[mapserver-users] huge performance problem with ECW

[Bart van den Eijnden]

Hi Frank, list,

I think I have tackled the problem. I was requesting an image on the 
verge/margin of the pixel size of the ECW file (1.5 m). If I adjust my 
boundingbox just a little bit I am on the fast side of the medal again. I 
saw some postings on the mailing list from a year ago that mapserver then 
returned a black image when requesting an image with pixels smaller than 
the ECW pixel size.

Question: how can I prevent users from requesting an image through WMS in 
which the pixel size is smaller than the pixel size of the original image? 
And is it a good idea to have this situation in mapserver instead of the 
black image when you get big performance issues instead?

image and bbox info:
x: 120630.915462629 - 121965.084537371 = +-1335 m
y: 487143.164317804 - 487670.835682196 = +-527 m

w: 890 pixels
h: 352 pixels

pixel size 1.5 m.

Best regards,
Bart

On Thu, 03 Jul 2003 17:53:41 -0400, Frank Warmerdam <warmerdam at pobox.com> 
wrote:

> Bart van den Eijnden wrote:
>> Hi all,
>>
>> I have a big problem when I zoom in to a high level on a big ECW file. 
>> The file is 4 Gb. I also have the same file in tiles of approximately 1 
>> Gb.
>>
>> When the boundingbox is (units is meters):
>>
>> bbox=171760,449736,173257,451421 (So approximately 1500 x 1700 m)
>> The image takes less than 1 second to be created and memory usage on the 
>> server of mapserv.exe is 6 Mb.
>>
>> bbox=171760,450736,172257,451421 (So approximately 500 x 700 m)
>> The image takes more than 34 second to be created and memory usage on 
>> the server of mapserv.exe is more than 50 Mb.
>> WITHOUT TILES: almost 100 Mb memory and more than 50 seconds to render 
>> image!!
>>
>> Can anybody help? Is this normal behaviour?
>
> Bart,
>
> This should not happen.  I am not sure how to go about debugging the 
> issue
> though since the file is clearly not easily transported over the network.
> I could try to talk you through some debugging options at your end, or I
> could try to reproduce the problem here by creating a similarly 
> configured
> file.
>
> Could you email me the output of the gdalinfo program run against this
> file?
>
> Best regards,
>



--