Optimizing FlexPro for Large Volumes of Data
FlexPro combines a high level of ease of use with high performance in a unique way and is ideal for a wide range of applications. In some areas, however, a higher level of ease of use and robustness can only be achieved when there is a loss of data throughput. FlexPro solves this problem by offering you different settings that make it possible for you to optimize the application for ease of use and data throughput. The default FlexPro settings are optimized for the highest level of ease of use and robustness. Please note the following if you want to optimize FlexPro for large volumes of data or for high data throughput.
Project Database Format and Operating Mode
The standard way that FlexPro operates is to work with a temporary copy of the FlexPro project database and to save it first where you specify by using the Save Project Database command. When working with large volumes of data, you should create the project database and save it in the In several folders and files storage format at the preferred location on your hard drive before actually starting to work with it, then open it in For direct editing mode:
1.Start FlexPro to obtain an empty project database.
2.Click on File > Save As.
3.In the Save As dialog box, select the hard drive folder in which the project database is to be saved.
Note: You should save the project database to a local hard drive and not in a folder on your network.
4.Now enter the desired name in the File Name field.
5.Select In several folders and files as the storage format and then click Save.
6.Now click File > Open.
7.In the Open dialog box click on Browse and select the project database that you just saved.
Note: The project database file extension is .fpd. This same folder contains a subfolder named the same, but with the file extension .db. This folder will later contain all objects stored as files.
8.For the Mode select Direct.
The project database has now been created and can be used.
Deactivating Global Link Correction When Renaming Objects
If you rename an object in a FlexPro project database, FlexPro searches the entire project database for objects that have references to this object in order to correct them. This can take a lot of time if the project database is large.
1.Click File > Options.
2.Click on the Operation tab.
3.Uncheck the option Search the complete project database when looking for references. FlexPro then searches only folders in which the renamed object is located.
Deactivating the Preview
The FlexPro Preview option may be deactivated to increase the data throughput. To do this, click Close in the title bar of the preview window.
Note: You can use View[Task Window] > Show > Preview to reactivate the Preview any time.
Preventing Recalculation of Images and Calculated Results
FlexPro can save images of diagrams and tables as well as calculated results of formulas in the project database. These are immediately available and do not have to be recalculated when you re-open the project database.
1.Click File > Options.
2.In the Options dialog box, switch to the Project Database tab.
3.Select the options Save calculated images of presentation objects in the project database and Save calculated results of formulas in the project database.
Data Import
If you want to import data into the FlexPro project database, you should only select the necessary data sets and import them in Copy mode. If you are working with equidistantly sampled signals, you should import them as signals.
1.Click Data[Import from Files and Databases] > Binary Data.
2.In the Import dialog box in the Channel selection field select the option Manually and in the Data field select the option Copy. Under Data structure select the option Signals.
3.Click Open to import the data.
Using Cursors for Large Data Sets
If you would like to use cursors for several channels of a large measurement, for instance, you can either use a single diagram with stacked Y axes or a worksheet object with one plane per diagram. The worksheet provides more options for zooming and scrolling the individual channels independently.
1.Select all of the data sets for which you want to use cursors.
2.Click on Insert[Containers] > Worksheet and click on the icon with the desired window layout.
The cursors are placed when positioning the mouse on the point closest to the position clicked. To do this, the distances for all XY value pairs for the curve are calculated, which causes a delay when working with large volumes of data. Therefore, you should switch to a mode that accounts for only the selected X position.
1.Click on the diagram for which you want to accelerate positioning using the mouse.
2.Right-click with your mouse on the worksheet and select Properties.
3.In the Properties dialog box click on the Cursor Settings (Active Diagram) tab.
4.In the Cursor positioning field select the option Use X position.
5.Repeat these steps for any additional diagrams.
When zooming to sections, you should select Spread X Range Between Cursors with rubber banding as the preferred method.
1.Click on the right side of the desired section to place the leading cursor at this position.
2.Press the TAB key to activate the origin cursor.
3.Click on the left side of the desired section.
4.Select Cursor Tools/Cursors[Zoom Curve] > Spread.
Note: You should select the smallest section as possible. Subsequent scroll operations will then be faster. To scroll, you should use the arrow on the scrollbar for shorter moves. You should only use the slider on the scrollbar for long jumps on the curve.
Optimizing the Analysis
You should set up and test your analysis smaller data sets and then apply the prepared analysis to a large volume of data. For instance, you can import a relatively small data set, use FlexPro's Reduce function to decimate the data, or create a section of the data using the Index operator.
You can optimize the amount of run-time environment memory on your computer required for the FPScript analysis language.
1.Click on File > Options.
2.Click on the System Settings tab.
3.Under Maximum size of data sets in memory enter a value in megabytes that is higher than the size of the data sets that you are using. If you have, for instance, data sets with 10,000,000 64-bit floating point values, you should enter a value that is at least than 80 megabytes (1 megabyte = 1,048,576 bytes), since 8 bytes are needed for each 64-bit floating point value.
4.Under Maximum memory allocation for data sets enter a value that equals no more than 70% of the computer's available memory.
5.Close FlexPro.
6.Now open Windows Task Manager by right-clicking on the taskbar and selecting Task Manager.
7.Now restart FlexPro and carry out a typical data analysis.
8.Follow the page file usage on the Performance tab of the Task Manager. The value should be displayed as the total value in the Physical Memory field and should not exceed more than 20% of the memory available on your computer.
9.If the memory requirement limit is exceeded significantly, you should set the value in the Maximum memory allocation for data sets field back to a lower value.
Additional Optimization Options
If you have considered all of the points mentioned above, you can increase FlexPro's data throughput considerably. However, before you start analyzing and managing large volumes of data, you should check to see if the amount of data cannot be reduced when you acquire it. You should set the sampling rate for the measurement in accordance to the bandwidth of the signals to be measured and not to the capabilities of the measurement system. If your system has a good anti-aliasing filter, it generally does not make much sense to oversample the signal considerably. FlexPro also has no trouble jointly processing and displaying data with different sampling rates. You can therefore also capture slower signals, e.g., temperatures, with a lower sampling rate than faster signals.
Additional options for optimization are available when you select the appropriate graphic format. For instance, for a 3D data set, a waterfall chart is much faster than a surface chart. You should also always take into account whether the data must be presented in its full resolution or if the data can be decimated in advance.
See Also