Working on a large-scale orchestral composition, with two movements (flows), and I’m generally finding that Dorico is really bogged down and slow to work in. Many times my clicks will take a couple of seconds to respond to. I may see a stopwatch cursor, or occasionally a spinning beach ball. The end result is that this project is taking much, much longer than it should, because Dorico’s constantly not ready to respond to my next command. Does anyone have any tips? I’m attaching a diagnostic report just in case it’s helpful. I’m on an M1 Max MBP, which shouldn’t be a slow machine.
I have the same issue and it seems to be part of the normal course of events. I guess Dorico does a LOT of processing under the hood. I have an 18 flow project in which it can take over a minute to process certain adjustments (with condensing on) and still considerable time with condensing off. [I’ve posted about this before and had good communication with the Dorico team about it, so I’m only offering it for context here.]
Does anyone have any insight into whether it could be worth splitting my multi-movement piece into separate files (i.e. one flow per file) at least for now while working on it? Does that speed things up? Presumably I could recombine later when all three movements are finished.
To answer the minor questions above:
I’m spending 95% of my time in Write mode and rarely using Engrave mode, since I’m still in the note input stage.
I’m not using condensing right now. (Again, I’ll do that after note entry is all done.)
I am using two tabs and it’s unavoidable. That’s because one tab is my MIDI sketches and the other tab is my full orchestral score.
I am using NotePerformer and could try switching to HALion during note entry.
I do use Galley View which is hugely helpful to me right now. I’m hoping to not have to work in Page View during note entry.
It is of course possible to work on one file per flow. At the end of note input you can then easily join them in one file.
Another thing that you can try is to create a full score Layout per flow, and deleting all the other layouts (you can recreate them easily) not sure thought of deleting them is really necessary, but if no layout exists than for sure Dorico will not need to calculate anything on it
So when you open the newly created “full score layout 1” for example, you and Dorico are only working on one flow, reducing the calculations.
Also I experienced a better reaction time when you lock the layout (in engrave mode/graphic editing/lock layout) so the calculations are happening only inside single frames.
As final consideration: I am working on a score with 23 instrument (on 12 players) with 20 flows and 54 minutes duration. (IMac 2020 3,6 i9 2TB SSD 128GB RAM) and Dorico works smoothly like butter, mostly no delay in note input/moving and only a couple of second when changing global options or properties.
@user450 if you like I could test your project on my system (also confidentially if you like to send as private message) I would be very curious how it reacts.
Hey, thanks for the lock layout tip, the difference is noticeable.
I think we need some tips like that for those who want to have a snappy experience while composing and care about engraving only at the end of a project (or not at all). I wonder what other stuff like that I might be missing.
@dspreadbury, I’ve read some posts of yours on this topic that were very helpful, are there any new tips regarding optimizing performance during the writing stage of a project? Like what sort of options we could disable to reduce the amount of calculations Dorico has to make.
Also, I’m curious if your team has identified the best PC build for Dorico or some general rules when building a computer just for Dorico and the performance difference between PC and Mac while using the software.
For example, are M1 processors ahead of current PC CPUs for Dorico usage?
In some calculation aspects of Dorico I think I have understood that they are linear: this means they do not take advantage so much of multicore processing but is most important to have a big single core frequency of the processor. What I am struggling about is to find somewhere clear information about how much are the M Apple silicon processor in frequency (and a comparison in single core frequency, I mean the actual number, between M1, M2 and M3…Then there a re the efficiency vs performance cores, then there are the Pro, Ultra, Mega etc. options… .and all that is bringing more confusion in my mind…). Maybe someone with a clearer mind of mine can give some advices based on actual usage (benchmarks are for YT click-baits IMO…)?
For the moment I am very happy with my i9 iMac (with 3,6 single core frequency), but of course I wish shamefully strongly to have an M Macbook Pro soon
Lock Layout is somewhat helpful, but I’d love to see Dorico go further and do something similar to the Pro Tools Freeze and Commit functions. If I have a project with 18 flows, and I’m working on flow #3, I’m usually not going to care at all about the layout of flow #14 while I’m on #3. If I could Freeze flows 5 through 18 to remove them from all calculations while I’m working on #3, I would imagine that would drastically speed things up. Alternatively, some sort of Freeze or Commit on Layouts could be good too. If I’m working on a score layout, I might not currently care about any of the part layouts, so I could Freeze them and they wouldn’t update at all until I removed the Freeze.
I was able to get Dorico to perform better, although not stellar. Just due to the time crunch, I didn’t isolate each fix to see what specifically was helping speed things up, but for reference, here’s what I tried:
I went to Engrave mode and clicked Lock Layout. In theory this prevents Dorico from constantly recalculating layout changes when I’m creating material in Write mode.
I disabled condensing.
I minimized tabs (I still needed two open – one with a full score and one with a small set of reduced instruments I used in my sketches).
I used Page View instead of Galley View for a bit, although I went back to Galley View and didn’t find much difference, so I don’t think this is a factor.
I switched to HALion for playback (ugh), but then I went back to NotePerformer 4, and didn’t find much difference, so I don’t think this is a factor.
I didn’t actually get around to removing the first flow from the file, so I continued working in a two-flow large orchestral project.
So my hunch is that it was one or both of the first two things in my list above that made it workable.
The M1 runs at 3.2GHz; the M2 at 3.5GHz; the M3 at 4.1GHz. As you say, efficiency cores are slower, somewhere around 2.0 for M1; 2.4 for M2; and c. 3.0 for M3.
(But note that there is not a direct equivalence between an x86_64 CPU at the same speed.)
Benchmarks are not just clickbait: they can be very good indicators of performance. Geekbench rates your iMac i9 at 1655 single and 8204 multi-core performance: the basic 8-core M3 is rated 3003 / 11596 …!!!
Thank you @benwiggy for the useful informations!
(I wonder from where you have them, since my searching was not very productive. Any suggestion where I can find such numbers, also to follow the future development of M series? Thanks)
[EDIT]: ok I found all the informations on geekbench dot com, nevermind…
Wooow! I need to start saving my money for the M3!