CoCreate User Forum  

Go Back   CoCreate User Forum > Applications > CoCreate Modeling

Reply
 
Thread Tools Search this Thread Rate Thread Display Modes
  #46  
Old 01-17-2003, 01:30 AM
clausb's Avatar
clausb clausb is offline
Registered User
 
Join Date: Nov 2002
Posts: 1,168
Hi all,

just sent a new summary to John for posting on the web site.

Fabrizio, your results are indeed kind of skewed since you ran the tests at 800x600 screen resolution. The same is true, by the way, for WuFan's test results for the Compaq Evo laptop. Quite obviously, screen resolution is still a major factor in performance these days. The viewbench test does not check the screen resolution. When we run it here in the lab, we almost always use the following screen resolution settings:

- 1280x1024 pixels (sometimes 1600x1200)
- TrueColor mode

Claus
Reply With Quote
  #47  
Old 01-17-2003, 07:27 AM
John Scheffel's Avatar
John Scheffel John Scheffel is offline
Administrator
 
Join Date: Sep 2002
Location: San Jose, CA
Posts: 1,288
I have posted the updated results from Claus to the web site.

http://www.cocreateusers.org/misc/viewbench_results.html

Note that if you have viewed it before you might have to refresh your browser to see the changes (I did).

Interesting comment about the screen resolution Claus. I used 1280x1024 for my tests, except the Omnibook which only goes to 1024x768. Maybe you should add a column for the screen resolution to the table just so we all know what it was.
__________________
John Scheffel
Reply With Quote
  #48  
Old 01-17-2003, 07:57 AM
clausb's Avatar
clausb clausb is offline
Registered User
 
Join Date: Nov 2002
Posts: 1,168
Quote:
Originally posted by John Scheffel
Maybe you should add a column for the screen resolution to the table just so we all know what it was.
I can only guess. viewbench records the size of the viewport in the result file, but not the screen resolution. The screen resolution is saved into the config file which most people didn't post along with the result files 8-(

Claus
Reply With Quote
  #49  
Old 01-17-2003, 03:15 PM
John Scheffel's Avatar
John Scheffel John Scheffel is offline
Administrator
 
Join Date: Sep 2002
Location: San Jose, CA
Posts: 1,288
Quote:
Originally posted by clausb
I can only guess. viewbench records the size of the viewport in the result file, but not the screen resolution. The screen resolution is saved into the config file which most people didn't post along with the result files.
But wouldn't the size of the 3D viewport be the factor that has the most effect on the result? I would think the pixels outside the viewport wouldn't matter much since they aren't changing. Maybe you could include a column containing the viewport size. It might provide some indication that the difference between two systems may have been influenced by smaller or larger viewports.

I will run some tests on my PC with different Modeling window sizes to see how much it affects the result.
__________________
John Scheffel
Reply With Quote
  #50  
Old 01-17-2003, 05:41 PM
Berney Coleman Berney Coleman is offline
Registered User
 
Join Date: Oct 2002
Location: Camarillo, CA
Posts: 6
I think the viewport size that is recorded by viewbench is wrong. Here's why.

I am using 1600x1200 resolution on my EVO N800W. I can start OSDM and shrink its window size down to about 1024x768 window with a 866x618 viewport inside that. After starting viewbench, it runs winmsd which fails on my system (I'm running XP). When I OK the error message from winmsd, the first real test (many_exams.pkg) resets my OSDM window to 1280x1024 with a viewport of 1211x852. The results in the perf_viewbench_hostname.txt file show the 866.0x618.0 viewport rather than the real value which is 1211x852.

Berney Coleman
CoCreate Consulting Manager
__________________
Berney Coleman
CoCreate Consulting Manager
Reply With Quote
  #51  
Old 01-20-2003, 02:36 AM
fabrizio fabrizio is offline
Registered User
 
Join Date: Jan 2003
Location: Parma - Italy
Posts: 34
about viewport size/screen resolution

I think John is right. In fact i ran the test at 1280x1024 with 16bit color depth and not at 800x600 as Claus guessed... but my Part Browser was really large (I was working with DataMgmt) and this caused the viewport to be too small for the test.
But I agree with John in the sense that we should specify a common resolution and color depth for all tests (as Claus suggested) but also a common viewport size. For example we could say that the test must be run at 1280x1024, 32bit color, OSD window maximized and with PartBrowser, commands menu and any toolbar switched off and the viewport maximized. This should guarantee results with more sense...

Fabrizio
Reply With Quote
  #52  
Old 01-20-2003, 04:48 AM
jmobley jmobley is offline
Junior Member
 
Join Date: Dec 2002
Location: Cary, NC
Posts: 25
Model: HP X4000
CPU: Intel Xeon 2 GHz
RAM: 1 GB
Graphics: ATI Fire GL2 64MB
OS: Microsoft Windows 2000 Professional 5.0 SP 2 Build 2195
Version: OSDM 11.60
Attached Files
File Type: txt perf_viewbench_hpx03685.txt (3.1 KB, 496 views)
Reply With Quote
  #53  
Old 01-20-2003, 05:04 AM
jmobley jmobley is offline
Junior Member
 
Join Date: Dec 2002
Location: Cary, NC
Posts: 25
Model: Premio (custom built)
CPU: Intel Pentium III 870 MHz
RAM: 512 MB
Graphics: NVIDIA Quadro4 550 XGL 64MB
OS: Microsoft Windows 2000 Professional 5.0 SP 2 Build 2195
Version: SD 9.01
Attached Files
File Type: txt perf_viewbench_pre03640.txt (3.1 KB, 479 views)

Last edited by jmobley; 01-20-2003 at 05:08 AM.
Reply With Quote
  #54  
Old 01-20-2003, 05:08 AM
jmobley jmobley is offline
Junior Member
 
Join Date: Dec 2002
Location: Cary, NC
Posts: 25
Model: Premio (custom built)
CPU: Intel Pentium III 650 MHz
RAM: 512 MB
Graphics: NVIDIA Quadro4 550 XGL 64MB
OS: Microsoft Windows NT Ver 4.0 Build 1381 SP 5
Version: SD 9.01
Attached Files
File Type: txt perf_viewbench_pre03606.txt (3.0 KB, 507 views)
Reply With Quote
  #55  
Old 01-20-2003, 05:11 AM
jmobley jmobley is offline
Junior Member
 
Join Date: Dec 2002
Location: Cary, NC
Posts: 25
Model: Premio (custom built)
CPU: Intel Pentium III 600 MHz
RAM: 512 MB
Graphics: NVIDIA Quadro4 550 XGL 64MB
OS: Microsoft Windows NT Ver 4.0 Build 1381 SP 6
Version: SD 9.01
Attached Files
File Type: txt perf_viewbench_pre03599.txt (3.0 KB, 488 views)
Reply With Quote
  #56  
Old 01-20-2003, 05:50 AM
jmobley jmobley is offline
Junior Member
 
Join Date: Dec 2002
Location: Cary, NC
Posts: 25
Model: Premio (custom built)
CPU: Intel Pentium III 650 MHz
RAM: 512 MB
Graphics: NVIDIA Quadro2 MXR 32MB
OS: Microsoft Windows NT Ver 4.0 Build 1381 SP 6
Version: SD 9.01
Attached Files
File Type: txt perf_viewbench_pre03643.txt (3.0 KB, 486 views)
Reply With Quote
  #57  
Old 01-20-2003, 06:05 AM
jmobley jmobley is offline
Junior Member
 
Join Date: Dec 2002
Location: Cary, NC
Posts: 25
(our oldest and slowest)

Model: Premio (custom built)
CPU: Intel Pentium III 600 MHz
RAM: 512 MB
Graphics: ELSA Gloria Synergy 8MB
OS: Microsoft Windows NT Ver 4.0 Build 1381 SP 6
Version: SD 9.01
Attached Files
File Type: txt perf_viewbench_pre02989.txt (3.0 KB, 498 views)
Reply With Quote
  #58  
Old 01-20-2003, 06:18 AM
jmobley jmobley is offline
Junior Member
 
Join Date: Dec 2002
Location: Cary, NC
Posts: 25
Model: HP Vectra VL 400MT
CPU: Intel Pentium III 860 MHz
RAM: 512 MB
Graphics: ELSA ERAZOR X ² 32MB
OS: Microsoft Windows 2000 Professional 5.0 SP 2 Build 2195
Version: SD 9.01
Attached Files
File Type: txt perf_viewbench_hpv03619.txt (3.1 KB, 492 views)
Reply With Quote
  #59  
Old 01-20-2003, 08:01 AM
clausb's Avatar
clausb clausb is offline
Registered User
 
Join Date: Nov 2002
Posts: 1,168
Hi all,

I just posted a new version of the viewbench code to ftp://ftp.cocreate.com/sdtestpackage/viewbench/viewbench_code.zip. The new version first sets a standard size for the viewport and *then* writes the viewport size into the logfile. This was a bug in the old code; thanks to Berney for pointing this out.

When we run the test locally, we always have the configuration file to verify the system configuration, so this was never much of an issue internally. Hope this fix will make it a little easier to correctly interpret the results.

I still hope for a rainy/boring Sunday afternoon to work on improved reporting for system configuration data and test results, but then, I haven't felt bored since a very long time now 8-(

I also updated the result list with all of jmobley's results and sent it to John for publishing it on the web site. Fabrizio, I also removed the note in the table which erroneously claimed that your test was run at 800x600.

Claus

Last edited by clausb; 01-20-2003 at 08:25 AM.
Reply With Quote
  #60  
Old 01-20-2003, 09:55 AM
May Kung May Kung is offline
Registered User
 
Join Date: Oct 2002
Location: San Jose, CA
Posts: 301
Claus, I tried rerunning the benchmark on the problematic system here, the one that kept hanging? There is one process eating between 30 to 50% cpu cycles (it fluctuates quite a bit) called SPOOLSV.exe. I don't recognize it and it won't let me kill it. When I don't have anything running, the normal cpu usage on this process is 0% (like right now).

Hope to run the benchmark on a Unix box or two sometime this week.
Reply With Quote
Reply


Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)
 
Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes Rate This Thread
Rate This Thread:

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT -8. The time now is 01:37 AM.



Hosted by SureServer    Forums   Modeling FAQ   Macro Site   Vendor/Contractors   Software Resellers   CoCreate   Gallery   Home   Board Members   Regional User Groups  By-Laws  

Powered by vBulletin® Version 3.8.4
Copyright ©2000 - 2024, Jelsoft Enterprises Ltd.
You Rated this Thread: