rec.autos.simulators

A person named Patrick sent me a benchmark method
he had used to evaluate his system with ICR2. I have
evaluated mine using the same method. I used VERY
conservative observations and maybe others would wish
to use this benchmark and share their findings.

In my opinion, anything above 20FPS is very raceable.
I had a 486DX2-66 until yesterday when my wife gave me
a plugin Intel OverDrive Pentium 83MHZ replacement
for the 486. Retail $299. I don't have a benchmark of
mine prior to the new processor, but it was VERY
driveable with some lowered detail in SVGA.

The benchmark method and the results for Patrick and
I follow:

================ from Patrick =======================

Hi,

  With the arrival of Indycar Racing 2, I am sure that
a lot of people will be considering an upgrade to their
system. I think I have found a pretty good way to measure
and compare the frame rates on different systems and for
different options and detail levels.

  Here's how it works: in the options/graphics panel,
set all textures to either ON or OFF (no AUTO). The
only exception: skids/paint. Set this one to AUTO.
Now determine a frame rate that you want to test.
Set both the min and max to that same value.

  Now go racing. If the skids/paint are on, that means
that your system is able to handle the textures you
have set to ON (including skids/paint) at the frame
rate you have selected. If the skids/paint continuously
toggle ON/OFF this means that the system can't achieve
that frame rate with skids/paint ON, but it can do it
if skids/paint is OFF. Consider this as meaning that
the system can't handle that frame rate. Try a lower
frame rate. I think skids/paint is a good one to use
since it is almost always visible and pretty much
everyone wants it ON anyways (I might be wrong?)

  I suggest to always use Michigan for the comparison,
because it is avaible both on the demo and in Nascar,
so we can even compare ICR2 and Nascar...

  Go in Race mode. Make sure you have cars behind you.
Get close to the cars ahead so that their textures are
drawn. Make a complete lap. If the skids/paint is ON
more that 50% of the time then that frame rate is
reached. Try a higher frame rate. If the skids/paint
are off, try a lower frame rate. Repeat until you find
the highest frame rate for a particular set of options,
number of opponents and textures.

If the different columns don't line up etc.,try a
different font.

============================================================

  < patrick >

  486DX2-66
  ATI Graphics Express Video
  UniVBE?

HARDWARE                      OPPONENTS GRAPHICS    FPS
CPU          VIDEO CARD G  S  N A B H  123456789  VGA  SVGA
------- --------------- -  -- - - - -  ---------  ---  -----
DX2-66  ATI Gr.Express. I  D 31 8 3 2  ..XXXXXXX <10     ?
DX2-66  ATI Gr.Express. I  F 15 4 1 1  .. .....X  17     ?
DX2-66  ATI Gr.Express. I  D  P        XXXXXXXXX  14     ?
DX2-66  ATI Gr.Express. I  D  P        ..XXXXXXX  18     ?
DX2-66  ATI Gr.Express. I  F  P        .. .....X   ?   <10

=============================================================

  Intel OverDrive(83MHZ) replacement  for my old 486DX2-66.
  (probably equivalent to a Pentium 60, due to motherboard.)

  Spider Tarantula Video (S3 Vision 964-4MB VRAM) VLB.
  16MB/256K Cache on NoName Mother Board
  UniVBE enabled in the demo.
  My normal race setting is #6-SVGA-23/fps.

HARDWARE                      OPPONENTS GRAPHICS    FPS
CPU          VIDEO CARD G  S  N A B H  123456789  VGA  SVGA
------- --------------- -  -- - - - -  ---------  ---  -----
OvDr-83 Spider Taran 64 I  D 31 8 3 2  ..XXXXXXX  24    16
OvDr-83 Spider Taran 64 I  D 15 4 1 1  .. .....X  30max 30max
OvDr-83 Spider Taran 64 I  D  P        XXXXXXXXX  30max 16
OvDr-83 Spider Taran 64 I  D  P        ..XXXXXXX  30max 25
OvDr-83 Spider Taran 64 I  D  P        .. .....X  30max 22
OvDr-83 Spider Taran 64 I  D 31 6 3 1  ..X...XXX   ?    23
OvDr-83 Spider Taran 64 I  D 31 6 3 1  XXXXXXXXX   ?    12

=============================================================

What it all means:
G: Game; I=Indycar 2, N=Nascar
S: Sound mode; D=digital, F=FM (-f option)
N: Number of RACE opponents; P=pre-season testing
A: Drawn ahead
B: Drawn behind
H: Heard
Graphics: X = ON, . = OFF
1: Grass texture
2: Asphalt texture
3: Object texture; blank = No objects
4: Grandstand texture
5: Wall texture
6: Horizon texture
7: Car texture
8: Smoke/Dirt texture
9: Skids/Paint texture
FPS: Highest reachable frame rate

================================================================
 SPAT upon the WIRES by:The Right Reverend Ed <R.REVV.ED>
 Elder Elder of the Pegram,Tennessee SOLO CLENCH  "Bob" IS Dobbs!

           Visit the MOTHER of ALL WEB SITES:
       http://www.racesimcentral.net/:80/subgenius/slack.html
================================================================

 I ran the benchmark test as you outlined on my Dell 133, 24 meg EDO
RAM, and a #9 FX motion 771 video card with the following results.
All test run in SVGA and sims set up as you had in #6 & 7.

                                                 24 FPS Nascar Tallad.

Does not seem this computer runs any faster Frame rates
than some of the others to me.  Maybe a Matrox Mill. Vid Card
would do better?????? Any way all test done Exactly As you outlined
and very conservative.  Any sign of blinking on or off, I went to the
next lower FR and ran stabilized.  Thanks for the Info on test.  

My dx2/66
16MB
Stealth32 2MB DRAM
with univbe (adds no more than 1 fps)
with digital (FM adds approx 1 fps)
SB Pro (same result with GUS)

 same                   I  D  P        .. ...X.X   ?   12
 same                   I  D  P        .. .X...X   ?   12*

*cuts skid/paint only when passing the grandstand.

Pentium 120 MHZ, ASUS P54/TP4 motherboard
        512K L2 Async Cache, 32 MB RAM.
Diamond Stealth 64 Video 3200 VRAM 2 MB.
Pro Audio Spectrum 16 sound card

With UniVBE enabled, it showed anywhere from 2 to 5 FPS increase.  The
least increase was apparent when the road texture was on.  This seems
to take a lot of CPU cycles to put the texture to the screen.

HARDWARE                      OPPONENTS GRAPHICS    FPS
CPU          VIDEO CARD G  S  N A B H  123456789  VGA  SVGA
------- --------------- -  -- - - - -  ---------  ---  -----
P120            DS64-U  I  D 31 8 3 2  XXXXXXXXX  30    18
P120            DS64-U  I  D 31 6 3 2  XXXXXXXXX        20
P120            DS64-U  I  D 31 8 3 2  ..X...XXX        30
P120            DS64-U  I  D 31 8 3 2  .XX...XXX        22
P120            DS64-U  I  D 15 4 1 1  XXXXXXXXX        19
P120            DS64-U  I  D 15 4 1 1  ........X        30
P120            DS64-U  I  D 15 4 1 1  ..XXXXXXX        30
P120            DS64-U  I  D 15 4 1 1  .XX.X.XXX        24

BTW, I just picked up the release version from Electronic Boutique.
The scenery around Elkart Lake is so nice, its almost difficult to
watch the road.  Sure is nice to have 40 gallons show up on the dash.

A really great racing Sim.

Dave.

least increase was apparent when the road texture was on.  This seems
to take a lot of CPU cycles to put the texture to the screen. <<

How do I know how many FPS I'm getting?

I've included the original message below incase anyone missed it and
wishes to try it.

It is accurate to within 1 FPS depending on whether you allow any on
or off painting of the skids to occur.

Pentium 120 MHZ, ASUS P54/TP4 motherboard
        512K L2 Async Cache, 32 MB RAM.
Diamond Stealth 64 Video 3200 VRAM 2 MB.
Pro Audio Spectrum 16 sound card

UniVBE was enabled. It showed anywhere from 2 to 5 FPS increase.  The
least increase was apparent when the road texture was on.  This seems
to take a lot of CPU cycles to put the texture to the screen.

The FPS numbers are derived by increasing the min/max FPS setting
until the skid marks were not on continuously, then decrementing by 1
FPS.

HARDWARE                      OPPONENTS GRAPHICS    FPS
CPU          VIDEO CARD G  S  N A B H  123456789  VGA  SVGA
------- --------------- -  -- - - - -  ---------  ---  -----
P120            DS64-U  I  D 31 8 3 2  XXXXXXXXX  30    17
P120            DS64-U  I  D 31 6 3 2  XXXXXXXXX        19
P120            DS64-U  I  D 31 8 3 2  ..X...XXX        30
P120            DS64-U  I  D 31 8 3 2  .XX...XXX        21
P120            DS64-U  I  D 15 4 1 1  XXXXXXXXX        19
P120            DS64-U  I  D 15 4 1 1  ........X        30
P120            DS64-U  I  D 15 4 1 1  ..XXXXXXX        30
P120            DS64-U  I  D 15 4 1 1  .XX.X.XXX        23

David Schick.