Hi, On Wed, May 2, 2012 at 8:59 AM, Thomas Mueller <muel...@adobe.com> wrote: > I guess the bottleneck is calculating the content hash (SHA-1 message > digest). That's expected. But maybe it's something else.
Another likely CPU consumer is full text indexing, especially if you're dealing with a complex PDF or Office document. The SHA-1 computation time should be a fraction of the IO time needed to copy a binary, whereas full text extraction can at times take much longer than the plain copy. BR, Jukka Zitting