View Single Post
Old 09-02-2014, 09:53 AM   #746
BetterRed
null operator (he/him)
BetterRed ought to be getting tired of karma fortunes by now.BetterRed ought to be getting tired of karma fortunes by now.BetterRed ought to be getting tired of karma fortunes by now.BetterRed ought to be getting tired of karma fortunes by now.BetterRed ought to be getting tired of karma fortunes by now.BetterRed ought to be getting tired of karma fortunes by now.BetterRed ought to be getting tired of karma fortunes by now.BetterRed ought to be getting tired of karma fortunes by now.BetterRed ought to be getting tired of karma fortunes by now.BetterRed ought to be getting tired of karma fortunes by now.BetterRed ought to be getting tired of karma fortunes by now.
 
Posts: 20,582
Karma: 26954694
Join Date: Mar 2012
Location: Sydney Australia
Device: none
Just for the giggle, I used the QC PI to swap last & first names on 137 authors, 181 books, 36 books had videos of 3-5GB - it was done in a second or two as I watched the author folders change before my very own eyes in my file manager. Given all the files were renamed I conclude that the files are moved using hardlinks, either by calibre itself or the file system - if they had been copying it would have taken about 4m40s

I conclude that there is nothing wrong with the QC PI, its working the same today as it was last week, last month... That one user in say 10% of 3.5M (Usage Statistics) has a performance problem, suggests to me that the problem is localised to that user's environment.

Hoods7070 if the drive that your library is sitting on is a FAT 32 drive, then swapping names will take longer because the actual files are copied and renamed. On an NTFS drive its done via a hardlink create/rename and hardlink delete which all happens in the MFT which is preallocated space and contiguous sectors - and hardlinks are very small, 128 bytes rings a bell - MFT entries are not sector bound.

BR

Last edited by BetterRed; 09-02-2014 at 10:02 AM.
BetterRed is online now   Reply With Quote