Cpu Usage Insane With 1.5 Rc1

machine specs:
Renoise: 1.5 RC1
P4 3.2 gig processor
1 gig ram
External Emagic USB asio device.

I have old songs that will simply not play in 1.5 RC1.
The CPU usage goes to hig to play the track, the audio gets choppy.
I had no problems writing this track with an older version.
Also right now my CPU usage is at 14% when the song is not even playing?

What gives, does anyone else have latency problems?

This should not be happening…
this seems to happen alot with new veresions and latency issues.

I made a comparison between the Renoise 1.25 and 1.5 and it did not show that 1.5 would be slower than the older version.

I think I will make more investigations…

Are you sure you have the same latency settings or sample rate as you did when you used earlier versions of Renoise?

Have you also installed any other programs that run in the background?

Have you upgraded to servicepack2? Could that make Renoise slower?

i do not have SP2 installed.
I may have installed some other programs since last year.
So i should reformat my machine to get renoise to work?

that seems a bit steep.
I have the latency set to 100ms which is the max for the emagic.

this is so annoying! I cant write music like this.

I made a more serious in test now.

And my conclusion is that Renoise 1.5 is less cpu heavy than atleast 1.25!

I turned off most background processes.

I checked both versions so that they hade the same settings.
I made a test song using some vsti’s and some multisamples instruments with instrument envelopes on.
I also used all Renoise internal effects.

I did not look at Renoise built in cpu meter but at how much cpu the Renoise process used on the computer in the activity window.

Now in Renoise 1.5 I even had the nice looking see next pattern effect on and I scrolled in the pattern editor. Also showed some automation envelopes. But still it did not go higher on the cpu than the 1.25…

So I think the reason could be other things…

what system are you using and can you send me the file you used for testing?

The only few differences with line47’s system are i have a PIV 2.8ghz and in internal soundcard.

I don’t experience a 14% cpu usage when nothing is playing. Does it consume 14% when a song is loaded or when it is started blank?
Do you have a template song with default vst-plugins?
What things are running in the background when using Renoise?

CPU consumption should be between 0.5 and 2% when nothing is going on.

i agree,

i have a t23 thinkpad with 1 gig of ram
windows xp pro with aston desktop manager
hardly anything else running when using renoise

and renoise rc1 runs like shit

Just did a couple of tests myself…

Hardware + OS:

  • Fujitsu Lifebook P5020D
  • 1 ghz Centrino M cpu.
  • Hitachi 60GB 7200rpm HDD
  • 1 GB RAM
  • Windows XP Pro w/ SP2
  • Optimised for background services


  • M-Audio Firewire Audiophile 4in/6out
  • ASIO buffer size 1024
  • 44.1khz used in Renoise w/ 26ms latency

Software running:

  • mIRC v6.16
  • AntiVir Guard
  • Notepad
  • FireFox 1.0 viewing Renoise forum :P
  • Renoise v1.281 / v1.5RC1 (separately, not at the same time)

Test #1:

  • 4 instances (not aliases) of FM7 using the “Warm Atmos” preset.
  • Each instance plays a 4-note chord on a separate track in Renoise.
  • (16 notes at once - quite a power chord!)

Test #2:

  • Same as Test #1 except using Vanguard on “PAD Anubis” preset.

Renoise v1.281 results

Using Renoise internal cpu meter:

  • Idle: 0.5%
  • Test #1 (FM7): 46.6%
  • Test #2 (Vanguard): 60.0%

Using Windows Task Manager process cpu meter:

  • Idle: 04%
  • Test #1 (FM7): 55%
  • Test #2 (Vanguard): 67%

Renoise v1.5RC1 results

Using Internal cpu meter:

  • Idle: 0.5%
  • Test #1 (FM7): 44.4%
  • Test #2 (Vanguard): 55.6%

Using Windows Task Manager process cpu meter:

  • Idle: 01%
  • Test #1 (FM7): 60%
  • Test #2 (Vanguard): 64%

Hope this info is useful somehow.

Edit: Updated this post to show both results from Renoise internal cpu meter results and windows task manager results.

Test songs: (v1.281 .rns files)

1.2x is performing better here as well, at least when comparing renoise’s internal CPU load measurements in the upper right corner.

but some other guys have mentioned it before in other related threads: renoise measures the CPU load differently and values from 1.2x are therefore not comparable with 1.5 - at least that’s what i’ve read here and there on this very board.

but to me, 1.5 subjectively feels a bit slower than 1.281, too.

ahh… d’oh… I knew that, but totally forgot about it when I did the tests…


I’ll try things again in a little while.

Edit: new results added to my original post.

I have a slow machine with a decent sound card (m-audio delta 66 with omni io). I have a 1ghz athlon thunderbird 512mb pc133 ram running xp pro with about 30 windows services disabled (and only 8 on automatic) in services.msc and only 3 programs in msconfig startup are even available because of a good cleanup of the run tree in the registry. (Being that my machine is only 1ghz, I do anything I can to speed it up)

After reading this discussion, I made a test song in 1.28 that brought the cpu up to about 85% usage in Windows Task manager. When I load the same song in 1.5rc1 the cpu usage is at 97-99% and I can barely stop the song from playing while the sound cuts in and out.

I’m using asio with the largest latancy setting, using 96000 on both versions.

I’ll test my version of 1.5 RC1 tonight with your test file and let you know what my results are.

I wonder if it could be a chip set issue?

hopefully we can get some good info to present to the development team.

You don’t need my testfile.
I don’t think there is any use in comparing our computers, they are too different.
What you need to compare is two different version of Renoise on your computer.

Dblue: Is there somewhere I get a legally free copy of the VST instrument used in the 2 test songs contained in renoise_test.zip? All I’ve been able to find is “€17.95” for Vanguard–I’m probably just looking in the wrong place. I suppose I could check bittorrent or DC++ sites, but I’d rather get the sounds in some legal fashion.

The songs should work equally well with the demo versions of both VSTs.

Demo of Vanguard can be downloaded here:

Demo of FM7 can be downloaded here:

You will obviously have to manually add the demo VSTs to the Renoise songs, replacing the previous instrument, then choose the correct preset that was used, but it only takes a moment to do that.

I only used these 2 VSTs because I thought they were quite popular and most people would have them. I was originally going to use the free Synth1 VST, but for some reason all my presets have disappeared, hehe. I guess I’ll try re-installing it later.

Everybody’s results will be different though because we all have different systems, different cpus, different amounts of ram, etc. The important thing was to be able to test the same song in different versions of Renoise, to see how they performed, to see which is faster/slower/etc. So, the actual VST being used is not that important in this example.

Thanks for the links. On the fm7 test, which, being that my processor is so slow I can only play one of the tracks (4 note polyphony on that fm7 instrument) at a time, on 1.28 it takes up 87-93% of the processor. In 1.5rc1 it only takes 75-88% of the processor. Quite different results than the test I made.

Heck, if someone wants to test with the stupid random keys (it’s definately not music) test I made using free sampletank 70’s electric piano sound, it’s at: http://www.infraxes.com/0.zip

If you are going to test Renoise in a good way, you (we) need to have a song that also plays one multisamples rni instruments, uses envelopes and also use the DSP effects of Renoise.

Because it could be that for instance the new automation of Renoise 1.5 slows it down. (Which does not at all seam to be the case).

The best thing would be to have the song cleary divided into parts which test the different aspects of Renoise. And then a part that tests them all at the same time in the end.

This way it would show what is slower and what is not slower.

Other things you have to think about when testing is to use the same settings and the same display of track scopes etc.

I have to say that renoise 1.5 is a hell of a lot faster that 1.28 on my machine. PIII 2.8GHZ, 512 DDR, Nvidia 512, EMU APS Soundcard.

If your using XP search (Tweak XP for Audio) on google and follow the steps if you have no aleady, it makes a HUGE diference. Also uninstall any anti virus software for testing. Maybe reinstall Renoise in case things didn’t go well the first time.

Good luck!!

I am now looking for some support from the renoise team to figure out this issue that i am having.

please renoise team contact me so we can figure out what is causing the issues i am having.


Setting up your windows machine is a tricky
piece of work.

If you have your fair share of firewalls viruskillers and whatnots running
on your machine the cpu usage shown in renoise is only the available cpu
looptime given by the system.

What you can do is to get yourself a script that shuts of all other nonessential progs
when you run renoise, or a program that executes this. They are plentiful!

I use a program called RAM Idle pro. It controls the windows core
and frees up looptime on the cpu.
Then you can set priority on all programs with this.
Say you start a program; then all others are set on low. Once you quit this
program (if it was configured this way) everything reverts to normal.
Its a bit work, but if you take your time, it’ll pay back!

Having a superfast machine does not always guarantee a fast result.
Also, tons of “services” that run under the coresystem can be shut down, and then restarted via “sevices”. This is the core controller in windows.
Here you can shutdown start and set any thinkable parameter.

A really fast HD is also smart to have. Say a raptor from seagate og WD
They are getting cheap these days, and they will run upwards of 5 - 25 times faster
than your trusty old 20gig. You install windows and programs on the fastest one.

An example:
If you start windows with all your normal progs, then renoise, and you choose
render file, you only get aprox 50% cpu, even if the window says 99.9%
If you shut down everything else you will notiice that the render process speeds up to ca twice the speed.