Testing Methology:

We will compare the DREAMCODE to our favourite test bed, the Abit BH6 + Celeron 300A.  To fully test this motherboard, we will test each of the three major areas:  Motherboard performance, TNT performance and Yamaha XG performance.  To do this we will run both test configurations through the following tests:

Overall System Performance

• SiSoft Sandra 99 Benchmarks

Riva TNT Performance:

• 3DMark 99 MAX
• Quake II (Demo1 and 3Finger's Crusher Demo)

Yamaha XG Performance:

• Our ears in various DirectSound3D tests
• ZDNet AudioBench 99 (CPU utilization under DirectSound3D)

Here is the system configurations used in this test:

Drivers used:

• nVidia Riva TNT Detonator driver 1.09 - for both DreamCode TNT and STB Velocity 4400
• Diamond Monster Sound II Driver - version 1.02 - A3D Driver 2.05
• Yamaha XG 724/740 Driver version 1.028

Important Notes:

Both TNT cards were tested with a core speed of 95MHz and a memory speed of 120MHz.  This was highest possible speed achieved without lockups.   Even though the PowerColor Onboard TNT is rated at 100/125MHz operation, ours would not make it up to this speed stably, despite the fact that it has a cooling fan on it.

In the SiSoft Sandra 99 CPU benchmark, the score between the BH6 and the DREAMCODE are almost identical, with the DREAMCODE only slightly edging out the BH6.  This has been a common trend with BX; they perform within 1-2% of each other.

The memory benchmarks are a little more separated, the DREAMCODE beats the BH6 in this category.  Again, these differences are only minor and probably won't translate into any noticeable performance increase but it is nice to see the DREAMCODE coming out on top, especially since its a newer board.

Here we see that at the same clock speed, the PowerColor's onboard TNT and STB's Velocity 4400 perform almost the same, just as we expected.

The Crusher demo results show a slightly different pattern.  Here we see the DREAMCODE's onboard TNT beating the STB Velocity 4400 by quite a significant margin: 13% faster at 1024x768 and 10% faster at 800x600.  The Crusher demo really strains the system to its maximum and perhaps this is where the integrated TNT makes a difference. Having the TNT on board means its a lot closer to the system bus and can perform at a slightly higher speed than an add-on peripheal. The Crusher demo is one of the best benchmarks, thanks goes out to 3Fingers for making this demo.

3DMark 99 is a great benchmark for testing your 3D graphics subsystem.  It also takes into account processor floating point performance and image quality.  As we can see from the results, the DREAMCODE once again beats the STB Velocity 4400 running on the BH6.  It looks like PowerColor has done a great job of optimizing this board.

lower scores are better

Since this is a gamer's board, 3D audio is definitely going to be utilized on the Yamaha chip so we decided see how the CPU utilization compares to an MX300 with the latest drivers.  The test was done using ZDBop's Audiobench 99 and only the worst score was taken.  In this case, the worst numbers were achieved with DirectSound3D at 44.1KHz, 16-bit static and even at this level, 10% CPU utilization is not too bad considering the old ISA sound cards which were much higher.

As you may have heard, the Yamaha XG chip supports DirectSound3D and A3D 1.x  A3D 1.x is emulated (just like on the SB Live!) through DS3D calls through the Sensaura 3D audio API.  When you load an A3D title, a neat little 3D Sensaura audio splash splashscreen pops up just like the A3D one does on Vortex based cards. Yamaha is also going to be adding EAX support through the Sensaura API; I'm not sure if that has been implemented as of yet. If anyone can tell me a way to test this, please let me know.

So, this chip is definitely 3D audio capable, but how good does it sound?   Well that is completely subjective, there is no 4-speaker support in this chip, so you have to rely on 3D sound from 2 speakers.  My MX300 does quite a nice job at this, so that is what I am going to compare it to.

When I ran ZDBop's AudioBench 99, I ran the subjective audio tests to see how well DirectSound3D was implemented and to my surprise, it worked quite well, almost as well as it did with my MX300.  I was able to pinpoint the direction of the DirectSound3D helicopter horizontal test 80% of time and the vertical test 100% of time! If you haven't tried this test, what it does is play a helicopter sound which is supposed to be circling around you and you have to chose which direction its circling.

As far as the Sensaura API goes, it seems to be quite good from my listening experience.  I'm not expert on this API so here are some links to get all the information you desire:

• Yamaha Sensaura Support Page
• Sensaura Home Page
• 3D Sound Surge Sensaura 3DPA Article

Next Page: Ratings and Conclusion

Hardware Pros Home Page

The Hardware Pros are not responsible from any damage resulting from overclocking.
All trademarks used are properties of their respective owners.
Copyright © 1999 Hardware Pros. All Rights Reserved.