Thanks! I especially find the last tip on that page very useful
And yes, I do for the most part find Usine quite multicore efficient so no complaint here, but now that we can create huge workspaces in 64 bit Usine, it definately highlights the bottlenecks in my system..
Statistics: Posted by gurulogic — 11 Aug 2014, 08:32
ps 1: If you carefully in MSWord you'll notice that loading a big document is single threaded... I can't do better than Microsoft ...
ps 2: Usine is actually one of the best audio soft for multithreading implementation. ie: loading a patch without glitch while the audio engine is on...
]]>2014-07-21T17:45:29+02:002014-07-21T17:45:29+02:00https://brainmodular.org/forums/viewtopic.php?t=4457&p=30044#p30044 I believe a lot of us has run into bottlenecks before regarding HollyHock and after awhile I noticed graphical performance at some point was more heavily dependent, so quite possibly gurulogic although you have a fast SSD and CPU as I do...maybe you are having some bottleneck performance on the GPU side? Just a thought as I have ran into issues before.
I have several other systems, AIO, tablets, and laptops which have decent processing and ram, however lack on the GPU and in the end my primary studio computer seems to always outperform with the upgraded GPU. Also have you updated to 1.1.008m yet, I have noticed over the last several updates my load and unload times were much faster on my main computer. On the others they still take a while.
-S
Statistics: Posted by sephult — 21 Jul 2014, 17:45
]]>2014-07-21T17:11:32+02:002014-07-21T17:11:32+02:00https://brainmodular.org/forums/viewtopic.php?t=4457&p=30041#p30041 I have been googling for more information about this subject but cannot seem to find much info on the subject. Any suggestions where I can can find info to satisfy my curiosity?
I did find this discussion which is not directly relevant http://stackoverflow.com/questions/3055 ... -windows-7 but a couple of comments made me wonder if something along these lines would be a possible workaround if programmed in Usine?
"A classic case is operating on sequential data in a big data file. Split the file into 4 smaller files, run the program 4 times, and you're done in 15 minutes instead of an hour"
"The most you could do (if relevant in your case) to better use your CPU is to slice your input into N chunks, and run the program multiple time on each chunk, then do something to combine the inputs once they are all done"
Another thing I know I mentioned elsewhere in the forum but I'll repeat, is an option for quick program exit would be nice (apart from the brutal seeming task manager process kill) In my large workspace, it takes about the same time to close HH as it does to load my workspace. In a situation as I experienced first hand last weekend on stage, I setup all my gear on stage, launched HH, launched my workspace and of course as murphy's law would have it, had some strange never before seen issue with my sound card not properly initializing so had to close HH, reboot my computer and lauch HH/workspace again, the whole process taking about 5 min despite my fast OS boot with SSD. Fortunately the DJ playing before me was able to continue to avoid interruption but regardless anything to speed up this process would be greatly welcome. Perhaps in this scenario I should consider a force quit of the Usine process? (would have been setup sooner but it took longer than expected to setup all equipment and get the rear speakers in place for surround, 1hr setup time in total!)
Statistics: Posted by gurulogic — 21 Jul 2014, 17:11
]]>2014-07-20T10:02:59+02:002014-07-20T10:02:59+02:00https://brainmodular.org/forums/viewtopic.php?t=4457&p=30033#p30033Loading and creation of graphical object, plugs, etc, can only be done in the what we call 'the main thread' which is by definition single threaded. I can't do anything to improve this bottleneck. The only solution is to send a feature request to Apple or Microsoft?:D But don't expect miracles before osx 15 or Win 19....
Thanks! I especially find the last tip on that page very useful
And yes, I do for the most part find Usine quite multicore efficient so no complaint here, but now that we can create huge workspaces in 64 bit Usine, it definately highlights the bottlenecks in my system..
Statistics: Posted by gurulogic — 11 Aug 2014, 08:32
ps 1: If you carefully in MSWord you'll notice that loading a big document is single threaded... I can't do better than Microsoft ...
ps 2: Usine is actually one of the best audio soft for multithreading implementation. ie: loading a patch without glitch while the audio engine is on...
]]>2014-07-21T17:45:29+02:002014-07-21T17:45:29+02:00https://brainmodular.org/forums/viewtopic.php?t=4457&p=30044#p30044 I believe a lot of us has run into bottlenecks before regarding HollyHock and after awhile I noticed graphical performance at some point was more heavily dependent, so quite possibly gurulogic although you have a fast SSD and CPU as I do...maybe you are having some bottleneck performance on the GPU side? Just a thought as I have ran into issues before.
I have several other systems, AIO, tablets, and laptops which have decent processing and ram, however lack on the GPU and in the end my primary studio computer seems to always outperform with the upgraded GPU. Also have you updated to 1.1.008m yet, I have noticed over the last several updates my load and unload times were much faster on my main computer. On the others they still take a while.
-S
Statistics: Posted by sephult — 21 Jul 2014, 17:45
]]>2014-07-21T17:11:32+02:002014-07-21T17:11:32+02:00https://brainmodular.org/forums/viewtopic.php?t=4457&p=30041#p30041 I have been googling for more information about this subject but cannot seem to find much info on the subject. Any suggestions where I can can find info to satisfy my curiosity?
I did find this discussion which is not directly relevant http://stackoverflow.com/questions/3055 ... -windows-7 but a couple of comments made me wonder if something along these lines would be a possible workaround if programmed in Usine?
"A classic case is operating on sequential data in a big data file. Split the file into 4 smaller files, run the program 4 times, and you're done in 15 minutes instead of an hour"
"The most you could do (if relevant in your case) to better use your CPU is to slice your input into N chunks, and run the program multiple time on each chunk, then do something to combine the inputs once they are all done"
Another thing I know I mentioned elsewhere in the forum but I'll repeat, is an option for quick program exit would be nice (apart from the brutal seeming task manager process kill) In my large workspace, it takes about the same time to close HH as it does to load my workspace. In a situation as I experienced first hand last weekend on stage, I setup all my gear on stage, launched HH, launched my workspace and of course as murphy's law would have it, had some strange never before seen issue with my sound card not properly initializing so had to close HH, reboot my computer and lauch HH/workspace again, the whole process taking about 5 min despite my fast OS boot with SSD. Fortunately the DJ playing before me was able to continue to avoid interruption but regardless anything to speed up this process would be greatly welcome. Perhaps in this scenario I should consider a force quit of the Usine process? (would have been setup sooner but it took longer than expected to setup all equipment and get the rear speakers in place for surround, 1hr setup time in total!)
Statistics: Posted by gurulogic — 21 Jul 2014, 17:11
]]>2014-07-20T10:02:59+02:002014-07-20T10:02:59+02:00https://brainmodular.org/forums/viewtopic.php?t=4457&p=30033#p30033Loading and creation of graphical object, plugs, etc, can only be done in the what we call 'the main thread' which is by definition single threaded. I can't do anything to improve this bottleneck. The only solution is to send a feature request to Apple or Microsoft?:D But don't expect miracles before osx 15 or Win 19....