Everest.
The drive performance here is bad, but could be due to the drivers we ended up having to use.
Stock HDD Performance
Again, Everest shows us more details than Sandra but still indicates that the Z8NA-D6C could have some drive performance issues.
Synthetic Tests - Part III
Overall System performance and Gaming
Here is where we dig out the FutureMark tests.
Version and / or Patch Used: 220.127.116.11
Developer Homepage: http://www.futuremark.com/
Product Homepage: www.futuremark.com
Buy It Here
For overall system performance we use PCMark Vantage. This is run in both x86 and x64 mode to give the best indication of performance.
Here we see that PCMark does not have the greatest support for dual CPU systems. Even with the extra 4 threads we are seeing very low scores.
Version and / or Patch Used: 1.0.1
Developer Homepage: http://www.futuremark.com/
Product Homepage: www.futuremark.com
Buy It Here
For synthetic gaming tests we used 3DMark Vantage, the industry standard and overlockers bragging tool. This is a test that strives to mimic the impact modern games have on a system. Futuremark went a long way to change from the early days of graphics driven tests to a broader approach including physics, AI and more advanced graphics simulations. 3DMark Vantage uses the DX10 API in addition to having support for PhysX. As we are no longer using an NVIDIA GPU for testing (at least until we can get a GTX 4xx card) you will only see the CPU based PhysX results in the scores. For testing we use the Performance test run.
Here we see a good score despite the slower CPU speed. We are seeing a pair of 2.4GHz CPUs performing just behind the 3.3GHz and higher CPUs.
Cinebench R11. x64
Version and / or Patch Used: R11.5 x64
Developer Homepage: http://www.maxon.net/
Product Homepage: www.maxon.net
Download It Here
Cinebench is a synthetic rendering tool developed by Maxon. Maxon is the same company that developed Cinema4D, another industry leading 3D Animation application. Cinebench R11.5 tests your systems ability to render across a single and multiple CPU cores. It also tests your systems ability to process OpenGL information.
Ouch is all I can say here. It is possible that the CPU Speed and lack of RAM are causing the performance we are seeing, as the scores are not that far off.
Real-World Tests - Part I
Real-world testing allows us to see how well a product will perform when used in the same manner as it would be in your house or office. It is an important side to performance testing as it can uncover hidden glitches in the way a product performs.
This is especially true when testing a mainboard; there are so many components of a board that have to interact that any problems between parts can cause a failure of the whole.
For real-world testing we use some common applications and functions. We test with LightWave 3D for rendering performance, AutoGK for transcoding from DVD to AVI, and two games for gaming testing.
Rendering of 3D Animation is a system intensive endeavor. You need a good CPU, memory and HDD speed to get good rendering times. For our testing we use LightWave 3D. This software from Newtek is an industry standard and has several pre-loaded scenes for us to use.
Version and / or Patch Used: 9.6
Developer Homepage: http://www.newtek.com
Product Homepage: http://www.newtek.com/lightwave/
Buy It Here
Lightwave 3D scores are pretty good in reality. What you are seeing here is a 2.4GHz CPU (well eight of them) running almost as fast as a 3.33GHz CPU. The extra two cores and 4 threads come into play here while the lack of memory per CPU and the slower HDD performance brings us back down.
Version and / or Patch Used: 2.55
Developer Homepage: http://www.autogk.me.uk/
Product Homepage: http://www.autogk.me.uk/
Download It Here
AutoGK stands for Auto Gordian Knot; it is a suite of transcoding tools that are compiled into an easy to install and use utility. It allows you to transcode non-protected DVDs and other media to Xvid or Divx format. For our testing purposes we use a non-DRM restricted movie that is roughly 2 hours in length. This is transcoded to a single Xvid AVI at 100% quality.
Our numbers here represent a few things. The first is the slower HDD speed, as the DVD has to move data through the same controller it is likely that this reduced the overall performance. We also have a feeling that AutoGK, although multi-threaded, cannot deal with the extra CPU. It would seem that we are only seeing the results from one 2.4GHz CPU.
Real-World Tests Part II
Here we have our real world gaming tests. Each of the games we chose uses multiple cores and GPUs. They are able to stress the system through use of good AI. Both have decent positional audio that adds impact to the sound subsystem of the board. We ran each game through the level or parts listed and recorded frame per second using FRAPS. This brings the whole game into play.
*** A word on gaming as a motherboard test; ***
Despite the fact that most games are very GPU limited, we are still noticing HDD and even audio creating issues in gaming performance. Because of this you may see differences in the number of frames rendered per second between different boards. Usually the difference is very small but occasionally, because of bad tracing, poor memory or HDD performance this difference is significant. The issues are often more prevalent in older versions of DirectX but can still pop up in DX10 and 11.
Call of Duty Modern Warfare 2 (DX9)
Version and / or Patch Used: 1.0
Timedemo or Level Used: First combat until the school is cleared
Developer Homepage: http://www.infinityward.com
Product Homepage: http://modernwarfare2.infinityward.com
Most of you know about the game Modern Warfare 2; it caused quite a bit of controversy in the latter half of 2009. The game is a first person shooter with a heavy combat emphasis. It follows the events in the first Modern Warfare very closely and brings back several characters from the original.
As with most games in the Call of Duty franchise, it features a heavy AI load. This is not because of a complex AI routine, but more due to the sheer number of enemies in any given combat situation. It is also our single DX9 based game in our testing suite. Settings are shown below.
Despite being a server/workstation board the Z8NA-D6C does pretty well. The dual CPUs do not hinder the gaming side of our testing so far.
Far Cry 2 (DX10)
Version and / or Patch Used: V1.00
Timedemo or Level Used: Clearing the Safe house through to the Rescue
Developer Homepage: http://www.ubi.com
Product Homepage: http://farcry.us.ubi.com
Far Cry 2 is a large sandbox style game. There are no levels here so as you move about the island you are on you do not have to wait for the "loading" sign to go away. It is mission driven so each mission is what you would normally think of as the next "Level". In the game you take the role of a mercenary who has been sent to kill the Jackal; unfortunately your malaria kicks in and you end up being found by him. Long story short you become the errand boy for a local militia leader and run all over the island doing his bidding. Settings we used for testing are shown below.
Again, the Z8NA-D6C does well; it is not up at the top but you certainly could use it for Far Cry 2 gaming.
Battlefield Bad Company 2 (DX11)
Version and / or Patch Used: V1.00
Timedemo or Level Used: From washing up on the beach to the mine fields.
Developer Homepage: http://www.ea.com/
Product Homepage: http://badcompany2.ea.com/
Battlefield Bad Company is another sequel and also another game "franchise." Bad Company 2 is also our DX11 Shooter game. The game follows a fictitious B company team on a mission to recover a Japanese defector. This puts you back in World War II (at least for the beginning of the game) while the multi-player game is centered on much more modern combat. For our testing we used the single player mode. Settings are shown below.
Once more, the Z8NA-D6C passes our gaming test with Bad Company 2. This time it is only one frame per second behind the leaders.
Although certainly not designed for it, we found that the Z8NA-D6C could game quite well. It opens up a few interesting options. You can setup the system with the right GPU to give you good workstation performance and still have power left over to play a few games after you are done for the day.
Power Usage and Heat Tests
We are now able to find out what kind of power is being used by our test system and the associated graphics cards installed. Keep in mind; it tests the complete system (minus LCD monitor, which is plugged directly into an AC wall socket).
Well look at that; one extra CPU and the power draw is quite acceptable. Of course this is due to the low power Xeons we are using.
As a new measure, we are now monitoring the heat generation from the key components on the motherboard; this being the Northbridge, Southbridge (if it contains one) as well as the Mosfets around the CPU. The results are recorded at idle and load during the power consumption tests.
Heat generation is another story. The 5500 series chipset gets a tad warm.
The ASUS Z8NA-D6C is a very interesting motherboard. It has some great potential, but I think for many graphics professionals the D6C will not be the choice; they will go for the D6 with its built in LSI raid controller or an add-in board.
However, as with most dual CPU (and multi-core) systems, you will run into applications that do not know how to use the second CPU or to get all of the cores available to work properly. We were impressed with the way the board handled games as well as the lower power usage (that was a very pleasant surprise). Although we did not touch on the way it would handle as a Linux or Windows server, we can see where the support is there for this role. Features like NUMA support, caching profiles, and remote management all combine to make this a very good product for the server/workstation market. Plus, when you consider the price tag of only $300.00 at NewEgg.com you are not getting a bad deal at all.
We will try to revisit the Z8NA-D6C with a faster pair of CPUs and more RAM in the near future (as well as a couple of additional workstation related applications), but for now we can tell you that the Z8NA-D6C is a very capable board for either a workstation or server that needs dual CPUs in a small package.