The good:
Exiting MM5 is a much cleaner process. When exiting normally (no bug was generated during the session), all background MM processes shown in task manager vanish cleanly and quickly (within a second or two). When exiting MM4 (4.1.14.1813) normally, it routinely takes 1:20 or longer for MM processes to end in task manager.
The current MM5 beta booted up in about the same time as MM4 (around 4 secs most times)
Needs improvement:
More memory-intensive processes were slower in the beta. My benchmark was to load a complex smart playlist. The scheme has many nested smart playlists within playlists. In MM4, the complex playlist takes around 30-40 secs to produce the output. In MM5, it varied widely, but the same playlist had a range between 1:03-2:30.
Test were performed with a clean boot of Win 10 and 8GB RAM with an Intel G3258 processor.
A few thoughts on the current Alpha (build 2057)
Moderator: Gurus
A few thoughts on the current Alpha (build 2057)
Last edited by jcfogerty on Fri Jan 13, 2017 10:32 am, edited 1 time in total.
Re: A few thoughts on the current beta
3:15 for my complex AutoPlaylist in MM5 (1.15 in MM4) on first load after starting MM. .28/.38 (MM5/MM4) the second time around.
Download MediaMonkey | License
Help: Knowledge Base | MediaMonkey for Windows 5 | MediaMonkey for Android
Lowlander (MediaMonkey user since 2003)
Help: Knowledge Base | MediaMonkey for Windows 5 | MediaMonkey for Android
Lowlander (MediaMonkey user since 2003)
Re: A few thoughts on the current beta
I thought the slower times may have been due to indexing not being complete, but I used MM5 for more than an hour before doing these tests. My library is 13,000 tracks.
My smart playlist scheme is memory intensive for both MM4 and MM5, but I found a clear difference in performance, at least for this particular function. Loading "all tracks" in both versions took around the same amount of time.
My smart playlist scheme is memory intensive for both MM4 and MM5, but I found a clear difference in performance, at least for this particular function. Loading "all tracks" in both versions took around the same amount of time.
Re: A few thoughts on the current Alpha (build 2057)
Hi,
the MM5 start time will be significantly improved in 2058 (there was actually a regression that was slowing it).
Re the auto-playlist load, we are going to verify it. This also sounds like a regression, probably related to sorting changes (auto-playlists are now sorted natively and not by SQL).
the MM5 start time will be significantly improved in 2058 (there was actually a regression that was slowing it).
Re the auto-playlist load, we are going to verify it. This also sounds like a regression, probably related to sorting changes (auto-playlists are now sorted natively and not by SQL).