5/13/2010 3:29 PM | |
Joined: 1/17/2007 Last visit: 12/5/2024 Posts: 1548 Rating: (538) |
Have you read the help file? I have attched it for you. I originally thought that the cycle time would increase / decrease depending on the number of connections and the traffic going over these connections. Looking at the help file, suggests that this is the case up to a point. The "Scan Cycle Load from Communication" setting sets the maximum amount of CPU time provided for handling communications jobs. If the required amount of CPU time to process the communications jobs is less that this setting, then the remainder will be used for processing the runtime code. This is the way I interpret the help file: If all communications jobs cannot be processed by amount of CPU time allocated to communications processing in a cycle (20% with the default setting), the communications jobs will need additional cycles to complete. So the CPU loading for communications will vary, but can never be greater than this setting. So this setting must be chosen carefully so that there is enough time alloacted to process the worst case communications load. Not too sure how you would go about trying to work out this setting. I suppose you would have to find out how fast the AS functions process the data (bits/bytes per second) and multiply that by the amount of data for each connection. Then add all these up for each connection. This would give you a time value. Then this could be compared with the required minimum scan time, and the % needed for communications could be ascertained. AttachmentScan Cycle Load from Communication.pdf (438 Downloads) |
Programming today is the race between software engineers building bigger and better idiot proof programs, and the universe producing bigger and better idiots. |
|
This contribution was helpful to3 thankful Users |
5/14/2010 3:12 AM | |
Posts: 118 Rating: (4) |
Thank You Very Much smiffy. Its nice explanation to get some good idea. I will try to calculate and set the CPU properties for communication load. |
Follow us on