The Test Rig
Once I was done messing around with some of the motherboard features, it was time to get down to business and perform some benchmarks. To test the M3A78-T, we built a brand new test rig around it. This test rig represents AMD’s new Dragon platform, and the parts roll call looks something like this.
- AMD Phenom II 720 2.8GHz Triple Core CPU
- OCZ 4GB PC2-8500 Gold Edition Dual Channel DDR2 Kit
- GIGABYTE Radeon HD 4850 512MB GDDR3 Video Card
- Western Digital VelociRaptor 300GB 10000rpm SATA2 Hard Drive
- HighSpeed PC Top Deck Tech Station
- Lian Li MAXIMA Force Extreme PS-A650GB 650W Power Supply
- Thermaltake SpinQ Heatpipe CPU Cooler
We tried to cover all the bases when testing the M3A78-T. There’s the standard appearance made by Futuremark’s benchmarking programs, as well as SiSoftware’s Sandra suite. We also performed a couple real world game tests, and even tested the clarity of the audio to make sure you are getting the full experience. But enough with the jibber-jabber, it’s time to see some results.
PCMark Vantage
For our first benchmark we turned to Futuremark and their PCMark Vantage benchmark suite. Though this suite is a synthetic test, it tries really hard to perform real world tests using applications and function similar to what you would do on any desktop system. There is a minor 3D component, which can be affected by the capability of your video card. With that in mind, we performed our benchmarks using both the onboard video found on the M3A78-T and our test video card. We picked three specific tests out of the myriad of different scenarios you can choose from.
Composite is a simple run through of most basic system functions, and is similar to the PCMark suites of the past. TV and Movies tests video playback, encoding, and rendering. Finally Productivity is…well…productivity. It tests things most office drones do all day, like surf the internet and play with Microsoft Word.
The results were favourable, with the difference in video card not having much of an impact on overall performance. As for the scores themselves, these are well within expectations for the configuration of our test rig. We would most likely see a results increase by upgrading the processor, as that seems to be the swing vote in terms of components.
SiSoft Sandra
Our set of tests were conducted using SiSoftware’s Sandra testing and diagnostics suite. This suite can be used for diagnosing and repairing ailing computer components, but it also comes with a built-in set of benchmarks. These are purely synthetic, and test the raw performance of component in the most straight-forward of fashions. Selecting a motherboard-specific test was difficult, as many of the tests rely on processing speed or the performance of other sub-systems. We settled on a simple memory bandwidth test.
In that realm of performance, the 790GX chipset found on the M3A78-T plays second fiddle to AMD’s own Xpress 1600. However performance is on par with the nVidia 680i SLI, and far and above the older Intel 955X. These are the four fastest chipsets found in the Sandra database, so I would say the 790GX chipset and the M3A79-T motherboard housing it faired pretty well.
Who Said Loud Music Damages Hearing? – RightMark Audio Analyzer
Before moving over to our 3D gaming tests, we decided to take a peek at the performance of the onboard audio found on the Asus M3A78-T. Specifically we are checking how much noise is present in the output, as this will affect overall sound quality. Though many claim that onboard audio has matured to a point that makes sound cards redundant, I still say the best experience can be found in a discrete sound card.
The testing performed using RightMark (aka: RMAA) didn’t vary from this assumption. The signal-to-noise ratio on the audio found onboard was a brutal -82dB. Most good sound cards start around -95dB, with the top end easily reaching -105dB or even -110dB. Thankfully Realtek seems to have fixed their problems with positional audio, so although there was a noticeable amount of noise on the audio output it was at least output properly.