- August 14, 2006
- 253
- 16
- Home Country
-
Aruba
- Thread starter
- #21
This should work yes, but only the events for MP Server, the other events are just for the computer on wich it is installed.
Aruba
Aruba
I've notice a couple of problems - maybe related. These were with COMswitch, so I updated to your LEDcontroller version and - at least the first issue still remains (not had time to see the second since)...
First, the resource use of COMswitch (and now LEDcontroller) seems very high, especially the CPU usage. See the attached Task Manager snapshot - in my case CPU continually hovers around 2-4% all the time! Is this because it's continually polling for state changes? If so, is it possible to maybe add a config value to extend the polling interval? I don't mind if the LED takes an extra 5 secs to turn on/off[As a side note, the memory usage seems high too, but not so important IMO]
Second, I get .NET error popups occasionally. From the details (see second snapshot), it looks like COMswitch was polling something and the net link was down. This tends to happen on wake from standby (maybe a race condition?), but I'm thinking a reduce poll might improve or cure this too. Note: this is ugly more than anything else, if I click "Continue" on the popup, most times everything seems to work fine.
Thanks!
The CPU usage is probably because it indeed checks every second, I will make an update so that it would be configurable, will also fix the error ASAP
Wich are the vents you are using?
I've fixe the bug and added your request.
The actual CPU usage isn't something I can reduce, but I have it running quite some time and never experienced nasty side affects so....
Aruba
I'll look into it, the added CPU usage is probably because the extra check that happens because of your bug when no network was avalable.
But with the time setting, if you set it to 5 sec, isn't the CPU usage high every 5 sec or still continiously?
Aruba
I'll look into it, the added CPU usage is probably because the extra check that happens because of your bug when no network was avalable.
But with the time setting, if you set it to 5 sec, isn't the CPU usage high every 5 sec or still continiously?
Usage stays high continuously - hovering around 6-8% for one event. Similarly, for 3 events it stays around 31% (plus or minus ~2%) all the time. Key thing is that the usage of the TVserver also goes up and stays up in the same circumstances - leading me to think that the polling is actually much higher than the configured 5 secs.
One other datapoint (though I can't see that it's relevant) is that I also have For the Record installed.
<EDIT> I snapshot the process monitor graph for a period of time (attached). The first three horizontal graph-segments (about 5 secs each) are with 3 events configured, the next 6-7 are with just 2 events config'd, and the last 3 segments are with just 1 event. The left hand core history shows there's a pretty consistent usage within these diff configured periods (one vertical segment = 20% utilisation of a single core).
hi,
would love to use your plugin, but my only com port is used for my momolight.
so, is there a chance that you add the use of the lpt port? Just select the port you want in the config.
Look at first page, this should solve your problems