You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
It takes several seconds to rasterize a ~100-slide deck. There is probably some low-hanging performance optimization fruit still:
Rasterization happens in a background thread, but only one thread
Rasterization starts over from files every time
It is tempting to display the first slide before they are all done rasterizing, but this would introduce the possibility of navigating to a slide that is not ready to show yet, which is the whole thing we're trying to avoid (vs. Preview).
The text was updated successfully, but these errors were encountered:
It looks like pdf2image has a lot of constant overhead, so shelling out one page at a time is inefficient. Boosted performance a little with chunking. Also was busy-waiting on the queue too much.
Probably this is about the best we can do with pdf2image. The next step would be using a pdf parser library so we can only pay the read/parse penalty once, and rasterize the same data structure at different sizes. Not sure if Poppler does this or not, but probably.
It takes several seconds to rasterize a ~100-slide deck. There is probably some low-hanging performance optimization fruit still:
It is tempting to display the first slide before they are all done rasterizing, but this would introduce the possibility of navigating to a slide that is not ready to show yet, which is the whole thing we're trying to avoid (vs. Preview).
The text was updated successfully, but these errors were encountered: