comp.lang.idl-pvwave archive
Messages from Usenet group comp.lang.idl-pvwave, compiled by Paulo Penteado

Home » Public Forums » archive » IDL Benchmark Anomaly
Show: Today's Messages :: Show Polls :: Message Navigator
E-mail to friend 
Return to the default flat view Create a new topic Submit Reply
IDL Benchmark Anomaly [message #17746] Thu, 11 November 1999 00:00 Go to previous message
MindSpring User is currently offline  MindSpring User
Messages: 5
Registered: December 1998
Junior Member
Greetings,


I saw the IDL benchmark tests and couldn't resist a run through of my
233 Mhz PII laptop
vs my Dual Xeon Linux box. The results were much like I expected.

In TIME_TEST3

Xeon: 7.06824=Total Time, 0.25336439=Geometric mean, 22
tests.
laptop: 28.4500=Total Time, 0.87814925=Geometric mean, 22
tests.

In disc access:

Xeon: 0.346815 1 Write and read 512 by 512 byte array 40 times.
laptop:1.21000 1 Write and read 512 by 512 byte array 40 times.

But in the graphics test the laptop killed the Xeon even though the Xeon
has a 32 Mb
state of the art video card (Nvidia TNT). Any ideas why? I would have
expected it to win this
by two to one not lose it. I'm sure the result is valid as I could
visibly see a difference in the
draw speeds. Has anyone else seen this? Do I need to tune IDL on linux
or my X-Server?

Xeon: 58.8926=Total Time, 1.7327737=Geometric mean, 8
tests.
laptop:21.2100=Total Time, 2.0917758=Geometric mean, 8
tests.

Full tests below.



******** TESTS ********
Running Tests, Please Standby...

% Compiled module: TIME_TEST3.
|TIME_TEST3 performance for IDL 5.2.1L:
| OS_FAMILY=unix, OS=linux, ARCH=x86
| Thu Nov 11 14:18:27 1999
1 0.141545 Empty For loop, 2000000 times
2 0.151196 Call empty procedure (1 param) 100000 times
3 0.181190 Add 200000 integer scalars and store
4 0.171662 50000 scalar loops each of 5 ops, 2 =, 1 if)
5 0.108775 Mult 512 by 512 byte by constant and store, 30
times
6 0.873211 Shift 512 by 512 byte and store, 300 times
7 0.361537 Add constant to 512x512 byte array, 100 times
8 0.341331 Add two 512 by 512 byte arrays and store, 80 times
9 0.347781 Mult 512 by 512 floating by constant, 30 times
10 0.643307 Shift 512 x 512 array, 60 times
11 0.568903 Add two 512 by 512 floating images, 40 times
12 0.185209 Generate 1000000 random numbers
13 0.176392 Invert a 192^2 random matrix
14 0.0710169 LU Decomposition of a 192^2 random matrix
15 0.180727 Transpose 384^2 byte, FOR loops
16 0.264680 Transpose 384^2 byte, row and column ops x 10
17 1.01350 Transpose 384^2 byte, TRANSPOSE function x 100
18 0.195215 Log of 100000 numbers, FOR loop
19 0.375858 Log of 100000 numbers, vector ops 10 times
20 0.369850 131072 point forward plus inverse FFT
21 0.240890 Smooth 512 by 512 byte array, 5x5 boxcar, 10 times
22 0.104464 Smooth 512 by 512 floating array, 5x5 boxcar, 5
times
Skipped read/write test
7.06824=Total Time, 0.25336439=Geometric mean, 22
tests.
% Compiled module: GRAPHICS_TIMES3.
% Compiled module: DIST.
|GRAPHICS_TIMES3 performance for IDL 5.2.1L:
| OS_FAMILY=unix, OS=linux, ARCH=x86
| Thu Nov 11 14:18:34 1999
1 7.15077 Simple plot, 30 times
2 9.06600 1000 vectors x 100
3 37.8072 Polygon filling
4 2.64451 Display 512 x 512 image, 10 times
5 0.705706 Surface 128x128, 2 times
6 1.05937 Shaded surface 128x128, 2 times
7 0.418945 Hershey strings X500
8 0.0400330 Hardware font strings X1000
58.8926=Total Time, 1.7327737=Geometric mean, 8
tests.
|I/O performance
% Compiled module: FILEPATH.
0.346815 1 Write and read 512 by 512 byte array 40 times.

__IS2R+
Thu Nov 11 14:19:34 1999
x86:linux:unix:5.2.1L:Aug 2 1999
TrueColor 24
Dell Inspiron Xeon
550
1000
32
Linux RH 6.0
321284 253364
7361572 1732774
346815
1761436
__IS2R- :w



************************** Laptop Starts Here
************************************

Welcome to IDLSPEC II. For best results, quit all other
running programs before continuing.

******** SYSTEM DESCRIPTION ********
Please answer a few questions concerning your system. If you don't
know the answer or to skip a question, just hit return.

Computer's Make and Processor Type (not speed): DEC HiNote Ultra 2000
PII
Processor Speed, in MHZ (for more processors use, e.g., 4x500): 233
Amount of RAM, in MB: 144
Video Hardware: 4
Any other OS/System Info: Laptop 4Mb Video

******** REVIEW DESCRIPTION ********
Make+Processor: DEC HiNote Ultra 2000 PII
Speed: 233 MHz
Ram: 144 MB
Video Hardware: 4
Other Info: Laptop 4Mb Video

Is this information correct? [Y]/N Y

********* CURRENT DIRECTORY ********
===>C:\RSI\IDL52
If this is not on a fast local drive (e.g. it's over the network)
or is not write accessible, enter another test directory.
(Maybe try C:\) [C:\RSI\IDL52]:

******** TESTS ********
Running Tests, Please Standby...

% Compiled module: TIME_TEST3.
|TIME_TEST3 performance for IDL 5.2.1:
| OS_FAMILY=Windows, OS=Win32, ARCH=x86
| Thu Nov 11 16:29:06 1999
1 0.550000 Empty For loop, 2000000 times
2 0.280000 Call empty procedure (1 param) 100000 times
3 0.270000 Add 200000 integer scalars and store
4 0.220000 50000 scalar loops each of 5 ops, 2 =, 1 if)
5 0.770000 Mult 512 by 512 byte by constant and store, 30
times
6 5.33000 Shift 512 by 512 byte and store, 300 times
7 2.58000 Add constant to 512x512 byte array, 100 times
8 2.30000 Add two 512 by 512 byte arrays and store, 80 times
9 1.54000 Mult 512 by 512 floating by constant, 30 times
10 3.18000 Shift 512 x 512 array, 60 times
11 2.64000 Add two 512 by 512 floating images, 40 times
12 0.440000 Generate 1000000 random numbers
13 0.880000 Invert a 192^2 random matrix
14 0.710000 LU Decomposition of a 192^2 random matrix
15 0.380000 Transpose 384^2 byte, FOR loops
16 0.820000 Transpose 384^2 byte, row and column ops x 10
17 1.49000 Transpose 384^2 byte, TRANSPOSE function x 100
18 0.330000 Log of 100000 numbers, FOR loop
19 0.770000 Log of 100000 numbers, vector ops 10 times
20 1.59000 131072 point forward plus inverse FFT
21 0.990000 Smooth 512 by 512 byte array, 5x5 boxcar, 10 times
22 0.390000 Smooth 512 by 512 floating array, 5x5 boxcar, 5
times
Skipped read/write test
28.4500=Total Time, 0.87814925=Geometric mean, 22
tests.
% Compiled module: GRAPHICS_TIMES3.
% Compiled module: DIST.
|GRAPHICS_TIMES3 performance for IDL 5.2.1:
| OS_FAMILY=Windows, OS=Win32, ARCH=x86
| Thu Nov 11 16:29:35 1999
1 1.10000 Simple plot, 30 times
2 3.02000 1000 vectors x 100
3 6.70000 Polygon filling
4 2.20000 Display 512 x 512 image, 10 times
5 4.01000 Surface 128x128, 2 times
6 1.43000 Shaded surface 128x128, 2 times
7 2.14000 Hershey strings X500
8 0.610000 Hardware font strings X1000
21.2100=Total Time, 2.0917758=Geometric mean, 8
tests.
|I/O performance
% Compiled module: FILEPATH.
1.21000 1 Write and read 512 by 512 byte array 40 times.

To enter this result, email the following text to:

idlspec@astro.cornell.edu

Be sure to include all text between the lines of +'s
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ +++
__IS2R+
Thu Nov 11 16:29:58 1999
x86:Win32:Windows:5.2.1:Jun 4 1999
TrueColor 16
DEC HiNote Ultra 2000 PII
233
144
4
Laptop 4Mb Video
1293182 878149
2651250 2091776
1210000
6419264
__IS2R-
[Message index]
 
Read Message
Read Message
Previous Topic: ATAN inconsistencies
Next Topic: Callable IDL and Cbuilder 3

-=] Back to Top [=-
[ Syndicate this forum (XML) ] [ RSS ] [ PDF ]

Current Time: Wed Oct 08 19:10:03 PDT 2025

Total time taken to generate the page: 0.00439 seconds