» Performance
This site relies heavily on Javascript. You should enable it if you want the full experience. Learn more.

Performance

This page collects information regarding potential optimization of your patches.

Detecting Performance Issues

In order to observe the performance of your patch you have two instruments:

Resolving Performance Issues

  • Learn how to control evaluation of subpatches
  • Hide (Alt+3) patches (not only dock them) to reduce the impact of vvvvs GUI on your CPU
  • Nodes that don't have an input like 'Enable', 'Read', 'Write', ... will see if their input is changed every frame and if so re-perform their task. With large spreadcounts this 'check-for-changed' can be be quite CPU-intesive and often you as a patcher actually know when the change has happened. In those situations use a S+H node before the CPU-intensive operation and only sample its input when you know it has changed.
  • MainLoop (VVVV) has Increase Timinig Precision which is on by default. While causing a smoother framerate in many cases it also causes higher CPU-needs in all cases. You'll have to find find out for your case if you want to change this setting. Here is a good read on the topic Windows Timer Resolution Megawatts wasted.
  • Boost (VVVV) can increase vvvv´s thread priority. In rare cases this may allow you to get the last quantum of performance out of your CPU.
  • Make sure to always use the latest drivers for your graphics card
  • Remove unneeded windows services

anonymous user login

Shoutbox

~1d ago

ggml: what a treat it is to sample the microdevil_intersect help patch from time to time

~6d ago

evvvvil: "FM-2030's messenger": Result of Tuesday's improvised live coding session on Twitch. 106 lines of shader code. https://www.shadertoy.com/view/3slyRX

~7d ago

metrowave: Thanks @Takuma, nice job!

~7d ago

u7angel: thanx cat, i would have missed the streaming...

~8d ago