Very fast, very stable and looks *very* good
Written: Sep 29 '00
|
Product Rating:
|
|
|
Pros: Outstanding picture quality--good performance for the visual quality it delivers
Cons: Nowhere near the FPS performance of a GeForce...
|
|
|
| psykosis_fc's Full Review: 3dfx Voodoo5 5500 AGP |
Voodoo5. Ahhh. While still using a Quantum3d Obsidian200sb (and a 100sb prior to that), I have to say that I've definitely enjoyed the cream of the crop when it comes to 3Dfx accelerators. <sigh> After the lengthy wait and enduring countless print ads, web banners, press releases and even TV commercials touting the speed and upcoming dominance of the VSA100, the delivery was less than spectacular. 3Dfx blew theri own horn just loud enough to not notice nVidia performing an end-around. Taking a lesson out of 3Dfx's own book, nVidia comes up with an architecture that's capable of pumping out *major* polys, and was a serious leapfrog in speed over everyone else. 3Dfx got spread pretty thin when they bought some chip fab plants with the intent of bringing manufacturing on complete cards back in-house (Quantum3d was spun off to do the same thing, but ended up developing very successful OpenArcade video solutions). With recent nVidia PR tactics equivalent to a presidential candidate smear campaign, 3Dfx has lost it's throne as the 3d acceleration king. It seems that all the reviews I can find are centered around Q3a performance with little else taken into consideration--great for the GeForce since it seems to be architected specifically for Q3. Again, another lesson learned from 3Dfx. But the question still remained: What does VSA100 bring to the table that V3 doesn't? You can read the white papers and feature sets, sure but those are essentially meaningless to me as numbers on paper don't relate to what I see in-game on screen, ya know? What's *really* sad is that nVidia is telling everyone that they spanked 3Dfx and 3Dfx replies with something akin to "oh yeah?!?? Wait 'till V5 6000 comes out!!" All I can say is, I'm waiting, and I've been waiting for over a year....
I finally replaced my aging V2 setup with a V3 and was pleased at the performance improvement. I built a system with a V5 and was waiting with abated breath to witness firsthand the firebreathing performance that I had been told about in those many advertisements.
My first exposure to the 5500 was at a LAN party; 2 guys had identically clocked systems (950Mhz) with virtually identical system specs and the same sony 21" monitors running (of course) Q3. One system had a Prophet2 GTS running at 1280x1024x32@90hz and the other was the 5500 running at 800x600x32x4xFSAA....all I could say was "WOW!". They were both lightyears smoother in both visual quality and framerates than my system with a v3 3500@200Mhz ever was--and my CPU is running at 1102Mhz! Impressive to say the least. But as I watched more and more I started to notice the differences--I saw that the v5 had noticeably brighter and more detailed texturing than the GTS. There was also signifigantly less banding of textures. You can probably see this in online comparison screenshots of you look at the sky texture and possibly flames such as the torches or something. Unfortunately, the pair was DM-ing all day and I didn't even have the remote chance of piddling with either setup.
Back to building the new system. The guy buying the thing had come to the same conclusions that I had regarding the 2 cards mostly by reading reviews--GTS is faster, but v5 looks better. But how much better? As it turns out, it is good enough that old games actually look good again. Games like SportsCarGT, Unreal/UT, the entire NFS Series and every flight sim I could throw at it looked remarkably better--so much so, it's kindof startling. After the 3rd or 4th title running at 4x FSAA, it starts to get less and less noticable and expected. Almost like going from a 266Mhz computer with 32Mb of RAM to a 1G with 128Mb...first hour or so are shocking at how much faster you are going, then it fades (until you go back to the older system, but that's another story).
As reported by the various hardware review websites, 4xFSAA *does* come at a signifigant performance hit, but there are several settings that can be tweaked in the 3dFX utilities that come with the drivers (hint: these also work with the v3 series) and are very complete, usable and easy to navigate/understand. After tweaking everything to the max for speed and everything to the max for eyecandy, in my experience, there really wasn't that big of a performance difference--MIP mapping and FSAA seemed to be the biggest determining factors to framerates. The lowest MadOnion score that I got with everything maxed at 4xFSAA @1024 was 3784 -- the highest score with everything tweaked for speed was 5570. A great score to be sure (especially for a v5 5500), but visual quality was horrible, akin to a TNT1 running in software emulation of D3D--ick. Strange to me was that the card is only clocked at 168Mhz. Overclocking to 190Mhz yielded another 10fps or so--not worth the heat and stress IMHO. This was an OEM card though, and with news of the recent ATI speed debacle it wouldn't surprise me if the OEM versions were signifigantly slower.
Almost a 2000 point difference in benchmarks would make you think that there is a huge performance gap, but in all practicality, it's not that noticable in gameplay. Realistically, you'd want to run the card in 4xFSAA at 800x600x32 for most applications. Personally, I'm used to 1024, but with 4xFSAA on, there's a little *more* detail than the v3 provides at 1024. My frags in Counterstrike exen went up a tad, but I think that it had to do more with the P3 running over a gig than the 5500 did.
All in all, the card supplied me with all the GLide goodness I could want from a Voodoo architecture. It's somewhat dissappointing how quickly you get used to not having jaggies on the screen and perfect alpha blends out to the draw-in distance. Even more dissappointing is going back to a v3 after playing on one for almost a month..... =(
Bottom line: This is the card to get if you play anything *BUT* Q3a. Hands down, you will be more competitive online of you have a GeForce instead of a V5--FPS, not visual quality get you the frags. But if you play driving sime, plane sims, or just about anything else the V5 5500 kicks some major butt. At 8x6 resolution with 4xFSAA, I never saw the FPS drop below 25 in any game (using the FRAPS utility to check) which in anything but an FPS is more than playable. Visual quality is better than anything outside of pre-rendering. Is it worth $200-300? Depends on how much visual quality is worth to you. For me, I'm mostly happy with my v3 3500tv, since I spent $140 on it and it comes close in performance to the 5500 now that it's overclocked to 220Mhz but I definitely notice the jaggies alot more now. I personally don't think the v5 6000 is going to be "all that", but Quantum3d makes a v5 with 16(sixteen) VSA100 chips and about a gig of RAM if you want the ultimate--you'd have to spend $80k on a new OpenArcade machine to get it, but hey--performance always comes at a price, right?
Word of warning--Don't even *think* about using this card in an AMD system that's clocked over 700Mhz without a power supply that can pump out more than 350watts--a CPU that sucks the juice along with a vid card leads to many random lockups and crashes... Actually, you shouldn't have an AMD without at least a 300w PSU anyway...
Recommended:
Yes
Amount Paid (US$): 209(OEM)
|
|
|
|
Epinions.com ID:
psykosis_fc
|
|
Member: Jeff
Location: Pasadena, CA
Reviews written: 42
Trusted by: 22 members
|
|
|
- Add to My Yahoo!: 
- Add to Google Homepage: 
© 1999-2009 Shopping.com, Inc.