Support Board
Date/Time: Wed, 02 Sep 2026 16:31:28 +0000
Post From: Feature-Suggestion.: Use Wavelet-Multiscale-Analysis, O(#bars*ln(#datapoints))-complexity
| [2026-09-02 12:10:59] |
| User408639 - Posts: 83 |
|
Moin Moin. As I am seeing, that changing a bar-size, or loading the data after a fresh start of SC64, or starting a replay is taking some time, seemingly depending upon the amount of data to consider for plotting the chart, may I suggest to use Wavelet-Multiscale-Analysis, in order to plot Charts with a bar-number-based complexity times a correction-term depending upon the data-amount, which is just increasing logarithmically. It is like OLAP for time-series.: 1. You can calculate, which data to retrieve, and 2. do not need the data inside of a pixel/bar. If You are interested, just holler. _Cheers, __Michael. P.S.: I am using that for my SCADA-systems, and they are 50k-times faster for a 1920-Pixel-Plot than just querying a Raw-data-SQL-DB with 1s sample-rate for a time-window over 1.5 years (and about 30% slower, if the plot is for a time-window over 14 years, because of the logarithmic term). |