|
2004-08-25 22:24
|
SpeedFan Chart.PNG (19,053 bytes) |
|
|
Which workaraound would you suggest? Zeroing stats upon resume from suspend? |
|
|
I would prefer to have the suspended time "missing" from the diagram. So the chart would look like the chart before suspend, followed with the data after suspend. If possible, a vertical line to indicate the suspend event would be nice. (I would like to see the data before and after suspend in the same diagram for comparison.) |
|
|
Similar to 229, I'd suggest to not draw anything when you do not have data. Here, this could mean, as Thomas already suggested, to omit the suspend interval from the chart, and to draw a one-pixel vertical marker instead. |
|
|
I've requested this once or twice myself (privately via email and once on this forum, if memory serves). I got the same reply from Alfredo then too. ;) Depending on how your data is stored and how your graphing code is set up, you might not have a timestamp for each data point. From the way the graphing code responds to various situations (like showing hard drive temps with slower samples with mboard temps), I would assume that you just have a FIFO array and just pass it to the graphing routine at each sample period. The graphing routine is just showing what the data has and does the best it can. I would assume you would have to add more data to keep track of timestamps per sample slice than you do now. The simplest approach might be to offer an option to just clear the entire history when the previous sample point is older than 5 minutes ago. I would even go for that. Just so the graph has something useful during the first 11 minutes of uptime. Then, all you have to store is the last sample point graphed and compare 'now' with that. Just an idea. |
- Anonymous
