Sean Meighan
Welcome => Latest xLights/Nutcracker software => Topic started by: sean on August 15, 2017, 06:12:23 PM
-
Release 2017.23 of xLights
Developers keep cranking. Be sure to thank them!
Latest releases are found at www.xlights.org (http://www.xlights.org)
Videos tutorials are found at videos.xlights.org (http://videos.xlights.org)
Forums are located at nutcracker123.com/forum (http://nutcracker123.com/forum)
Issue Tracker is found here: www.github.com/smeighan/xLights/issues (http://www.github.com/smeighan/xLights/issues)
XLIGHTS/NUTCRACKER RELEASE NOTES:
---------------------------------
2017.23 Aug 15, 2017
-- enh (dkulp) Add ability to specify starting location and direction for circle model
-- enh (keith) Add ability to manual edit effect timing - right click on an effect to access it
-- bug (keith) Fix model group panel no longer to able to repeatedly add/remove models using button
-- bug (keith) Fix scheduler does not properly restore from an interrupted schedule
-- bug (keith) When adding timing tracks only add them to master + current view
-- bug (keith) Fix some drag and drop anomalies on display elements
-- bug (keith) Add some SDL logging to track down OSX audio not playing bug
-- enh (keith) Fine tune AC UI ... try to make it more intuitive
-- bug (keith) Fix handling of expansion boards on the falcon
-- enh (keith) Allow AC sequencing on nodes
-- enh (keith) Some check sequence additions
-- bug (keith) Fix model rename breaks submodels
-
WOW! I don't know what was done been v20 to 23 but the rendering is super fast. I was frustrated at .20 so I backed up to .17 and it was still slow. I went from 2+ minutes down to 17 sec. Thanks developers.
-
Yes. I just upgraded from 2016.10 to 2017.23....I LOVE what you did with the scheduler! Pulling it out of the main program and allowing it to run on its own has been great! Only suggestion I would have is allow the "schedule" to be sorted by date and time rather than playlist. As rando1957 said- rendering is super fast and over all performance seems to be "smoother" and more efficient - especially on an older machine. GREAT JOB!
Sent from my iPhone using Tapatalk
-
I looked at all the checkins and nothing was really done to make it render faster... This is like when the singer tells the sound guy to turn her up in the mix and he spins the knob the light on the board and then she says that's much better...
-
I looked at all the checkins and nothing was really done to make it render faster... This is like when the singer tells the sound guy to turn her up in the mix and he spins the knob the light on the board and then she says that's much better...
Shh.... quiet!!! If the users say it's running faster you are supposed to say something along the lines of "yea... we worked really hard, toiled for hours to fine tune some of the algorithms..." :)
-
I removed the make it seem slow function.
-
Gil; he is going from 2016.10 to 2017.23. Did you maybe think 2017.10?
he went from March 2016, to Aug 2017.
I do believe we have had speed improvements in the last 18 months. the pinwheel effect has the new and old algorithm.
Here are some i saw from the release notes:
-- enh (dkulp) Optimise blur - speeds rendering
-- enh (dkulp) Optimize the redraws on timer events a bit.
Drops CPU usage from 95% to about 60% on my machine.
-- enh (dkulp) Optimizations for OpenGL for WaveView and model/house previews to pass all vertices/colors/textures in one shot
-- bug (dkulp) Bunch of optimizations, more gl error logging, some fixes based on the error log
-- enh (cjd) Rewrite pinwheel effect to handle larger buffer sizes cleanly
-- enh (keith) Refactor output processing. This should generally be invisible other than some performance enhancements and
-- bug (keith) Performance fix for animated GIFs. Example: "This is Halloween" went from 300 second save to 65 seconds.
-- enh (dkulp) Start all jobs that don't depend on anything above them first, then start the rest. 30% performance improvement (helps performance, particularly when multiple previews are shown)
-- bug (dkulp) If using pixel style "Solid Circle" or "Blended Circle", performance with OpenGL 2/3 was really bad.
-- enh (keith) Move audio file loading to its own thread to improve audio load time in scheduler. Eliminates the need to load and convert the entire file before the audio starts to play.
We now over 64bit versions of xlights. In 18 months there are probably some other speed improvements.
-
Well I was looking at the first reply that said he went from v20 to v23 and rendering became super fast.
-
sorry gil. i mixed up threads. a person updated from 2016.10 to 2017.23 in anothe rtrhead. next time i will read the thread again.
'
i agree, nothing obvious should have sped up peoples sequence save in the last few weeks.
-
Well you were right the second reply was the user going from .10 to .23.
-
So it must have been the install. In that case, I would love to thank myself for such a great install. Your install skills are outstanding and looking forward to more installs you perform in the future. Keep up the good work! ;D ;D
-
Atta boy Randy. Your the man.
-
It was me - I went from 10 to 23. The later versions of 2016 for some reason would not run well on my older computer. (I was having memory issues and a bunch of other issues - I cannot pin everything on xlights) but what ever you guys did - it appears to solved the issues I was having with the later versions from late last year.
As I run tests here in the Christmas light show testing room, I am not having any issues ( other than learning what all the "new" buttons are for). Once I am pleased with how it is looking (I am tweaking as I am testing) I will move my test boards down to the garage and hook up the controllers and see how it does.
Again so far so good and GREAT JOB on the improvements!
Sent from my iPhone using Tapatalk
-
Just found the AC mode - loving it! My whole display is AC - I have no pixels, etc (maybe in the future) and was wondering how I would do a "twinkle" like I have done in one of those other programs, this mode looks to solve this issue. Again GREAT job...thanks for all the hard work on this!
Sent from my iPhone using Tapatalk