rec.autos.simulators

I've got to say, I'm very confused with the V5 in GPL.

At Monza, sat at the back of a full grid (19 opponents) with all the details
turned on and with the in-car view, I used to get 28 fps on my V3.  I now
get 25 fps with the V5.  This is with FSAA turned off, of course.
It's definitely nothing to do with previous V3 drivers either, because I
completely deleted (deltree windows etc) and re-installed windows last
night.
I've checked Eric's fix, and with the mirrors totally disabled, the frame
rate goes up to 32.  With the mirrors set at every other frame I get 28 fps,
and with them set at every third frame I get 30 fps.
Also, as i've said I get 32 fps with the mirrors turned off, and when I run
the V5 in single VSA mode under the same conditions I get 30 fps, so where
is the extra speed from having two chips going?
With everything else, including other Glide games, the V5 has produced a big
increase, without exception.

Are we looking at a driver problem here?  Something like the nVidia cards
used to get in OpenGL.  Maybe it's because GPL uses Glide2, which is a bit
ancient now.

(note- these frame rates are only for the start of full grid races)

The display adpater is "3dfx voodoo5", so i've already got the right drivers
on.  I've also tried the latest leaked drivers.  All of the drivers i've
tried, including the "voodoo series" ones give me exactly the same
performance.
By the way, the performance problems are only with GPL.
Can you tell me whether the "agp aperture size" effects the V5 and what to
set it to?  Also my board is only AGP 2x, so could that be the reason?
Also, do I gather that you get performance way in advance of what I get in
GPL?

Sounds kinda strange?  My framerate stays about maxxed out with everything
on etc. with a V3 and P3700, probably the processor more than the card.
JJ

Sorry, :)  P3-650.  Also my board is only AGP 2x (be6-2).

Try to set the AGP Aperture size to half of your RAM.  However, the size shouldn't matter
since the card does not do AGP Texturing.  The card runs at AGP1X.  BTW, try to run GPL
under a lower resolution 640 x 480 to 1027 x 768 or higher.  The fps shouldn't differ
much if the CPU or system is the bottleneck.  Otherwise, at a lower resolution you should
get a constant maybe 36 fps all the way.

You haven't stated what your CPU is?

I haven't checked the fps at the back of a full grid yet,
but I think I would be getting better than that with my V5.
I was getting better scores with my V3 than you are.

Michael

[snip]

I've now just found out that running "Single chip" is faster than the
"fastest performance" setting for GPL using Glide.  Single chip = 27 fps,
fastest performance = 25.
Does anyone else get this result?

What are you talking about??????

Vintook

  Simon what SCREEN RESOLUTION are you running your tests at.  I have seen a
long thread on the V5 newsgroup that is hinting at the 1600x1200 screens and
up don't work well.  They too think the 2nd chip isn't being used at the
highest resolution screens.  It could be a driver problem, but I haven't
seen any confirmation or a fix yet.
news.3dfx.com for the 3dfx newsgroups and there you'll find the V5
newsgroup.
dave henrie

I'm running in 1024*768, and I also get similar results at 800*600.

You don't get those results?  Or you don't understand my message, in which
case read the whole thread.  :)

He better be using Glide, otherwise I will have to "dog" him :-)

--

"> > Spoiled for choice I call it...

   Hey Greg....I use the D3d patch with my V5.  That allows me to use the
2xFSAA option at 1024x740.  With glide, 2xfsaa locks up.  4x works but you
have to go down to 800x600 to get it enough fps.  The only problem with D3d,
A V5, and GPL is I see lines on the tracks occasionally marking the border
of textures and the flags are not always shown the proper
color...but...since I'm pretty sure that when I see a flag waved at
me...It's gonna be blue...so I know what to do.. :)
dave henrie
  For Simon,  do you get these results with FSAA enabled?  That should show
some difference..
deh

Hehe
Yeah, I'm using Glide.  I've also tried the D3D patch, but as expected it
was slower than glide.

No, without FSAA.  I'm really confused, the dual chip (fastest performance
setting) is way faster than the single chip setting in OpenGL in Q3A, but in
Glide and D3D they both seem to be either the same speed, or else the single
chip is faster.  It must be the drivers.  I'm still testing D3D and Glide
though, so I'll let you know if it happen with everything, or just a few
games.