Page 1 of 1
					
				What's the point of 'Use all formats available'=OFF ?
				Posted: Mon Jun 26, 2006 11:24 pm
				by Olivier_G
				I haven't been able to mesure any startup speed/memory difference when turning it ON and OFF (in Options>General>Startup).
Has anyone experienced any improvement by unchecking that option?
...If not, I believe we should remove that setting (ie: permanently ON).
			 
			
					
				Re: What's the point of 'Use all formats available'=OFF ?
				Posted: Tue Jun 27, 2006 6:27 am
				by xnview
				Olivier_G wrote:I haven't been able to mesure any startup speed/memory difference when turning it ON and OFF (in Options>General>Startup).
Has anyone experienced any improvement by unchecking that option?
...If not, I believe we should remove that setting (ie: permanently ON).
Perhaps on old computer? And perhaps less crash?
 
			 
			
					
				
				Posted: Sat Oct 14, 2006 5:35 pm
				by Olivier_G
				I had a deeper look at that option:
-> Saving list is reduced from 63 to 50 filetypes.
-> Performance
(CPU application time and Private memory measured with 
Process Explorer; HD timing measured with 
FileMon, which gives a good estimation of real duration)
OS restart, then lauching XnView Browser in a folder with 36 large Thumbnails of various images, videos (+other files non listed) already in cache (two measurements each):
Code: Select all
              CPU      |    HD    |   RAM
WITH     0.640s-0.687s | 2.3-2.4s | 16 320 KB
WITHOUT  0.640s-0.687s | 2.3-2.4s | 16 320 KB
 
Same folder, but XnView had already been started before:
Code: Select all
           CPU  |   HD  |   RAM
WITH     0.640s | 0.78s | 12 792 KB
WITHOUT  0.608s | 0.86s | 12 792 KB
 
Single XnView cycle:
Code: Select all
start Browser in empty folder
           CPU  |   HD  |   RAM
WITH     0.453s | 0.69s | 7 636 KB
WITHOUT  0.406s | 0.70s | 7 636 KB
enter folder with 2500 photos (6MP 99%JPEG 1%RAW) unCached
          CPU  |   HD    |   RAM
WITH    7mn41s | 11mn16s | 390 824 KB
WITHOUT 7mn42s | 11mn21s | 390 824 KB
re-enter folder cached
          CPU  |  HD   |   RAM
WITH     10.0s | 18.2s | 366 216 KB
WITHOUT  10.0s | 18.3s | 366 216 KB
enter 1st mix folder unCached
          CPU  |  HD   |   RAM
WITH     17.8s | 29.3s | 19 276 KB
WITHOUT  17.8s | 28.4s | 19 836 KB
re-enter folder cached
           CPU  |  HD   |   RAM
WITH     0.312s | 0.27s | 19 808 KB
WITHOUT  0.437s | 0.28s | 20 356 KB
 
Bottom-line: RAM is identical, CPU and duration HD is extremely close and balanced (probably due to small environment differences).
The Save filetypes reduction exists but becomes quite negligeable with Favorite savelist (which should be improved, though).
=> I think this option should be removed (=always on). More opinions?
Edit: I believe Alpha/Beta is an efficient way to test a change (stability...) on a larger scale, without as much commitment as in a final version.
 
			 
			
					
				
				Posted: Sat Nov 18, 2006 8:31 pm
				by Olivier_G
				Not implemented in 1.90 Beta 1...  
(and not a single comment on my lengthy tests...  
 ) 
			 
			
					
				Re: What's the point of 'Use all formats available'=OFF ?
				Posted: Sun Nov 19, 2006 11:15 pm
				by marsh
				Olivier_G wrote:
Has anyone experienced any improvement by unchecking that option?
Yes.  I confirm that program seems more stable.  It also makes some options panels less crowded.  And, you might not want some extensions seen as images? (ex: ".cat").
 
			 
			
					
				
				Posted: Sun Nov 19, 2006 11:15 pm
				by XnTriq
				Olivier_G wrote:=> I think this option should be removed (=always on). More opinions?
I have to confess that I never fully understood the purpose of this setting in the first place. Whenever I install XnView from scratch and go through the configuration process, this proves to be one of the major stumbling blocks for me, because I either don't remember or can't figure out what this is good for. Judging from the posts we get in the “General Support” forum I'm not the only one who gets confused by this option.
I'm convinced that it had its justification at some point, but now that we have the raw numbers in terms of performance I support Olivier's proposal to remove it in Beta #2.