Sorry for the error in reporting the ExtremePC link......the "e" in front of
the word "extreme" should not be there (guess my fingers wanted to type it
correct anyway): www.xtremepcuk.com/guides/gfg/index.shtml
To answer a few questions posted in one place:
1. The "speed settings" for D3D are designed to run consistent scores on
3DMark200x (which runs in D3D, not OpenGL). In addition, you would not want
to use these settings for actual "game play" since they do not produce a
quality 3D graphic, they are designed for speed..... not image quality.
2. When benchmarking N4, I use the same settings on the video card as I use
for running N4....which for all Nvidia cards should be OpenGL. I would not
find any purpose to benchmark my computers in D3D for N4...since I wouldn't
ever run N4 in D3D in the first place. The picture quality is substantially
lower than the OpenGL API....along with frame rates in the 50 % range of the
OpenGL frame rates on the same system. "Your mileage may vary."
3. I have not experienced any change in the GPL image quality or frame
rate....and wouldn't have expected any really (the FPS in GPL is maxed at 36
FPS....both of my "race boxes" already max out at 36 FPS....with plenty of
head room I would expect). I have not had any crashing in GPL, since the
new D3D rasterizer was released. And, although I have only had these
21.81's installed for a few days.....there has been no crashes in the
approximate 2 hours I have used them in GPL so far........and I wouldn't
expect any. I did test them for the presence of the OpenGL "black screen"
bug in GPL.....its still there.
4. I too was surprised and suspicious of the frame rate jump on the PIII
800. But, I have used the "static" N4 benchtest ever since N4 was released,
and I take care not to alter the video card settings from test to test.
However, I do not actually use the PIII 800 for sim racing.....so it is
possible the settings on a the video card were changed accidentally......but
I doubt it since I always check them before running the benchtest. And,
certainly....I do not have Vsync turned on for any of the
benchtests....that's too obvious of a mistake to not catch.....and I
specifically check for this since a few Windows programs somehow manage to
change the Vsync settings from time to time (I have no clue how....but I
have seen my Vsync reset before....and I hadn't reset it). I provided the
PIII 800 benchtest results because I have them recorded anyway.........for
anyone's use here on r.a.s. that has a similar computer. The only
benchtests I really care about are on my race box....which is System #2 in
my report.
5. You can not use benchtest scores to make direct comparisons between
video cards only (or video card drivers only) on otherwise different
computers. You must compare "apples to apples" and that would include the
cpu platform and ops system, along with ram and chipset (including chipset
drivers). The benchtests are valuable to me, because they tell me what and
how my computer's performance changed due to the addition of the new video
drivers (all else being identical to previous benchtest runs). You can not
use them for that same purpose....to compare new drivers on my system....to
what new drivers on your system would perform like....not in terms of direct
comparison of test scores. What you can ascertain.....is that if a new
driver increases frame rates on my system, and yours is similar, there's a
good chance it will do the same on yours. That's all you can ascertain.
I've had several email asking what Nvidia "utility" I use for making the
card and screens settings. I have been using Riva Turner for about six
months.......and like it better than NvMax....but the difference isn't
totally compelling.....its subtle. You can find either of these and almost
any of the Nvidia reference driver sites.
I hope this provides a bit more detail for those who wanted it. I didn't
originally include any of this stuff since in a way, that was kind of an
"off topic" post. So I wanted to limit the dialogue on it.
Regards,
Tom
