View Full Version : Xabre reviewer can't read numbers

06-29-2002, 09:57 AM
was just reading the review of the Xabre refrence card.

"Max Payne being based on DX8 technologies gives the Xabre a bit more of a advantage. Still unable to beat the nVidia in the core area, it does make up some ground, especially in FSAA"

now whats this about the xabre gaining back ground in the FSAA test??

it actualy looses more ground:

77/85 = .9 so its only 9 10ths of the performance of a gf4mx
20/26 = .76 so its less than 8 10ths of the performance of a gf4mx, and as were talking about 20 V's 26 FPS the difference is even more important than the 8 frames at full speed, they don't make any difference to playability, but 6 frames when your below 30 is the difference between being able to play for 1 minute and 5.

06-29-2002, 11:03 AM
The percentages between the FSAA off and FSAA on changed so the "on" was closer to the GeForce. Yes, it is still behind, but the percentage between them is slightly better than with FSAA off. Cameron Johnson is correct in his assessment of the Xabre gaining some ground.

06-29-2002, 11:05 AM
no with it off the xabre was 90% of the gf4mx
with it on it was only 76%, how does that mean it closed the gap when you turn FSAA on?

06-29-2002, 11:17 AM
Oh, waitaminute. I got screwed up, even read that part myself:rolleyes: My bad, I see you spotted something...I didn't. Maybe its' some dyslexia kicking in:cackle:

06-29-2002, 11:19 AM

06-29-2002, 11:50 AM
You're right if we are talking on percentages, Cameron Johnson was talking raw frames.

06-29-2002, 02:42 PM
Soooo, technically, both sides are right:cackle:

06-29-2002, 03:30 PM
but when the scale has changed by a factor of 2 you can't continue comparing difference, you have to use frame rates.

if you read the text it sounds like your less likly to notice the difference between the Xabre and the gf4mx if you turn FSAA on, which is BS

06-29-2002, 03:40 PM
It says make up some ground.

06-29-2002, 03:58 PM
when its performance is getting even worse

06-29-2002, 04:13 PM
We have already determined the benchmarks could be read both ways and are both correct. End of story....