REQ Dual \ Quad Core Support

Ideas for improvements and requests for new features in XnView Classic

Moderators: XnTriq, helmut, xnview

ckit
XnThusiast
Posts: 2467
Joined: Tue Feb 17, 2004 1:11 am
Location: Cabarlah, Australia

REQ Dual \ Quad Core Support

Post by ckit »

XnView needs Multithreading support for dual-core and quad-core CPU's!
Also making use of SSE,SSE2 and SSE3 if they available would be nice too!!
Foobar2000, PhotoFiltre, Vivaldi Browser (standalone), Windows Eleven Pro x64 22H2 (MPO disabled), XnView MP and XYplorer Pro
ckit
XnThusiast
Posts: 2467
Joined: Tue Feb 17, 2004 1:11 am
Location: Cabarlah, Australia

Post by ckit »

Added Quad-Core.
Foobar2000, PhotoFiltre, Vivaldi Browser (standalone), Windows Eleven Pro x64 22H2 (MPO disabled), XnView MP and XYplorer Pro
ckit
XnThusiast
Posts: 2467
Joined: Tue Feb 17, 2004 1:11 am
Location: Cabarlah, Australia

Post by ckit »

I'm hoping XnView 2.0 will have multi-threading for All batch operations, the rest of XnView won't benefit from multi-threading unless you're resizing really massive resolutions :)
Foobar2000, PhotoFiltre, Vivaldi Browser (standalone), Windows Eleven Pro x64 22H2 (MPO disabled), XnView MP and XYplorer Pro
Dstruct
Posts: 230
Joined: Fri May 02, 2008 9:27 am

Post by Dstruct »

Mmh, just did a conversion job here and XnView (1.95.4) just made use of one core :(


Would be great if we could specify number of threads in the preferences so that XnView could convert say 4 images at the same time on a quadcore ...
rob
Posts: 7
Joined: Fri Jun 26, 2009 7:35 pm

Re:

Post by rob »

ckit wrote:the rest of XnView won't benefit from multi-threading unless you're resizing really massive resolutions
Sorry but I disagree. Stepping through my 6MP photos on a Core2Duo E6600 could be much faster if the image/jpg loading mechanism would use several cores.
AFAIR JPG would be the perfect candidate for a multi core loading as each jpeg block is independent of the others. Therefore the image could be divided easily into 2/4/8 parts which could be then loaded in parallel.

Rob