| Re: IDLSPEC [message #11689] |
Sat, 09 May 1998 00:00  |
rivers
Messages: 228
Registered: March 1991
|
Senior Member |
|
|
In article <3549019D.ABC78B98@astrosun.tn.cornell.edu>, "J.D. Smith" <jdsmith@astrosun.tn.cornell.edu> writes:
> Visit http://astrosun.tn.cornell.edu/staff/jdsmith/IDLSPEC.html for
> recent IDLSPEC results, along with some graphs of most entrants'
> performances, thanks to Terry Herter. These graphs really make the
> relative performances of various machines quite clear.
I would like to add a caution about the results of the graphics times reported
in this survey. While respondents reported what graphics hardware was in the
system, they did not provide information on what MODE that hardware was running
in, specifically how many bits (8, 16, 32, etc.) or what resolution the display
was set to. I have just found that this makes a HUGE difference on my system.
I have a Pentium II machine with a DIAMOND Stealth 3D 2000 Pro video card.
If I set my display for 1024x768 16777216 colors I get the following results
for graphics_times2:
19.7790=Total Time, 3.3734063=Geometric mean, 4 tests.
However, if I set my display for 1024x768 256 colors I get the following
results for graphics_times2:
3.04400=Total Time, 0.51242809=Geometric mean, 4 tests.
Thus, there is more than a factor of 6 difference between the times reported
depending upon whether I use 8 bit mode or 24 bit mode. Without knowing what
mode the machines were set to in the survey, the results are hard to compare.
____________________________________________________________
Mark Rivers (773) 702-2279 (office)
CARS (773) 702-9951 (secretary)
Univ. of Chicago (773) 702-5454 (FAX)
5640 S. Ellis Ave. (708) 922-0499 (home)
Chicago, IL 60637 rivers@cars.uchicago.edu (e-mail)
or:
Argonne National Laboratory (630) 252-0422 (office)
Building 434A (630) 252-0405 (lab)
9700 South Cass Avenue (630) 252-1713 (beamline)
Argonne, IL 60439 (630) 252-0443 (FAX)
|
|
|
|
| Re: IDLSPEC [message #11773 is a reply to message #11689] |
Wed, 13 May 1998 00:00  |
J.D. Smith
Messages: 214
Registered: August 1996
|
Senior Member |
|
|
Mark Rivers wrote:
>
> In article <3549019D.ABC78B98@astrosun.tn.cornell.edu>, "J.D. Smith" <jdsmith@astrosun.tn.cornell.edu> writes:
>> Visit http://astrosun.tn.cornell.edu/staff/jdsmith/IDLSPEC.html for
>> recent IDLSPEC results, along with some graphs of most entrants'
>> performances, thanks to Terry Herter. These graphs really make the
>> relative performances of various machines quite clear.
>
> I would like to add a caution about the results of the graphics times
> reported in this survey.
> ____________________________________________________________
Quite right, indeed. I had thought I had made that point clear, that
graphics times were dependent on where idl was running (e.g. over a
network X connection), and what type of graphics were running. If I had
thought about it, I would have included this as part of the
questionnaire to begin with, but alas, I did not. General statements
can be made about which machines were faster/slower on the graphics
tests, given that 8-bit is always faster, and likely half of the people
were running 8-bit.... but come to think of it, this might explain why
the Creator 3-D graphics didn't outperform a PC. If the users of these
machines were using another graphics type, and would like to resubmit,
using 8-bit graphics, I'll be happy to update the survey.
JD
--
J.D. Smith |*| WORK: (607) 255-5842
Cornell University Dept. of Astronomy |*| (607) 255-4083
206 Space Sciences Bldg. |*| FAX: (607) 255-5875
Ithaca, NY 14853 |*|
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
