HARDOCP - Subsystem Testing, Test Systems & Memory Benchmarks

Date:: Tuesday , October 11, 2005
Author:: Morry Teitelman
Editor:: Kyle Bennett
Google +1

Albatron K8SLI

Are you looking for a small, no-nonsense, and inexpensive SLI motherboard? If so, you might want to consider the K8SLI, which is Albatron’s entry into the nForce4 SLI arena. How well does Albatron's board stack up against the competition from ABIT, MSI, and ECS?

continued...

Subsystem Testing

NOTE: For all Subsystem Testing, an AMD Athlon 64 socket 939 FX-53 CPU (12x200MHz) with a 200MHz FSB and Corsair XMS 3200XLPro memory running at 400MHz was used.

Audio – Subjective Listening

One of the easiest ways to determine the quality of the audio subsystem is via a subjective sound test. Ideally, a sound test requires audio covering the entire spectrum, from subtle to intense. For this test, I chose selected tracks from the Drowning Pool album Desensitized.

Audio playback was clear, with no discernible distortion detectable.

Audio – Microphone Port Testing

The MIC-IN input was tested using a Labtec Desk Mic 524. Spoken words were recorded and played back using Microsoft Sound Recorder, with the Microphone Boost option disabled and enabled. The Microphone Boost option is found within the Advanced menu under the Microphone section within the Volume Control menu.

The record audio pickup and playback were good with and without the Microphone Boost option enabled. However, input volume setting is critical in achieving an adequate recording.

Audio – In Game Testing

In addition to CD or MP3 playback, users most often rely on the audio subsystem for gaming, whether it be for stand-alone first person shooters or online deathmatching. To test the quality of the audio subsystem adequately during game type scenarios, we took benchmark measurements with sound enabled and disabled using the following benchmarks: Jedi Knight 2 and Comanche 4.

The in-game audio performance was very good, with both games suffering little more than a 10% loss with the in-game audio enabled. As expected, the D3D game outperformed the OpenGL game, but not by much. This is an adequate audio solution that you wouldn’t be disappointed with.

Drive Performance

To test the capabilities of the on board USB 2.0 and IEEE 1394 connections adequately, we chose to use an ACOMDATA HD060U2FE-72-USB 2.0/FireWire HDD connected to the USB port only, since the K8SLI does not have built-in support for IEEE 1394. SATA and IDE drive tests were performed using Maxtor 40Gb ATA 133 model 6E040L0 hard drives on the IDE headers and Seagate 80 GB Barracuda SATA hard drives on the SATA headers. The SATA drives were used for testing in a RAID 0 16k block size configuration on the NIVIDA RAID controller. IDE drives on the same ATA-133 port were tested in a RAID 0 16k block size configuration, as well as a mixed IDE /SATA drive RAID 0 16k block size configuration on the NVIDIA controller as well. Testing was also conducted using a standalone SATA drive and an IDE drive connected in a primary slave configuration on IDE1. All drive benchmarks were performed using the open source Iometer program.

As expected, the pure SATA RAID array scored top marks, with the mixed SATA/IDE array coming in a close second. The IDE primary slave drive performance came in a bit lower than expected, while the external USB 2.0 drive performed within expectations.

Network Utilization Tests

Hagel Technologies’ DU Meter software was used in conjunction with Windows Task Manager to measure the performance of the onboard NVIDIA Gigabit NIC. DU meter was used to measure bandwidth, while Windows Task Manager was used to monitor the CPU utilization on the test system. For the test itself, a 750MB archive file containing various sized .WMA audio files was used for the large file transfer test. For the small file transfer test, 750MB worth of various sized .WMA audio files were used. Both tests used an Intel PRO/1000 Gigabit NIC on the host system and a crossover cable to connect the host system to the test system. A crossover cable was used to rule out any possible bandwidth losses due to hub or switch passage.

NVIDIA GigE controller

For the small file transfer tests, the upload speed bested that of download by about 8 MB/s, with neither test showing more than a 20% CPU utilization while running.

The large file transfer test showed similar differences between tests, with the download besting the upload results by about 8 MB/s. The download speeds averaged an impressive 30 MB/s, but not without cost. During download, CPU utilization averaged between 30-40%, while upload only caused a CPU utilization of approximately 20%.

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:

Motherboard

CPU(s)

System Memory

Video Card

Storage

Operating System

Albatron K8SLI (nForce4 SLI)

Athlon 64 X2-4800+, Athlon 64 4000+

1GB (2x512MB) Corsair XMS DDR 3200LL Pro DDR400 (2,2,2,5)

NVIDIA NV45 (400/550 with v61.45 drivers)