Osirus/Ostirus: high idle usage

[borkfomb]

Hello, the idle cpu usage for these emus is very high.

Osirus 1.3.14 CLAP

• Idle: 4.80%
• Active: 5.25% (default patch playing a chord)

Ostirus 1.3.15 CLAP

• Idle: 8.60%
• Active: 10.15% (default patch playing a chord)

In comparison,

• TAL-J-8, 0.10%
• DIVA, 0.40%
• FB-7999, 0.25%
• SURGE XT, 0.30%
• PAINOTEQ, 0.35%
• SPIRE, 0.10%
• KONTAKT with a 70gb library, 2.40%

Latency is high too but I'm sure that's known.
REAPER 7.22 x64 Windows

[grawlinson]

None of the synths that you’ve listed are emulating a full blown DSP chip, though.

[borkfomb]

The entire u-he and fullbucket vst collection (over 50 plugins) uses less cpu in total than 1 instance of Ostirus, a straight clone of a 30 year old dsp, and the response is deflection? Ok, good luck

[dddux]

These true hardware emulations are working quite differently to all these VST synths you mentioned. These have to run DSP chip emulations in the background (1,2 or more chips), and emulating Motorola 563xx chips cost a lot of CPU in comparison to a normal VSTi that doesn't need to run a DSP chip emu, to run the native DSP code. The only way to optimise these plugins further is to optimise the 563xx chip emulation, which has been done many times already. Maybe they can squeeze out a bit more, but they would still use far more CPU than normal VSTi.

In Reaper, there is an option to mute the track and turn off the VSTi. That helps with these plugins as the CPU usage drops almost to zero. Mute has to be on and record button on the track turned off. There's an option for this in Reaper settings if you have it turned off.

[insolace]

Native code will always outperform emulation on different architecture, 24/31bit addressing.Completely different instruction sets.On Sep 9, 2024, at 7:06 AM, borkfomb ***@***.***> wrote: The entire u-he and fullbucket vst collection (over 50 plugins) uses less cpu in total than 1 instance of Ostirus, a straight clone of a 30 year old dsp, and the response is deflection? Ok, good luck —Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>