I have a 4000XT here now but no system to put it in as of yet, and a 4500 on the way. One thing I have found when researching Kyro/Kyro II cards, though, is not to let synthetic benchmarks prejudice you against them. They actually perform better in games than the benchmarks would seem to indicate because of how they render. They only render what you can actually see on the screen at any given moment instead of rendering every object in the scene, many of which you may only see partially or not at all. Ignoring objects or parts of objects that aren't actually visible saves a lot of time and keeps the framerates higher than you would expect. I haven't seen too many reviews that contained both a Kyro and a Voodoo, though. Most reviews put Kyros and Voodoos against an assortment nVidia and ATi cards instead of against each other but I'm sure you could find at least one review that has both if you look hard enough. Both seem to test slower than a GeForce 2 GTS or Ultra, if that's any help, and neither has a hardware T&L engine built in. They probably are somewhere in the TnT2/Geforce SDR/DDR range.

Kyro is extremely efficient but also has relatively little fillrate. If it had as much raw fillrate as a GeForce 2 GTS, it would completely blow it away. NV15 (GF2 GTS and friends) is very inefficient and wastes a lot of fillrate and memory bandwidth.

This is why the Radeon DDR could keep up with GeForce 2 GTS in 32-bit color. Radeon is a lot more efficient but again has less than half of the pixel fillrate.

GeForce 4 MX is really the pinnacle of that GeForce 2 tech from an efficiency standpoint. It's even more efficient than Radeon 7500.

PowerPie5000 wrote:

Ok which is better in terms of picture quality, performance and compatibility? (not including Glide).

Kyro2 easily. At least its drivers are officially newer by 3 more years. 😉

It will have severe problems with 2002 games and later.

It seems the Kyro II totally annihilates the V5 in this benchmark (if it's true):

http://www.techspot.com/reviews/hardware/vivi … ividxs-10.shtml

It makes me wonder if one of the VSA-100 chips was disabled without them realising... or is the Kyro II really that good? It must have been a real underdog back then.

PowerPie5000 wrote:

It seems the Kyro II totally annihilates the V5 in this benchmark (if it's true):

http://www.techspot.com/reviews/hardware/vivi … ividxs-10.shtml

It makes me wonder if one of the VSA-100 chips was disabled without them realising... or is the Kyro II really that good? It must have been a real underdog back then.

Village Mark is a synthetic benchmark (which was written by the creators of the Kyro chip!). You should always take those with a grain of salt. Look at the actual Quake 3 scores. They aren't that far apart. The Serious Sam scores are also within a frame or two of each other. Real game performance often means a lot more than benchmarks. Now, since the Voodoo 5 performs similarly but with two GPU's, I would say the Kyro II is a better chip, BUT the Voodoo does Glide so that tips the balance in it's favor. I'd like to see the Kyro II against the single chip Voodoo 4, though, just to be sure that the Voodoo 5 wasn't running on just one GPU. The review also doesn't say what they tested on. A faster CPU might make a big difference if either of the cards is CPU bound. From other reviews I have gathered that the Voodoo 5 is more GPU bound than CPU bound so I would have to guess that the Kyro II would benefit the most from a faster CPU.

sliderider wrote:

Look at the actual Quake 3 scores. They aren't that far apart. The Serious Sam scores are also within a frame or two of each other.

I think those game benchmarks are only showing Kyro-II results using different filters and texture modes... I don't see it being compared to the V5 in these benchmarks 😖

KYRO isn't good with bilinear filtering precision. Play Operation Flashpoint and watch that ground block and jitter!

Villagemark is an extreme case of overdraw and the point is to show how great PowerVR's tech is at not drawing what isn't necessary. The tiled deferred renderer is apparently 100% efficient at removing overdraw, which is awesome of course.

But, like leileilol says, it's not exactly free of problems and bugs. And the chip just doesn't have enough pure grunt to beat the brute force competition. It could be that they intended to go up against GeForce 256 instead of GeForce 2. Voodoo5 and Radeon had that problem too.

I remember when they were new. It's too bad they were never continued, could've been interesting to see how far they could've gone.

I never owned one but I'd definately love to have one 😁

Tetrium wrote:

could've been interesting to see how far they could've gone.

Don't get your hopes up. The card stopped being cool to use when NOLF2 and Battlefield 1942 came out. Those games gave the Geforce2 its purpose for the first time. 😀

leileilol wrote:

Tetrium wrote:

could've been interesting to see how far they could've gone.

Don't get your hopes up. The card stopped being cool to use when NOLF2 and Battlefield 1942 came out. Those games gave the Geforce2 its purpose for the first time. 😀

We can always play those games on our Win XP/Vista/7 machines... I think a Kyro-II for a retro Win9x rig would be pretty decent 😀

Also wondering how does the ATI Rage Fury Maxx compare with both the V5 5500 and Kyro-II? Just deciding if i should hang on to it or not 😕

PowerPie5000 wrote:

Also wondering how does the ATI Rage Fury Maxx compare with both the V5 5500 and Kyro-II? Just deciding if i should hang on to it or not 😕

Rage Fury Maxx is slower. It's a good card to keep, though, because there doesn't seem to be a lot of them for sale. They seem to come and go in bursts. You could go a year without seeing one for sale then all of a sudden there's 10 of them on ebay.

sliderider wrote:

PowerPie5000 wrote:

Also wondering how does the ATI Rage Fury Maxx compare with both the V5 5500 and Kyro-II? Just deciding if i should hang on to it or not 😕

Rage Fury Maxx is slower. It's a good card to keep, though, because there doesn't seem to be a lot of them for sale. They seem to come and go in bursts. You could go a year without seeing one for sale then all of a sudden there's 10 of them on ebay.

I've decided to let it go as i only have 1 retro PC which contains my V5 5500... so there's no need to keep any other old graphics cards. I've recently had a big clearout on all my old components, computers, consoles and games. I've only kept the stuff that i'm actually using (i can now move around the house 🤣).

Anyway i've listed the Rage Fury Maxx here: http://cgi.ebay.co.uk/ws/eBayISAPI.dll?ViewIt … =STRK:MESELX:IT

Rage Fury Maxx is from the worst era of ATI driver quality. There are a lot of unsolved issues. I ran into some really bad stuttering with a few games. The Rage 128 chip also has perhaps the worst 16-bit color dithering.

Oh and if you want to run XP/2K, only one of the two chips on a Maxx will function. It's also likely that the driver quality is much lower than in 9x. ATI dropped driver support in 2001.

swaaye wrote:

Rage Fury Maxx is from the worst era of ATI driver quality. There are a lot of unsolved issues. I ran into some really bad stuttering with a few games. The Rage 128 chip also has perhaps the worst 16-bit color dithering.

Oh and if you want to run XP/2K, only one of the two chips on a Maxx will function. It's also likely that the driver quality is much lower than in 9x. ATI dropped driver support in 2001.

Yeah and from what I read the issue with 2k/XP isn't something that can be solved with a driver update anyway. From what I read ATi themselves gave up on trying to make the MAXX work because 2k/XP doesn't handle dual GPU's in the way that ATi implemented them so no amount of fiddling around with the drivers is going to help. You'd have to hack Windows itself and that's just too much work even for the most dedicated hobbyist to support a video card that almost nobody has anyway. 😒