williamcopper
Max Output Level: -68 dBFS
- Total Posts : 1120
- Joined: 2014/11/03 09:22:03
- Location: Virginia, USA
- Status: offline
speed optimization idea?
It takes a long time to load sonar filled with VSTi instruments. Generally, that's probably not Sonar's fault ... however would it be feasible to NOT show the name of every single instrument as it is loaded? On my system, it's all Kontakt, and there are approximately 5 x 16 x 40 = 3200 instruments loaded (kontakt .nki files, organized in kontakt .nkb banks). While the names flash by very quickly, still that's a lot of display output for no benefit at all -- they go by too fast to read easily, but it takes a long time to put it all on screen. This might be a Kontakt function about which Sonar has no control - don't know.
|
John
Forum Host
- Total Posts : 30467
- Joined: 2003/11/06 11:53:17
- Status: offline
Re: speed optimization idea?
2015/12/23 20:59:27
(permalink)
I may not have noticed those names but I don't recall seeing them. What I have seen is the loading of samples that go very fast as a rule. Then I may have up 10 instruments in Kontakt nowhere near the number you are loading.
|
williamcopper
Max Output Level: -68 dBFS
- Total Posts : 1120
- Joined: 2014/11/03 09:22:03
- Location: Virginia, USA
- Status: offline
Re: speed optimization idea?
2015/12/23 21:16:33
(permalink)
I imagine loading a few goes so fast you don't even see the screen. Maybe the way I configure instruments is out of the ordinary -- there tend to be very many lightweight instruments so that articulations can change very frequently depending on the music.
|
williamcopper
Max Output Level: -68 dBFS
- Total Posts : 1120
- Joined: 2014/11/03 09:22:03
- Location: Virginia, USA
- Status: offline
Re: speed optimization idea?
2015/12/23 21:18:48
(permalink)
Besides, if this were improved it would get me more seldom on this forum ... the 5-10 minutes of loading time are too long for making a pot of coffee, so I have to do something.
post edited by williamcopper - 2015/12/23 21:39:07
|
charlyg
Max Output Level: -72 dBFS
- Total Posts : 950
- Joined: 2015/02/27 22:10:47
- Location: West Hills, CA
- Status: offline
Re: speed optimization idea?
2015/12/23 21:49:46
(permalink)
All hail the block tool.....
|
sharke
Max Output Level: 0 dBFS
- Total Posts : 13933
- Joined: 2012/08/03 00:13:00
- Location: NYC
- Status: offline
Re: speed optimization idea?
2015/12/23 22:07:49
(permalink)
I don't think that showing the Kontakt loading screens in any way slows down the loading of projects.
JamesWindows 10, Sonar SPlat (64-bit), Intel i7-4930K, 32GB RAM, RME Babyface, AKAI MPK Mini, Roland A-800 Pro, Focusrite VRM Box, Komplete 10 Ultimate, 2012 American Telecaster!
|
Anderton
Max Output Level: 0 dBFS
- Total Posts : 14070
- Joined: 2003/11/06 14:02:03
- Status: offline
Re: speed optimization idea?
2015/12/23 23:08:20
(permalink)
The time required to load instruments comes from pulling the samples (probably at least several hundred megabytes) from where they're stored and either a) shuttling them into RAM if you stream from RAM, or b) separating just the attacks and shuttling them into RAM if you stream from hard drive. The display is just running in parallel with the loading process.
|
bitman
Max Output Level: -34 dBFS
- Total Posts : 4105
- Joined: 2003/11/06 14:11:54
- Location: Keystone Colorado
- Status: offline
Re: speed optimization idea?
2015/12/23 23:19:13
(permalink)
Anderton The time required to load instruments comes from pulling the samples (probably at least several hundred megabytes) from where they're stored and either a) shuttling them into RAM if you stream from RAM, or b) separating just the attacks and shuttling them into RAM if you stream from hard drive. The display is just running in parallel with the loading process.
And displaying strings is negligible anyway in terms of overall processing time.
|
JonD
Max Output Level: -39 dBFS
- Total Posts : 3617
- Joined: 2003/12/09 11:09:10
- Location: East of Santa Monica
- Status: offline
Re: speed optimization idea?
2015/12/23 23:30:07
(permalink)
Move your most often-used libs to a solid state drive. Otherwise, be prepared to take a lot of coffee breaks.
SonarPlat/CWbBL, Win 10 Pro, i7 2600K, Asus P8Z68 Deluxe, 16GB DDR3, Radeon HD5450, TC Electronic Impact Twin, Kawai MP11 Piano, Event ALP Monitors, Beyerdynamic DT770 Pro, Too Many Plugins, My lucky hat.
|
Sidroe
Max Output Level: -55.5 dBFS
- Total Posts : 1954
- Joined: 2010/11/10 18:59:43
- Location: Macon,Georgia
- Status: offline
Re: speed optimization idea?
2015/12/24 07:44:55
(permalink)
As discussed in the IK forum, the new Miroslav 2 library now takes 5-6 minutes to load just the keyswitching versions of the string sections! OUCH! The same project with the old Miroslav library loads in about 10-15 seconds! A deal breaker if you rely on loading and un-loading those librarys on stage. Maybe we can change the way live performance is done!!! Instead of a 15-20 minute break at the end of a set we could just take a 5 minute break between songs while the next song loads. It could average out for the night for the same amount of time if we're lucky!!! LOL!
Sonar Platinum, Sonar X3e, Sonar X2a , Sonar X1 Expanded and 8.5.3 (32 and 64 bit), Windows 10 on a Toshiba P75-A7200 Laptop with i7 @ 2.4 quad and 8 gigs of RAM and secondary WD 1 Tb drive, Windows 10 desktop, Asus i5 @ 3.2 quad, 12 gigs RAM, 1 Tb drive, 1 500 gig drive, MOTU 24io, 2 Roland Studio Captures, Saffire 6 USB for laptop, Soundtracs Topaz Project 8 mixer, Alesis Monitor 2s, Event BAS 20/20s, Roland Micro-Monitor BA-8s, and 45 years worth of collecting FX, Mics, Amps, Guitars, and Keyboards!
|
williamcopper
Max Output Level: -68 dBFS
- Total Posts : 1120
- Joined: 2014/11/03 09:22:03
- Location: Virginia, USA
- Status: offline
Re: speed optimization idea?
2015/12/24 11:38:20
(permalink)
Just for fun I timed loading up Sonar with approximately 40 GB of samples, all from an SSD drive and presumably into memory only (?), and copying approximately the same amount of data from an SSD drive to a regular SATA drive. It took sonar just under 7 minutes to load; completing the copy took just under 6 minutes. Now clearly writing to a regular drive is going to be way slower than reading from an SSD drive ... so ... what's taking all that time with sonar? I've got a suspicion: Sonar, or kontakt, or Windows, is writing all that loaded data to a working file located somewhere on my system disk (regular SATA). I don't believe it's going to the paging file, it is set at a rather low size on the SSD manufacturer's recommendation (basically that the system should not use a paging file): 800 MB on a system with 64 Gb of memory. (edit) -- confirmed this last, by changing to a very large paging file entirely on an SSDD drive --- no significant change in load time.
post edited by williamcopper - 2015/12/24 11:58:28
|
williamcopper
Max Output Level: -68 dBFS
- Total Posts : 1120
- Joined: 2014/11/03 09:22:03
- Location: Virginia, USA
- Status: offline
Re: speed optimization idea?
2015/12/24 12:23:12
(permalink)
nm .. bad lead
post edited by williamcopper - 2015/12/24 12:38:05
|
Anderton
Max Output Level: 0 dBFS
- Total Posts : 14070
- Joined: 2003/11/06 14:02:03
- Status: offline
Re: speed optimization idea?
2015/12/24 15:05:25
(permalink)
williamcopper It took sonar just under 7 minutes to load; completing the copy took just under 6 minutes.
Loading requires parsing the data - e.g., assigning to keyboard ranges and MIDI channels, separating attacks if you're streaming from disk, etc. A straight copy is just a bit-for-bit file transfer, no intelligence required.
|
williamcopper
Max Output Level: -68 dBFS
- Total Posts : 1120
- Joined: 2014/11/03 09:22:03
- Location: Virginia, USA
- Status: offline
Re: speed optimization idea?
2015/12/25 08:52:40
(permalink)
Here's another idea: some sorting routine used by the software is using, by default, a temporary disk file. That could well be Kontakt but it might be Sonar too. Nothing in memory could take that long, no matter how much parsing is required.
|
Anderton
Max Output Level: 0 dBFS
- Total Posts : 14070
- Joined: 2003/11/06 14:02:03
- Status: offline
Re: speed optimization idea?
2015/12/25 11:04:00
(permalink)
☄ Helpfulby charlyg 2015/12/25 12:06:40
From storagereview.com: An SSD does not have a mechanical arm to read and write data, it instead relies on an embedded processor (or “brain”) called a controller to perform a bunch of operations related to reading and writing data. The controller is a very important factor in determining the speed of the SSD. Decisions it makes related to how to store, retrieve, cache and clean up data can determine the overall speed of the drive. We won’t get into the nitty-gritty details for the various tasks it performs such as error correction, read and write caching, encryption, and garbage collection to name a few. Yet, suffice to say, good controller technology is often what separates an excellent SSD from a good one.
|