|
|
ASUS V9900 Ultra 21st June 2003
The GeForce FX 5800 (NV30) range have been around for a few months now but have only recently become available in quantity. This has co-incided with the release by Nvidia of their GeForce FX 5900 (NV35) displacing them from the top spot. now that there is a higher range above them and they are being mass produced, both factors which will bring the price of the NV30 down we though it was about time we reviewed one of these cards. The review we are bringing you today is of the ASUS V9900 Ultra which is based on the GeForce FX 5800 ultra and is the fastest card in the NV30 range.
Here are the manufacturers specifications:
This card uses the expensive DDR2 memory which is what gives the high end NV30 products their extra muscle power.
On the left we have Nvidia's latest NV35 offering, the GeForce FX 5900 Ultra which you can see is considerably larger than the V9900 Ultra. They both take up the PCI slot closest to the AGP slot but the newer 5900 doesn't use the back plate for air intake/venting. We prefer the alignment of the molex socket on the V9900 Ultra as it involves less stretching of cables.
Looking at the back the
current trend of large surface area heat sinks is continued with the 5900 Ultra
card having almost twice as much coverage.
Here we can see why two slots are taken up by the card. The V9900 Ultra takes air directly from the intake at the back and exhausts hot air from the outlet. Having the intake and outlet so close together is a crazy design from an aerodynamic point of view with a lot of the hot air exhausted being sucked right back in. Nvidia would have been a lot better off with having just an intake or an exhaust but not both. We think Nvidia have realized this and their new cards use a conventional approach to cooling. Having said that the cooler is still large enough to take up an extra slot as can be seen in the above picture.
We used our standard test system based around AMDs Athlon XP3200+ which is the fastest CPU we have. Clicking on the image above will give full details of the system. The most noticeable thing about the card was noise it made when in 3D mode which was akin to a vacuum cleaner. The situation has improved since the first incarnation of the FX 5800 and the noise is barely audible in normal (non-3D) use which is a big improvement. We compared the ASUS V9900 Ultra with the FX 5900 Ultra pictured above and a Ti4400 (no picture needed as they have been around for a while now).
Still
the most important 3D benchmark and a good score is achieved.
The
V9900 Ultra is almost as fast as the newest Nvidia offering and a big step up
from the Ti4400.
A very good
score thanks in part to Nvidia's latest optimized drivers.
The architectural improvements since the GeForce4 are so obvious the V9900 Ultra is actually three and a half times faster than the Ti4400 but less than two percent below the NV35 card.
We like this benchmark and prefer it over Quake 3 (once you get over 400 FPS in Quake 3 it's difficult to have a meaningful comparison). When Doom 3 and Half Life 2 come out we'll see the cards weep but until then we can only really stress them by using high resolutions and/or 4x full scene anti-aliasing. Here we can see that any of the three cards are quite capable of handling this benchmark.
Things
change when we engage 4x FSAA to improve quality. The Ti4400 takes quite a hit
at the higher resolutions whereas the NV30/35 card don't suffer as much. The
V9900 Ultra is actually faster at the lower resolutions but this is probably due
to lack of driver optimization for the FX 5900 Ultra.
The actual botmatch figures show the game is playable at any resolution with any of the cards.
Things change with 4x FSAA and only the NV30/35 cards are really running at an acceptable speed at the highest resolution. Again the V9900 Ultra is ahead in the lower resolutions of the more expensive FX 5900 Ultra.
This test is designed to stress the cards by throwing lots of polygons at them. Standard settings were used but 4x FSAA was selected from the display settings rather than the applications to ensure it was active.
There isn't much between the NV30 and NV35 cards.
Switching to 4x FSAA really stresses the cards and the NV35 pulls ahead. None of the cards could complete the higher resolution tests without the application crashing. This is due to a bug in the benchmark which is getting a little long in the tooth.
The settings used for this test were as follows: Which yielded the following results:
As expected there is little difference between the cards (still waiting for Doom 3 to sort the wheat from the chaff).
The ASUS V9900 Ultra (GeForce FX 5800 Ultra core) did very well in our tests, outgunning the newer FX 5900 Ultra in some cases and being virtually neck and neck in others. Current 3D games don't really use enough advanced features to slow them down but that will all change with a slew of new cutting edge games later this year. It can be noisy when running but thankfully the fan only switches to full speed when using 3D processing and is fairly quiet under normal use. Should people upgrade to the V9900 Ultra? That depends on your goals. If money is no object then definitely or even the FX 5900 Ultra if you can find one. That will ensure future proofing for a while at least. If you are budget conscious but need something that will play future games with all the eye candy turned on then the V9900 will do that and as soon as the FX 5900 hits the market the price of the V9900 Ultra will drop making it an attractive buy. Those who tend to play strategy games and think they do not need 3D power should look at games like Medieval: Total war where a Ti4400 struggles in larger battles. Basically it would be unwise to buy a Ti4400 / Radeon 9000 as an upgrade unless your budget will allow nothing fancier and it may quickly become obsolete if DirectX 9 features start becoming standard in games. The V9900 gives top end performance without the top end price tag. We
would like to thank ASUS UK for the review sample V9900 Ultra.
All trademarks are the property of their respective
owners. |
|
||||||||||||||||||||||||||||||||