Mark, Thanks for the response, I greatly appreciate it. Unfortunately at this point in time switching to something else would be a huge headache and frankly I don't have the time. If I had to stick with it, what would you recommend in terms of getting the most performance out of it as a I can. Here is the exact scenario I am dealing with: I have two windows: a main window and a second window. Both windows have multiple graphs and tables displaying real-time data (coming from a queue from another process that is doing all the calculations). The main window also has two GLViewWidgets that are displaying GLMeshItems that are being translated (This is not an issue as it is a constant cost per frame and no matter what, they only moves once per frame). Here are my questions: 1. Would it be better to have each window in its own process? 2. Are there any under the hood ways I can use the QPath's to make appending data faster? I have tried doing some things similar to this http://stackoverflow.com/questions/17103698/plotting-large-arrays-in-pyqtgraph 3. I have noticed an append function in the source for the 'CurveDataItem', but its simply an empty function. Was there/is there a plan for implementing that and if so how would that be done? Thanks, Paul
On Tuesday, February 28, 2017 at 1:17:38 PM UTC-6, Paul Gross wrote: > > Hi, I am working on displaying real-time telemetry data using pyqtgraph. I > am quite pleased with the visual results however I am having issues with > the frame rate dropping as more data is plotted. I am receiving about 100 > data points per second. At the beginning plotting is quite fast but the > frame right dives rapidly as more data is being displayed. > > > I have fixed the size of the plots and auto-ranging is disabled, I have > tried down sampling and it helps, but not quite enough. I am looking for > the fastest way to plot a large amount of data points in real-time, as I > receive them. > > > I have found other posts about speed and some of them reference > ‘arrayToQPath’, but I am not sure if that is the best way to approach my > issue. My biggest concern is with plotting the data points as quickly as > possible as to not slow down the rest of the event loop. The current way > that I am plotting data in the ‘PlotWidget’ is setting the data of a > ‘PlotDataItem’ in the following manner, where the arguments are either a > numpy array or python list: > ‘plot_data_item.curve.setData(data['time'][self.x_min:], > data[data_type][self.x_min:])’. I have tried plotting a new curve each > time, but that seemed to be quite slow. In an ideal world I would just be > able to append points without having to connect all of the previous points > to each other. At some point I want to reset the plot to an empty plot, > but that really isn’t an issue because this is a rare occurrence in my > program. I simply want to display in real time several minutes of data with > multiple different plots, each of fixed size with no scrolling, panning, or > auto-resizing. > > I would not be opposed to threading some of the work out if that is an > option on top of any suggestions you have. Something I am not entirely > clear on is if the GIL is an issue when it comes to QThreads since they are > C++? Will I get any performance boost by using QThreads in the update > function? I am reading from a data queue in the update function, so ideally > that would be taken out of the loop. That however isn’t my main concern > because if I reset the plots (change the range of data that is taken from > my data list/array the frame right jumps right back up). > > > I am using the Dev Branch because my application requires PYQt5 and python > 3.5. > > Any suggestions you have would be greatly appreciated! > -- You received this message because you are subscribed to the Google Groups "pyqtgraph" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/pyqtgraph/4b83454c-9c7f-4887-ace9-7d21dc92ef94%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
