[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference bulova::decw_jan-89_to_nov-90

Title:DECWINDOWS 26-JAN-89 to 29-NOV-90
Notice:See 1639.0 for VMS V5.3 kit; 2043.0 for 5.4 IFT kit
Moderator:STAR::VATNE
Created:Mon Oct 30 1989
Last Modified:Mon Dec 31 1990
Last Successful Update:Fri Jun 06 1997
Number of topics:3726
Total number of notes:19516

2879.0. "Problem with Rubberbanding...." by VMSDEV::KALER () Tue Jun 05 1990 11:01

  I just heard about a problem that a customer is having and thought I would
 seek advise....


  The customer has some code that implements a "rubber band".  His application
 works fine on a color machine.  If he takes the same code over to a B&W 
 machine, the code works, but it is ungodly slow.

  He has a GC and he sets to XOR mode and sets the forground and background
 to his foreground and background pixel values.  He uses the default plane
 mask.

  If he extracts out the decwindows stuff from his application and runs it,
 it seems to be faster, but inside his application it is slow.

  Now he is running VMS 5.1.  He is planning to upgrade to 5.3 in the near
 future, but he has customers that cannot upgrade.  

  So the questions are:

	Is there something magical about the plane mask that could be causing
	the problem?

	Was a "similar" problem resolved in DecWindows V2?

	Any clues?

   Thanks for any help,

	Christopher
T.RTitleUserPersonal
Name		DateLines
2879.1DECWIN::FISHERPrune Juice: A Warrior's Drink!Wed Jun 06 1990 13:4522
Hmm.  Usually the complaint is with a color system.  Why do you suspect the
plane mask?  If the customer is always using the default, then it should not be
a problem.

However, what s/he *should* be doing is:

In the gc, function to XOR and set the plane mask and the foreground to 

(actual foreground ^ background (i.e. xor them).

This is assuming that there are only two colors on the entire window.  With real
color it is much harder.

But in any case, with an MFB system, the plane mask should not make any
difference, since there is only one plane to begin with, so I don't understand
why there are big performance differences.

Any chance the customer is doing this on an Ultrix system with backing store?
In that case, if the application is asking for backing store, and b.s. has been
activated, the server will be double-drawing, cutting the speed in half.

Burns
2879.2xor modeDECWIN::NORCROSSWed Jun 06 1990 17:3416
    What types of graphics are they trying to display during 
    rubber banding...zero-width-lines, non-zero-width-lines, arcs?
    If they are drawing non-zero-width is it possible that
    they can display them with zero-width while rubberbanding?
    Zero-width is much faster.

    I think the reason why you are seeing a difference in performance
    on monochrome vs. color is that in XOR mode items are first drawn
    to a pixmap.  The pixmap is then used as a stencil to display the 
    image in the window.  On systems such as GPX the copy of the pixmap to 
    the window is supported in the hardware.  On a monochrome system
    it is not supported in the hardware so each scan-line of the
    pixmap must be traversed to see if it has any data to be displayed
    in the window - which is slower.  
      
    -AnnMarie
2879.3clarificationDECWIN::NORCROSSWed Jun 06 1990 18:217
    I just want to clarify that pixmap scheme described in .2
    is used only for wide lines and dashed lines.  For zero-width 
    solid lines the line is drawn directly to the window.  If your 
    customer is already drawing zero-width lines then I'm not sure 
    why there is such a big difference between mfb vs. color.
    
    -AnnMarie
2879.4none of the above.....VMSDEV::KALERThu Jun 07 1990 10:5920
RE .0

	Didn't really think that plane mask would change anything, but I usually
	set in and the customer didn't... so it was worth a try.  But it didn't
	make any difference.

RE ALL

	The hardware is a VaxStation 2000 B&W with 6meg and a VaxStation 3100
	with 8-10m.  Both exibit this behaviour.  Nothing is being drawn except
	your basic zero width line (actually no changes to the line width or
	style of the GC) and nothing else graphically,  just a "little old
	rubber band"

	I find this quite interesting.  Any further help would be greatly 
	appreciated.


	thanks, Christopher