I have been using Windows 10 as my Roon Server platform for a number of years now. Initially it was running on a well spec’d i7 desktop machine but this was over-powered, noisy and power hungry so I replaced it with a NUC 8i3BEH which is sized much more sensibly for my library.
Yesterday I decided to do away with Windows 10 (1903) on the NUC and replace it with ROCK. The migration went smoothly with no issues and I’m up and running again.
However I am noticing two things that are of concern;
TIDAL content is loading more slowly, tracks take longer to start and I am seeing more tracks skipped. At least I have done so far, maybe it will settle down.
The performance indicator in the Signal Path is definitely showing poorer performance on the same tracks versus the Windows 10 core. The same 16/44 tracks that used to show ~32x performance in Windows Signal Path now show ~28x. 24 bit content shows a similar drop, under Windows it used to be 18x, now it is ~12x.
My understanding was that ROCK was specifically optimised to maximise Roon performance, so what could be the issue here? The hardware is absolutely identical in every way, including music storage, so why would the performance be different? I would expect it to be the same as Windows, at least.
You replaced an i7 with an i3 and depending on the generation of the i7 you just removed like 4 additional cores. I dont think they can optimize for less processing power. How big is your library? Make sure you have at least 8gb of ram and try a M.2 SSD drive, much faster than standard.
Given the snippet below from the 1.7 release notes I’d raise a support ticket to ask.
Performance improvements on Roon OS
Starting with Roon 1.7, when running on Roon OS, Roon will use the Microsoft .NET Core Runtime, which offers higher performance and improved memory efficiency.
Library management aspects of Roon typically run 30-40% faster, use less power, and generate less heat. This also enables larger music libraries on the same hardware.
Roon OS has always been the best way to run Roon for most users, but as of this change, we are confident that it is the best option for everyone, regardless of library size. Roon on Roon OS is now more efficient and performs better than Roon on any other platform or product.
Maybe the OP wasnt clear, I was running the NUC 8i3BEH on Windows 10 1903 and then I replaced the OS with ROCK. So I am not comparing an i7 to an i3. I probably should have just left out the first bit about the i7! I just wanted to show that I was familiar with the Core on Windows, as opposed to ROCK which is completely new to me.
@ged_hickman1 It was specifically that 1.7 changelog entry that I was thinking of. I am definitely running 1.7 on ROCK and on all my controls etc.
@Jonathan_Ramos I dont think a reinstall is likely to help since I only set this all up on Roon OS yesterday? I have a windows image backup I can go back to, but with the 1.7 changelog in mind I’d like to understand why it isnt performant in ROCK.
Paging @support do you have anything that might clarify here please?
Yes good thought. I did wait for it to finish before I tried to play anything, it only took a few minutes. So I dont think that was it. I am still seeing the performance difference now, 24 hours later. I know it isnt a huge difference, but it wasnt what I was expecting. I guess I thought there would be parity at least.
I thought about this too, if the Nx value is a measure of how many times faster than normal playback speed (which would be 1x) it is rendering the stream, then it should be directly comparable across systems. That may not be the actual measurement though, its not entirely clear what is it tbh.
Strangely enough I had the opposite issue happen switching from Windows 10 to ROCK on my tiny desktop. When I open the remote on my phone or other devices, its lightning quick to contact the Core and music starts up right away. I have not really paid much attention to performance in the signal path though. That may be the next thing my mild OCD tackles now. My device is hard wired to Ethernet, not sure if that makes a difference.
Since this is exactly the same hardware shouldn’t the signal path performance be comparable in this case? The post from @brian does seem to imply that, if I’m reading it right.
And you’re trying to compare two different operating systems which are using completely different drivers, task schedulers, power management settings, … .
Not necessarily, because it’s a different OS with different drivers etc.
Likewise, environmental factors, or the speed at which your CPU has decided to run today based on energy saving or thermal management rationale may pull the number around, sometimes dramatically. This logic can change with OS or BIOS or Driver updates, too.
If what @brian says is correct, 4x speed means one second of audio is processed in one quarter of a second. So by that measure alone my hardware is performing worse running ROCK than it was running Windows 10.
I can’t see that it is possible to reach any other conclusion.
My issue is that this flies in the face of what we are told is one of the key benefits of ROCK since v1.7, as @ged_hickman1 quoted from the 1.7 release notes
The above claim is the reason I ditched Windows as my core platform after almost three years of use. I could have saved myself the bother!
@Tim_Rhodes I have a similar theme going in my thread about ROCK on NUC8i3 seeming quite sluggish. My experience has been similar to yours - slow navigation, delayed artwork, spotty Tidal performance. I was coming from an i5 iMac which while being faster than the NUC isn’t such a speed demon so as to fully explain the difference. Your experience with ROCK vs windows on the same hardware is raising my suspicion that something isn’t right with ROCK performance…
As far as I know ROCK is not RoonOS. Those optimisations apply only to RoonOS at this juncture, but will be integrated more broadly in a future release.