If, like me, you have a fairly decent setup and went with Rendition for
N2, you were probably a bit disappointed with the framerates at
Martinsville. Because Rendition is pretty good with it's textures,
turning the biggies on and off has little effect. Here's something I
found out yesterday by accident when I was trying standard N2 for screen
dumps (BTW thanks, Eric B).
When I tried M/ville (standard N2, remember) without track, grass or sky
textures (I think all the rest are pretty much essential just for
atmosphere) I got very noticable framerate increases over RENDDMA, about
20-25 fps compared to the 15 or so fps previously.
I don't know why RENDDMA M/ville is so bad considering all other tracks
are better than the unaccellerated version. Maybe Papyrus or Rendition
can answer that?
Just thought I'd share that one, hope it works for you.
System
P200, VX M/b, 512 L2, 32 EDO, 3D Blaster
Mark
ps Sorry if this is already common knowledge.
Quoted text (20 lines)
>If, like me, you have a fairly decent setup and went with Rendition for >N2, you were probably a bit disappointed with the framerates at >Martinsville. Because Rendition is pretty good with it's textures, >turning the biggies on and off has little effect. Here's something I >found out yesterday by accident when I was trying standard N2 for screen >dumps (BTW thanks, Eric B). >When I tried M/ville (standard N2, remember) without track, grass or sky >textures (I think all the rest are pretty much essential just for >atmosphere) I got very noticable framerate increases over RENDDMA, about >20-25 fps compared to the 15 or so fps previously. >I don't know why RENDDMA M/ville is so bad considering all other tracks >are better than the unaccellerated version. Maybe Papyrus or Rendition >can answer that? >Just thought I'd share that one, hope it works for you. >System >P200, VX M/b, 512 L2, 32 EDO, 3D Blaster >Mark >ps Sorry if this is already common knowledge.Quoted text (20 lines)
> If, like me, you have a fairly decent setup and went with Rendition for > N2, you were probably a bit disappointed with the framerates at > Martinsville. Because Rendition is pretty good with it's textures, > turning the biggies on and off has little effect. Here's something I > found out yesterday by accident when I was trying standard N2 for screen > dumps (BTW thanks, Eric B). > When I tried M/ville (standard N2, remember) without track, grass or sky > textures (I think all the rest are pretty much essential just for > atmosphere) I got very noticable framerate increases over RENDDMA, about > 20-25 fps compared to the 15 or so fps previously. > I don't know why RENDDMA M/ville is so bad considering all other tracks > are better than the unaccellerated version. Maybe Papyrus or Rendition > can answer that? > Just thought I'd share that one, hope it works for you. > System > P200, VX M/b, 512 L2, 32 EDO, 3D Blaster > Mark > ps Sorry if this is already common knowledge.In my personal (Reactor) experience to date, it's only *with* the
Rendition that Martinsville (only) is noticibly slower RELATIVE to all
the other tracks, but hey, it sure is pretty. Without Rendition, it's
performance is similar to the performance you get with all the other
tracks.
Bearing that in mind, performance improvement and graphics quality with
Rendition is ....well, a qualitative or even a quantitative explanation
is simply inadequate. You simply have to witness it yourself. The
graphical improvement was greater than I expected, and outweighs a
processor upgrade in terms of both framerate and $$$$$$$ value by far.
For a pleasant change the advertising claims are not an exagerated
aspect of someone's fanciful imagination.
ie: Rendition bus mastering increases the *available* video I/O path tfr
rate 4x (from 20Mb/sec FIFO to 80+ Mb/sec DMA), whilst the onboard RISC
processor reduces the demand upon (up to 50% claimed) and need to move
data to/from the PC's CPU, thus further reducing the bus load and data
handling.
In a word STUNNING! Without question, one of the best value and
satisfying purchases I've made in recent years of computing.
*
*
*


































