|
Thanks for the responses. There was nothing obvious on the idle
queue. Although no one individual constraint adds much overhead, it
looks like the sheer size of my application caused the aggregate load
to overwhelm the CPU. I've got about 60 complex screens and 50 dynamic
datasets. One screen is visible with the rest hidden. Replacing my
many grids and other complex, dynamically scaled widgets significantly
lowered CPU utilization, but not enough. I wound up splitting the
application into a main screen with several popup overlays. This
removed the CPU burden at the price of a mild performance hit for the
user to raise the external screens. Henry Minsky wrote: There might be some tasks which you set on the idle queue which are sitting around eating cycles, you could put a print tool like this to tell you if there is stuff sitting there |
- [Laszlo-user] Are percentage-based size specifications eval... Charles Watt
- Re: [Laszlo-user] Are percentage-based size specificat... Henry Minsky
- Re: [Laszlo-user] Are percentage-based size specif... Charles Watt
- Re: [Laszlo-user] Are percentage-based size specificat... P T Withington
