pro_latest v1.3003 - Performance Update (Patreon)
Content
Hi all,
I've just released 1.3003 into pro_latest which I've been working on for the last couple of weeks or so - it offers some huge performance improvements across several core aspects of Tdarr when using it with large libraries (30,000+ files).
For those not in the Discord room, these changes were mainly implemented due to morpheus2n2 having trouble with a 100,000+ file library on Windows whereby the system would lock-up indefinitely, especially when running a 'Scan (Find new)' on a library. Now he can run 10+ file scanners at the same time with the system not locking up which is great. Smaller libraries should be buttery smooth. For fun, I tested a mock 1 million file library and all was working well with the Tdarr queues being built in sub-10 seconds while using 4088MB of RAM (reasonable for a library that ridiculously large). For comparison, building a queue that size on an older version of Tdarr would have taken 10+ mins while locking up the whole system.
With that aside, I can now focus on the worker issues (when using an Output Folder). I'd also like to offer more value for Patreon supporters - think themes, more advanced stats, custom worker style etc. If you have any good/fair ideas then send me a message.
Welcome to the new pledgers and thanks to all for supporting the project!
HGG
- Beta v1.3000 release [27th September 2020]:
- Massively increased performance across application (10* faster queue builder, 10* less DB usage, up to 200* faster file checking on startup, ~20% less RAM, far more responsive UI especially when scanning files)
- Fixed UI bug when creating a backup (backup size showing much lower than it should be)
- Fixed folder watch bug (error saying txt file doesn't exist)
- Moved scheduled hourly 'Scan (Find new)' to its own option instead of being enabled when 'Scan on start' enabled.
- Beta v1.3001 release [1st October 2020]:
- Massively reduced system load when running 'Scan (Find New)' or when folder watcher adding lots of files (~150* faster scanner prep time with 100,000 file library)
- Fixed DB initialisation memory bug on startup if 30,000+ files
- Beta v1.3002 release [3rd October 2020]:
- Fix DB initialisation bug
- Beta v1.3003 release [3rd October 2020]:
- Added options to prioritise transcodes or health checks
- Added option to ignore schedules