In the latest Picard running on El Capitan, the memory usage is off the charts. I did choose my whole media/music folder but have tried similar in other products. Right now after almost a day of processing, it’s up to 8GB of RAM. Last time it took all of the application memory and locked up the machine. Anybody else experience this?
I don’t specifically monitor it, but Picard isn’t really intended to be run on an entire library at once, so I don’t.
Perhaps it should just have some built-in limit on the number of files you can add or scan at a time? Until the server load is reduced, you’re going to be mostly refreshing matched albums anyway.
I would expect the memory to be big if you are pulling tens of thousands of album cover artwork into memory.
Don’t load up so much stuff Picard was not designed for that.
Wasn’t aware it couldn’t handle it. For the record, the memory could be managed and released but if the product isn’t designed to ever be used this way…I hear you. Wasn’t aware you had to do things piecemeal.
Part of the issue is that the UI is not really great for thousands of files. If you need to review the changes Picard wants to make and for some difficult albums Picard scattered the tracks over multiple releases it really gets hard to spot and correct this if the list is too long. I personally do only 20-30 albums at once and for difficult compilations I often load only 2 or 3, but that’s pretty personal figure. Some people might easily deal with a few hundred albums, but there is definitely an upper limit to what makes sense. Another point is the loading speed since Picard is limited by the web service request limit enforced by the MB server.
Regarding the memory usage: Yep, it is rather high. This is partially caused by the fact that Picard keeps the metadata for each file and each loaded track in memory, and usually has multiple copies so it can show you the changes and revert changes made. Also cover art from files and the web is kept in memory. Together with all the structure the UI needs this sums up pretty quickly. There are for sure strategies that could be applied to lower this, but since this wouldn’t fix the fact that Picard is not really meant to be used this way nobody has put effort into this.
Unfortunately there are frequently users who try to load huge collections at once and get frustrated. Not sure how to best deal with this, maybe Picard should display a warning
120,000 Songs. Program crashes