rec.autos.simulators

i currently have a voodoo 3000 and am unfortuantely need to look
elsewhere for improved performance for playback of my dvd's and the
upcoming world sports car , any info would be very well appreciated.

thanks

Rich

Twice the performance of the MX, and no more than $80 more......

An MX card will be a step BACKWARDS for GPL. Your mileage will vary in other
sims, but in GPL an MX card is definitely slower than a V3 - been there and
tried it!
                         Phillip...

-Gregor

First N4, 1024 x 768 with shadows and reflections etc. off, medium car
detail and half way on the sliders for draw distance and world detail.
V4 D3D: 40 to 60 fps, with a drop to about 35 in a pretty big accident. Open
GL is pretty much unusable on a Voodoo card in N4, with framerates between
18 and 25 fps.
GE Force MX Open GL: 40 to 70fps but in an accident the drop was to about 25
fps. The performance was slightly less in D3D, but not by much.

Next GPL, 1024 X 768, rear of the grid at Monaco, full detail slider all
graphics maxed out.
V4 Glide: When entering track 30fps, dipping to the high 20's at the start
but maxing at 36 by the end of T1 and staying there unless a big accident
happens.
V4 D3D: 25 fps when entering the track, dipping just below 20 but  maxing at
36 by the time the field has spread except for the odd dip to 34 or 35.
GE Force MX D3D: 20fps on entering track, dipping slightly on the start, but
again maxing out after the field spreads, except in accidents or when the
field bunches etc.

Finally for race sims MBTR which has it's own benchmarking utility. I tested
both cards at 800 x 600 using the recommended settings in the config
utility.

V4 4500: gave a benchmark of 70fps.
GE Force MX: gave a benchmark of 35fps.

I then tried them both in Unreal tournament, the V4 running in glide and D3D
gave me an average framerate of about 60 fps. The Geforce running in D3D
gave an average of just under 40fps. The slowdown in UT was unbearable at
1024 on the GeForce card, but was quite stable on the V4. Thinking about it,
I should have given the GeForce a try in Open GL , but it was getting late,
and I'd pretty much decided that it wasn't the card that I had hoped for.

In conclusion, I had hoped for a good showing from the GeForce card, but
felt very let down after hearing all the good reports about them in this and
other newsgroups. The only sim it out performed my V4 was NASCAR 4, with
other sims being slightly better on the V4 and MBTR and UT seriously
knocking the performance of it. So i'll stick with my V4 for now (the
GeForce has been returned from whence it came), and hopefully there will be
a good video card manufacturer to take the fight to NVidia, possibly ATI
with the Radeon, but at the moment they are rather expensive here in UK.

--
Ian P
<email invalid due to spam>

thank you for providing a thorough comparison. I would imagine both the
cards being quite on par in performance. As you've shown it, it really
depends on the games you try to run.

N4 takes good advantage of T&L and this is probably where the slight
edge came from. It would probably be even more obvious on a lower level
machine as a faster processor could even be better off trying to do the
T&L calcs itself instead of passing them to the card, depending on the
situation, of course. I believe CMR2.0 also gains quite a lot from this.

However, as these are mostly 'budget' cards (thogugh how high we've set
our standards as to what 'budget' is in the recent times is mind
boggling!), I don't expect them to be replaced by their owneers as soon
as the next big little thing comes out, or they would have bought
something better at this point. Therefore, there is a slight risk
involved in sticking with the Voodoo's as when the first titles
requiring DX8.0 will come out, there could be possible driver issues
with the 3dfx cards. I would advise anyone going for a budget 'card' at
this point to consider this issue as well.

-Gregor

Ash

| Please bare in mind the MX cards are using SDR ram not DDR.  This make a
big
| difference in performance.  Any MX card is going to be much slower then
the
| DDR counterp[art. Thus the price difference.
|


| > I bought a GE Force MX 32MB (hercules prophet) on Friday to replace a
| Voodoo
| > 4 4500 .
| > Here are my results on an Athlon 750 with 256MB RAM.
| > I used the Hercules version 6-35 drivers for the GeForce, the refernce
| > drivers wouldn't install correctly and gave me an error message when
| trying
| > to access the advanced properties page, and version 1.04 drivers for the
| V4.
| > I found the results rather shocking !! ;)
| >
| > First N4, 1024 x 768 with shadows and reflections etc. off, medium car
| > detail and half way on the sliders for draw distance and world detail.
| > V4 D3D: 40 to 60 fps, with a drop to about 35 in a pretty big accident.
| Open
| > GL is pretty much unusable on a Voodoo card in N4, with framerates
between
| > 18 and 25 fps.
| > GE Force MX Open GL: 40 to 70fps but in an accident the drop was to
about
| 25
| > fps. The performance was slightly less in D3D, but not by much.
| >
| > Next GPL, 1024 X 768, rear of the grid at Monaco, full detail slider all
| > graphics maxed out.
| > V4 Glide: When entering track 30fps, dipping to the high 20's at the
start
| > but maxing at 36 by the end of T1 and staying there unless a big
accident
| > happens.
| > V4 D3D: 25 fps when entering the track, dipping just below 20 but
maxing
| at
| > 36 by the time the field has spread except for the odd dip to 34 or 35.
| > GE Force MX D3D: 20fps on entering track, dipping slightly on the start,
| but
| > again maxing out after the field spreads, except in accidents or when
the
| > field bunches etc.
| >
| > Finally for race sims MBTR which has it's own benchmarking utility. I
| tested
| > both cards at 800 x 600 using the recommended settings in the config
| > utility.
| >
| > V4 4500: gave a benchmark of 70fps.
| > GE Force MX: gave a benchmark of 35fps.
| >
| > I then tried them both in Unreal tournament, the V4 running in glide and
| D3D
| > gave me an average framerate of about 60 fps. The Geforce running in D3D
| > gave an average of just under 40fps. The slowdown in UT was unbearable
at
| > 1024 on the GeForce card, but was quite stable on the V4. Thinking about
| it,
| > I should have given the GeForce a try in Open GL , but it was getting
| late,
| > and I'd pretty much decided that it wasn't the card that I had hoped
for.
| >
| > In conclusion, I had hoped for a good showing from the GeForce card, but
| > felt very let down after hearing all the good reports about them in this
| and
| > other newsgroups. The only sim it out performed my V4 was NASCAR 4, with
| > other sims being slightly better on the V4 and MBTR and UT seriously
| > knocking the performance of it. So i'll stick with my V4 for now (the
| > GeForce has been returned from whence it came), and hopefully there will
| be
| > a good video card manufacturer to take the fight to NVidia, possibly ATI
| > with the Radeon, but at the moment they are rather expensive here in UK.
| >
| > --
| > Ian P
| > <email invalid due to spam>
| >
| >



| > > >


| > > > > anyone had any problems with this type of card in gpl with the
| > > > > direct 3d patch. what sort of performance do u get on demanding
| tracks
| > > > > like 3rd party ones?
| > > > >
| > > > > i currently have a voodoo 3000 and am unfortuantely need to look
| > > > > elsewhere for improved performance for playback of my dvd's and
the
| > > > > upcoming world sports car , any info would be very well
appreciated.
| > > > >
| > > > > thanks
| > > > >
| > > > > Rich
| > > >
| > > > An MX card will be a step BACKWARDS for GPL. Your mileage will vary
in
| > other
| > > > sims, but in GPL an MX card is definitely slower than a V3 - been
| there
| > and
| > > > tried it!
| > > >                          Phillip...
| > >
| > > I must second that; an MX a lot better in other racing games (CMR 2.0,
| > > N4 demo), but for GPL I think the Voodoo 3 is probably excellent.
| > >
| > > -Gregor
| >
| >
|
|

--
Ian P
<email invalid due to spam>

--
Ian P
<email invalid due to spam>

-Gregor

--
Ian P
<email invalid due to spam>

   I think ONE of the areas that 3dfx ALWAYS lagged behind was their
memory choices.  When the world was trying Vram, they stuck with Dram.  When
SDram began to move, they stuck with Dram.  Finally when ddr became
available
they were just moving to sdram and selling the company.
  dave henrie(still glad to have a V5 btw)