|
Gunmetal is a DirectX 9 based game that has pretty good graphics. The design of the game is quite interesting, as you get to play as a bot, and a plane. Lets see how the FX5900 Ultra performs in this new test.

| Video Card |
Minimum FPS |
Average FPS |
Maximum FPS |
| MSI FX5900 Ultra Oc'd: |
17 |
33.36 |
58 |
| MSI FX5900 Ultra: |
17 |
27.62 |
55 |
| Radeon 9000 Pro AiW: |
10 |
17.63 |
44 |
| Matrox Parhelia: |
8 |
17.14 |
45 |
There are a couple of notes for this test. Both the Radeon and the Parhelia were not able to render the scene correctly. The Radeon could not provide all the textures that were needed. The Parhelia didn't have the vertex shader version 2.0 required by DirectX 9. Otherwise the FX5900 performs very well, as during the demo I was able to play the game on the FX5900 Ultra without problems or serious slowdowns. What about adding AA and ansio to this test, what happens to the performance then?

| Video Card |
Minimum FPS |
Average FPS |
Maximum FPS |
| MSI FX5900 Ultra Oc'd: |
19 |
29.11628 |
53 |
| MSI FX5900 Ultra: |
16 |
27.39416 |
49 |
| Radeon 9000 Pro AiW: |
4 |
10.20053 |
33 |
| Matrox Parhelia: |
8 |
14.8062 |
44 |
When AA and ansiotropic filtering is added, we see that the frame rate loss is much less than one would expect. The results are still fairly playable for the FX5900 Ultra, especially when overclocked. On the lower end of the performance scale the Parhelia beats the Radeon and has about half the performance of the FX5900 Ultra.
X2 is a different kind of game from most benchmarks. It is a space simulation based on DirectX 9, and offers some very nice graphics, as one would hope from a newer game. It also has a benchmarking option, which we used. The only difference from the standard settings was to enable the shadow support. Lets see the results.

| Video Card |
Minimum FPS |
Average FPS |
Maximum FPS |
| MSI FX5900 Ultra Oc'd: |
N/A |
64.117 |
N/A |
| MSI FX5900 Ultra: |
N/A |
61.824 |
N/A |
| Radeon 9000 Pro AiW: |
N/A |
25.921 |
N/A |
| Matrox Parhelia: |
N/A |
18.997 |
N/A |
The results here show there is a difference between a DirectX 8 and DirectX 9 based hardware. The FX5900 Ultra is over two times faster than the other hardware that we tested here. There is a slight difference in performance (4%) when we overclock the FX5900 Ultra. Surprisingly the Radeon 9000 Pro is faster than the Parhelia, which is more a DirectX 9 card than the Radeon 9000 is. What happens with AA and ansiotropic filtering turned on?

| Video Card |
Minimum FPS |
Average FPS |
Maximum FPS |
| MSI FX5900 Ultra Oc'd: |
N/A |
47.911 |
N/A |
| MSI FX5900 Ultra: |
N/A |
44.668 |
N/A |
| Radeon 9000 Pro AiW: |
N/A |
7.643 |
N/A |
| Matrox Parhelia: |
N/A |
16.504 |
N/A |
We see that nothing is really changing in this test, the FX5900 still dominates the tests. The Parhelia is much faster than the Radeon 9000 Pro, but that is saying that one is more unplayable than the other. The FX5900 Ultra puts on a very good performance losing about 34% of its non-AA performance in this game. Even at 1024*768 with 4X AA and 8X ansiotropic filtering the FX5900 Ultra is very playable in this game.
Conclusion
So having looked at the card, the packaging, the drivers, video in/out, and plenty of performance tests, what can we conclude?
First the box and contents. The nBox itself really gives a very nice first impression, that of class and high quality. The inner box is pure black with glossy black and green lettering. Inside the box everything is placed very well, the card itself has a level all to itself, that of the first one you see. Underneath the card is a layer containing all the other stuff you get. The inclusion of three games that are commonly played now is very nice to see, and a welcome change to MSI's standard 14 games that are really of little use to one. The inclusion of the mouse was a little odd, as most who buy this card don't need a mouse, much less a small optical one. However during use I found it very responsive, though too small for my hands, so for those with small hands this is a very good mouse.
The drivers are nothing to write home about, though I did not have any problems with them during testing and gaming. The menus are very logically laid out and all the features that you expect to be there, are. The most notable menu is that of the temperature menu, which gives you the actual temperature of the video card. The 3D quality of the Antialiasing and the ansiotropic filtering was very good, with the 8X AA almost reaching the quality of the 16X FAA of the Parhelia. The ansiotropic filtering was by far better than any of the other cards, including the Radeon 9000 Pro's 16X, though the sample could be from an area that the ATi card didn't filter.
The cooling system is next. MSI did a masterful job of keeping the card cool while keeping it almost silent, and by almost I mean that you wouldn't notice it unless you were close to the fan and no other fans were on. Turning off one of the two fans quiets it down even more so from 63dBA to 56dBA (from 3' away) almost as silent as the Radeon 9000 AiW, which has a smaller fan. Directly in line with the cooling of the chip is that of the overclocking of the card. With both fans, and memory heatsinks the card reached 50MHz/50MHz (100MHz) core/memory faster than the stock 450MHz/425MHz (850MHz), which is nice and helped performance in UT2003 almost reach 60fps where it couldn't before.
Now on to the performance of the card. For the most part in the synthetic tests the FX5900 Ultra and the Radeon 9800 Pro performed very similarly. In SPECviewperf the MSI 5900Ultra did much better on the DRV-09 test than the Radeon 9800 card, but worse on the DX-08 and UGS-03 tests than the Radeon 9800. In OpenGL games the MSI 5900 Ultra was as fast as the Radeon 9800 Pro in all tests, though with overclocking at the highest resolution in Jedi Knight II it provided slightly better frame rates. In Serious Sam SE we didn't have results from the Radeon 9800 but to say that the 5900 Ultra beat the Parhelia and the Radeon 9000 Pro would be an understatement. Moving to DirectX 8 games we see that the MSI FX5900 Ultra enjoys a decent lead especially at 1600*1200 both with and without AA and ansiotropic filtering enabled. In Splinter Cell we see that without overclocking the MSI FX5900 isn't as fast as the Radeon 9800, and at 1600*1200 it is a couple of fps slower than the 9800 even when overclocked. Looking at future games based on DirectX 9 we see very little as the tests weren't conducted on the Radeon 9800 as well. In Gun Metal the game was playable with the FX5900 Ultra at 1024*768 without AA and ansiotropic filtering. In X2 the FX5900 pretty fast providing over 60fps on average with that card, without AA and ansio.
Lastly we will look at the value of the card. The card is priced pretty high at , but it is pretty close, within about $100 (US) to its comparable 256MB cards, both ATi and nVidia based. The inclusion of the games which come to close to 3/4 of the difference helps bring the card more in line with its competitors, and the video capture/editing software and the mouse really bring it on par with comparable offerings. Compared to the Radeon AiW 9800 Pro the FX5900 Ultra is overpriced as it doesn't include a TV-tuner or a remote control. The major advantage that the MSI nBox FX5900 Ultra has over the Radeon AiW 9800 Pro is the three games mentioned previously. Good Points
- Fast performer
- Good quality AA and ansio
- Video In and Out
- Temperature Monitor
- Decent Overclocking
- Very Quiet
- Three very good games
- Very nice box and mouse
Bad Points
- Expensive comparatively
- Not much faster than Radeon 9800 Pro Currently
- 2D quality isn't the absolute best
- Big card
- TV-Out isn't very good
- TV-In isn't very good either
Final Words
This card has all the features that most would want, minus the TV-Tuner, including some actually played games and apart from some less than adequate TV quality it is pretty good. The Radeon 9800 AiW is a better value, but if you are looking at FX5900 cards, this is the best I've seen around, so I'm giving it a recommended award.

If you got any comments, be sure to hit us up in our forums.
HOME
|