Soyo SYKT400 Dragon Ultra vs. Shuttle AK37GTR

Soyo SY-KT400 Dragon Ultra vs. Shuttle AK37GTR
Showdown at the KT400 Corral!

By, Robert Maloney
November 6th, 2002

TESTING METHODOLOGY: To help fully explain the scores we listed in the following benchmarks, we felt it was necessary to explain how the systems were setup before running the benchmarks. On all of the boards, we started off by manually optimizing the BIOS settings to the most aggressive RAM settings and system options. The memory frequency was manually set to DDR333 for the first round of tests, and then to DDR400 on the KT400 boards for another round.  The hard drive was formatted each time, and Windows XP Professional with Service Pack 1 was installed. After the Windows installation was complete, we installed the VIA 4-in-1 chipset drivers, version 4.43, and installed the drivers for the rest of the components, using the drivers supplied on the CD, except for the GeForce Ti 4600.  For the GeForce, we downloaded and installed the latest NVIDIA reference drivers, version 40.71.  Auto-Updating, Hibernation, and System Restore were disabled, and then we set up a 512MB permanent page file. On these test systems we set the visual quality to "best performance" in system performance as well as in the video driver settings. Lastly, we installed all of the benchmarking software, defragged the machine, and rebooted one last time.  We brought out the Asus A7V333 board that we had reviewed back in August for a reference point.  If you remember, this board fared very well, and we thought it would be good to compare the difference between a well-rounded KT333 and the newer KT400s. The Hot Hardware Test Systems Can the KT400 replace the KT333?     TEST BOARDS: Soyo SY-KT400 Dragon Ultra (KT400) Shuttle AK37GTR (KT400) Asus A7V333 (KT333)   COMMON HARDWARE:   AMD Athlon XP 2200+ 256MB Corsair PC3200 DDR Chaintech GeForce 4 Ti 4600 On-board AC'97 audio Western Digital WD200BB ATA100 7200rpm 20GB Hard Drive Creative Labs 52x CD-ROM Windows XP Professional with Service Pack 1 VIA 4-in-1 Drivers, version 4.43 NVIDIA Detonator Drivers, version 40.71   Performance Comparisons with SiSoft SANDRA 2002 Sandra, Sandra, Sandra!

SANDRA (the System ANalyzer, Diagnostic and Reporting Assistant) is an information and diagnostic utility put out by the folks at SiSoftware.  It's a quick and easy way to compare results from any system against an internal database of similar systems and drives. These benchmarks are theoretical scores, and can't necessarily be measured in ?real-world? terms, but provide a good way to make comparisons amongst like components.  For each test that we ran, we chose components from the database list that we thought would be found in mainstream PCs.  We ran a set of tests at both the default speeds, and then again when overclocked.  The two columns on the left are from the Soyo Dragon Ultra while the two on the right come from the Shuttle AK37GTR.

Soyo CPU Arithmetic at 133MHz FSB

Soyo CPU Arithmetic at 145MHz FSB

Shuttle CPU Arithmetic at 133MHz FSB

Shuttle CPU Arithmetic at 141MHz FSB

In both sets of charts we can see that the boards are capable of putting up some high scores in the CPU tests.  It's hard to really call a winner as the scores were very close to each other, but the Soyo did come closest to the Athlon XP 2200+ score from Sandra's database, only missing by 5 points.  When overclocking the boards, we were able to get the Soyo stable at 145MHz, while the Shuttle topped out at 141MHz.  In each case, benchmarking when overclocked produced results that were well above all of the database scores.

Soyo CPU Multimedia at 133MHz FSB

Soyo CPU Multimedia at 145MHz FSB

Shuttle CPU Multimedia at 133MHz FSB

Shuttle CPU Multimedia at 141MHz FSB

The CPU Multimedia tests were pretty much in line with what we saw with the CPU Arithmetic, although the delta between the two boards was a little more apparent.  The Soyo board produced scores that were close to, yet slightly behind, the scores listed in Sandra for an Athlon XP 2200+ CPU.  The Shuttle board was still up there, but lagged behind the Soyo board by another 40 points or so.  Overclocking the boards produced some extra points, although obviously a bit more for the Soyo, which was able to get an extra 4MHz over the Shuttle when overclocked.

Soyo Memory Bandwidth at 133MHz FSB

Soyo Memory Bandwidth at 145MHz FSB

Shuttle Memory Bandwidth at 133MHz FSB

Shuttle Memory Bandwidth at 141MHz FSB

Once more, we saw the Soyo board produce numbers that were almost exact with Sandra's score for a KT333 using DDR333 memory, however the Shuttle board could not keep pace.  It was about 40 points behind the Soyo board, and behind all of the other database scores for the other chipsets.  Overclocking the Shuttle board allowed us more bandwidth, eclipsing the KT333 scores by a bit.  The Soyo board, when overclocked, produced a 10% difference between scores.

Both boards produced scores well above the database score for an ATA 100, 7200rpm hard drive using FAT32.  The Soyo board, again, held a slight lead over the Shuttle, getting an extra 100 points in file system performance.

Performance Comparisons with PC Mark 2002 CPU, Memory and Hard Drive Testing

For our next overall PC comparison, we chose MadOnion?s PCMark 2002. PCMark 2002 performs a series of CPU tasks such as image compression, text searches, and audio conversion to give us three scores: CPU, Memory, and Hard Disk Drive (HDD).   It is a relatively quick process for comparing the performance of two or more systems.  We ran the benchmark for all three boards using DDR333 memory timings, optimized as best we could, and then again after raising the memory speed to DDR400. Finally, we ran another set after getting a stable overclock, 141MHz for the Shuttle board and 145MHz for the Soyo.  On both boards, however, we needed to lower the memory timings to obtain a stable operating system.

We got a bit of a shock when we saw these scores, as neither of the KT400 boards outperformed the Asus board, based on the KT333.  In fact, you will see this trend in most, if not all, of the following graphs as well.  Also notable was the drop-off when switching between DDR333 timings and DDR400.  This could partially be explained by the fact that the stick of RAM we used was rated at CAS 2 at 166MHz, but only CAS 2.5 at 200MHz.

Here the performance drop-off was even more evident when switching to DDR400.  It is quite apparent why VIA backed off from DDR400 support, as the performance gains are just not to be found.  Running the memory speeds that far off from the FSB just do not provide good results.  The Asus A7V333 topped the charts again, beating out the Soyo by 11 points, while the Shuttle board pulled into a distant third.  Overclocking the Shuttle board, however, brought the greatest gain.  It raised the Shuttle score by 170 points, narrowing the gap.  The Soyo board, on the other hand, only managed to gain a mere 55 extra points, even though it was overclocked higher.

All of the scores were neck and neck, finishing within 40 points of each other.  Predictably, the file system performance scores dipped slightly when overclocked, as IDE devices don't really seem to take well to running out of spec.