Rage Pro Does Curves

Im not going to make any decision until someone has opened a shirnk wrapped box bought 1 month after release and benchmarked THAT.

Hmm, on second though they do look a little out of line. Oh well, hehe. Hey what\'s that over there!!??

The fps in crusher demo are a mistake (P3/640/TNT2U). If you take a little time to compare it to the three other demos it should be clear to everybody. (everybody?, hope so, haha)

#43 You\'re calling me a fool but you fail to grasp a few simple facts:
1) An Athlon 650 is undispudedly faster than a P3-600
2) The result on the Athlon system was slower than that for the P3-600
Hence, since the video cards were different, the video card is logically what slows the slower system down. Pretty simple eh? As for supposing that the GeForce is really faster and that they just fudged up that benchmark, the best you can do is speculate.

We\'ll see what happens when Steve gets his review together, I would personally be surprised if the GeForce was slower, but who knows, it\'s had really shitty scores up till now.

I don\'t know, the Athlon 650 in that review impressed me more than the GeForce.

Oh, and please tell us if the GeForce actually gives constant framerates on a lower end system (if you have a celery 400 or something).

Steve, could you try benchmarking q3demo1 using the high geometry complexity settings that Carmack specified in one of his .plan updates? I think it was r_lodbias -2 and r_stencilshadows 1. Pretty sure about the first setting, not sure about the second though.

Steve, benchmark the fps in q3test1 standing behind the quad looking down toward the RL exit, it\'s the slowest room in q3 because it has close to 5000 triangles. If geometry acceleration is enabled, there should be a difference in the lower resolutions (512x384 or lower).

kandyman

http://www.the-trench-coat-mafia.com/

What kind of useless site is that? Trying to ride on the shooting\'s publicity?

#38 The only one who is a fanboy is one who would dismiss a bad result as a simple \"typo\". It looks to me like the GeForce could have a problem at low resolutions. It\'s possible that they screwed up too, but hey, they have a GeForce and you don\'t.

Hahaha. Notice in that review, the TNT2U on a P3-600 *beats* the GeForce on a K7-650. How lame is that? You get a kick-ass CPU like the Athlon, and you get the most expensive video card out there, and that video card actually SLOWS YOUR CPU DOWN. What kind of crap is that???

T&L is supposed to offload your CPU, not slow down your CPU. I hope they get this bullshit sorted out soon, I don\'t want to buy a POS video card this fall.

Getting whipped in public because of a hardware review....that would blow, lol.

Well, #30, that hardware site is in Singapore. I think he might get censored/fined if he says fucking instead of freaking.

Good thing I don\'t live in Singapore :)

Heh,

The guy who did that didn\'t have a TNT2 ultra available. If you read his explaiantion, that TNT2 benchmark was from 2 months ago. It seems to me that the DDR version will be the one to get if you plan on playing at 32 bit. $300 , bleh! That TNT2 is looking better and better.

Oh yeah, and the BitBoy seems to be right on with the memory bandwidth comment. Look at the 32 bit performance vs. the 16 performance.

sigh, notice they have geforce benchmarks on the athlon 650, but the benchmarks they compare it to for the tnt2u are run on a p3 600.

how about this.
some freaking benchmarks side by side of the tnt2u vs geforce on the same damn cpu?

ill have those posted tomorrow.

Go read that hardware site\'s review of the cl-annihilator

http://www.hardwarezone.com/reviews/video/CL-GeForce256/cl-geforce256.html

These guys are HardwareZONE, not HardwareONE, and they have benchmarks that actually show substantially DIFFERENT results! :)

Actually, now that I look at q2 numbers, it\'s as fast running on a p2-350 at 800*600 as a TNT2 ultra running on a P3-600. Not bad, considering NEITHER cards were fillrate limited at that point.

So hardware T&L seem to work pretty well in q2, too bad q3 is what we\'ll be playing. Let\'s hope id and Nvidia get their acts together on hardware t&l for q3.

I think the idea was to show that the nv10 could run at a little extra beyond the normal high detail and not take a speed hit. Of course they didnt have any benchmarks, so it would seem they are just hitwhores at this point.

Turns out it does do ~170 fps on q2 at 800*600.... on an Athlon 650, yeh!

So their drivers are optimized for the \"Tree demo\" but they\'re not optimized for a real game like Q3. Or Q3 is not \"optimized\" for T&L. I don\'t know if I buy that, what Creative seems to want people to believe is that at some point after you buy their cards, there will be a magical driver and a magical Q3 setting that will make the card fulfill all its promises.

Carmack never said that T&L is disabled in Quake3. In fact, if Quake2 can use T&L *right now*, why shouldn\'t Quake2? Is Creative\'s idea of an \"optimized Quake3\" just a detail slider that will make Quake3 2x slower on non-T&L cards but only 10% slower on the GeFarce? Screw that, what I\'ve been sold on is the idea that the GeForce will accelerate Quake3 at the default (already high) detail setting, setting it apart from existing cards without having to cram lots of polygons where none are needed.

Bottom line: I want a card that will let me play Quake3, out of the box, FASTER than any other card, without enabling some detail settings that are hardly noticable. I think most people here feel the same way.