http://permalink.gmane.org/gmane.comp.graphics.qwt.general hourly 1 1901-01-01T00:00+00:00 http://gmane.org/img/gmane-25t.png http://gmane.org http://permalink.gmane.org/gmane.comp.graphics.qwt.general/2442 <pre> The zoomer doesn't know about the situation of the scales and if you don't set the parameter to false it will always trigger a complete replot. Uwe ------------------------------------------------------------------------------ Download Intel&amp;#174; Parallel Studio Eval Try the new software tools for yourself. Speed compiling, find bugs proactively, and fine-tune applications for parallel performance. See why Intel Parallel Studio got high marks during beta. http://p.sf.net/sfu/intel-sw-dev </pre> Uwe Rathmann 2010-03-16T11:35:29 http://permalink.gmane.org/gmane.comp.graphics.qwt.general/2441 <pre> Yes, I too noticed a flickering in that case Yes, but I call it after changing the scales.. Thanks a lot. Giacomo. ------------------------------------------------------------------------------ Download Intel&amp;#174; Parallel Studio Eval Try the new software tools for yourself. Speed compiling, find bugs proactively, and fine-tune applications for parallel performance. See why Intel Parallel Studio got high marks during beta. http://p.sf.net/sfu/intel-sw-dev </pre> Giacomo 2010-03-16T11:24:08 http://permalink.gmane.org/gmane.comp.graphics.qwt.general/2440 <pre> Yes, a known effect with Qt4. The effect depends on the paint engine: f.e when starting refreshtest with "- graphicssystem raster" you won't notice such a difference - because the performance of painting with a pen width of 0 is slow as well. On X11 you can enable the "Paint On Screen" option on the "Canvas" page. On my system this doubles the refresh rate - but is flickering. QwtPlotZoomer::setZoomBase() explicitly triggers a replot before initializing itself from the rectangle of the scales. This replot is necessary to recalculate the scales from previous QwtPlot::setAxisScale calls ( or from autoscaling ). In your situation you can simply call setZoomBase(false) - but this should have been obvious from reading the docs. Uwe ------------------------------------------------------------------------------ Download Intel&amp;#174; Parallel Studio Eval Try the new software tools for yourself. Speed compiling, find bugs proactively, and fine-tune applications for parallel performance. See why Inte</pre> Uwe Rathmann 2010-03-16T08:25:45 http://permalink.gmane.org/gmane.comp.graphics.qwt.general/2439 <pre> I Identified the problem(s): - setting a pen with width &gt; 0 (e.g. 1.0), doubles or even triples the CPU usage; - also the setZoomBase() on the zoomer is CPU consuming. - calling setAxisScale followed by replot(), even if the scale hasn't changed, is CPU consuming. Last point apart, I have not investigated the reason why setZoomBase is so expensive, while the first point is quite annoying. Thanks Giacomo. Elettra Synchrotron Radiation Facility - Control System group. Trieste, Italy. ------------------------------------------------------------------------------ Download Intel&amp;#174; Parallel Studio Eval Try the new software tools for yourself. Speed compiling, find bugs proactively, and fine-tune applications for parallel performance. See why Intel Parallel Studio got high marks during beta. http://p.sf.net/sfu/intel-sw-dev </pre> Giacomo 2010-03-16T07:37:30 http://permalink.gmane.org/gmane.comp.graphics.qwt.general/2438 <pre> Don't think so - and the refreshtest example doesn't do anything tricky ( it paints all points each time ). You don't need to port to 6.0.0 - the performance bottleneck should be in your code or in the way you have configured the Qwt classes. Uwe PS: I read the list - you don't need to send the CC to me. ------------------------------------------------------------------------------ Download Intel&amp;#174; Parallel Studio Eval Try the new software tools for yourself. Speed compiling, find bugs proactively, and fine-tune applications for parallel performance. See why Intel Parallel Studio got high marks during beta. http://p.sf.net/sfu/intel-sw-dev </pre> Uwe Rathmann 2010-03-15T19:24:00 http://permalink.gmane.org/gmane.comp.graphics.qwt.general/2437 <pre> Thanks I tried it out and it runs fine. There are 2 things I would like to ask now: between 5.2.0 and 6.0.0-svn, have any improvements been made in order to optimize this situation, or should I expect the application to run smootly also in the 5.2.0 release? Since I am not able to build refreshtest on 5.2.0, is there any documentation about 6.0.0 available so that I can make a test or porting of my old application to 6.0.0? Is there any place where I can read news about the new release? Thanks for all. Giacomo ------------------------------------------------------------------------------ Download Intel&amp;#174; Parallel Studio Eval Try the new software tools for yourself. Speed compiling, find bugs proactively, and fine-tune applications for parallel performance. See why Intel Parallel Studio got high marks during beta. http://p.sf.net/sfu/intel-sw-dev </pre> Giacomo 2010-03-15T09:28:41 http://permalink.gmane.org/gmane.comp.graphics.qwt.general/2436 <pre>Hi Uwe, There is an American expression, "confusing the map for the terrain". I think this is what is going one here. Many users seem to have a misunderstanding that the QwtPlotCurve is *not* the data, it is a picture of the data that is managed by the QwtPlot. &gt; It would be impossible to derive QwtPlotItem from QObject, &gt; what is planned for a future release. Very nice. So it will be possible for a plot item to emit signals when the mouse comes near a data point, for example? This will make life so much easier. Regards, David _______________________________________________________________ David Stranz, Ph.D.david_stranz&lt; at &gt;MassSpec.com Sierra Analytics, Inc. 5815 Stoddard Road, Suite 601 Modesto, CA 95356 Tel: (209) 545-8508 http://www.massspec.com _______________________________________________________________ Uwe Rathmann wrote: ------------------------------------------------------------------------------ Download Intel&amp;#174; Parallel Studio Eval Try the new software tools for yours</pre> David Stranz 2010-03-14T17:05:52 http://permalink.gmane.org/gmane.comp.graphics.qwt.general/2435 <pre> Nice name dropping, but where do you see MVC here ? If you insist on this design pattern you could declare QwtData as model, but QwtPlot and all plot items ( including QwtPlotCurve) are part of the view. The best match you could find for a plot item is the delegate from the MV design pattern Qt is using. But as you can see from the comparable QGraphicsView design doesn't fit very well. In practice almost everyone implements classes containing View + Controller and Models ( f.e MV design of Qt). But to be honest who needs a design pattern for something so trivial as isolating the data from the widgets ? Nevertheless MVC is the only design pattern on the market - obviously there is not much to say about designing user interfaces. A couple of features would break ( f.e all caches in raster items ) and It would be impossible to derive QwtPlotItem from QObject, what is planned for a future release. In the end it is important to find a design, that is understandable for the application developer. </pre> Uwe Rathmann 2010-03-14T12:21:02 http://permalink.gmane.org/gmane.comp.graphics.qwt.general/2434 <pre>------------------------------------------------------------------------------ Download Intel&amp;#174; Parallel Studio Eval Try the new software tools for yourself. Speed compiling, find bugs proactively, and fine-tune applications for parallel performance. See why Intel Parallel Studio got high marks during beta. http://p.sf.net/sfu/intel-sw-dev_______________________________________________ qwt-interest mailing list qwt-interest&lt; at &gt;lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/qwt-interest </pre> Grissiom 2010-03-14T08:57:56 http://permalink.gmane.org/gmane.comp.graphics.qwt.general/2433 <pre> You can't - but you can implement QwtData objects, that connect the same set of points to different curves. Uwe ------------------------------------------------------------------------------ Download Intel&amp;#174; Parallel Studio Eval Try the new software tools for yourself. Speed compiling, find bugs proactively, and fine-tune applications for parallel performance. See why Intel Parallel Studio got high marks during beta. http://p.sf.net/sfu/intel-sw-dev </pre> Uwe Rathmann 2010-03-13T08:27:06 http://permalink.gmane.org/gmane.comp.graphics.qwt.general/2432 <pre>------------------------------------------------------------------------------ Download Intel&amp;#174; Parallel Studio Eval Try the new software tools for yourself. Speed compiling, find bugs proactively, and fine-tune applications for parallel performance. See why Intel Parallel Studio got high marks during beta. http://p.sf.net/sfu/intel-sw-dev_______________________________________________ qwt-interest mailing list qwt-interest&lt; at &gt;lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/qwt-interest </pre> Grissiom 2010-03-13T04:45:46 http://permalink.gmane.org/gmane.comp.graphics.qwt.general/2431 <pre>Hi Uwe, Hi to all in the mailing list. First of all, is the mailing list still alive? I have stopped receiving mail from qwt-interest since a week. I have a QwtPlot that replots a curve of 1000 points at 10Hz. When the plot window is small, the cpu usage is reasonable. Increasing the application window (the plot area), the CPU goes 100% and the application hangs. I underline that the application CPU usage is low, but X goes 100%. In the case of appending one point at a time, using the low level draw() method and the canvas attributes indicated in the realtime plot example, the performance is much better. Is there any way to improve my plot in the case of full vector update? The mail I sent to the mailing list last friday is attached below. Can someone help with that too? Thanks -------------------------------------------------------------------------------------------------------------------------------------------------------- Hi to all. I have an x axis representing timestamps. I want to divi</pre> Giacomo 2010-03-10T11:21:03 http://permalink.gmane.org/gmane.comp.graphics.qwt.general/2430 <pre>Hi Giacomo, We have implemented that for our TauPlot. It draws the timestamps using human-readable strings that change dynamically depending on the range (e.g., it skips putting the year if you are looking at a minutes range) Have a look at the code for the TauTimeScale class (and its parent FancyScaleDraw). It is python, not C, but you may get an idea: https://tango-cs.svn.sourceforge.net/svnroot/tango-cs/tau/trunk/lib/tau/widget/qwt/scales.py Hope it helps. On Wednesday 10 March 2010 12:21:03 Giacomo wrote: </pre> Carlos Pascual Izarra 2010-03-10T11:41:21 http://permalink.gmane.org/gmane.comp.graphics.qwt.general/2429 <pre>Hi to all. I have an x axis representing timestamps. I want to divide the available space on the axis according to the dimension (in pixels) of the axis scale, and set the number of ticks proportionally to the number of pixels of the scale. I'd like to know which is the right place (also in terms of efficience) to calculate the scale division and apply it. I was also looking for, if someone has already implemented it, an algorithm to nicely distribute the time labels on the x axis. So far, I reimplement QwtScaleDraw::label(double ) virtual function to convert timestamps into a string representing hours mins and secs. Any suggestion on how to divide scale will be appreciated. Thanks Giacomo. ------------------------------------------------------------------------------ Download Intel&amp;#174; Parallel Studio Eval Try the new software tools for yourself. Speed compiling, find bugs proactively, and fine-tune applications for parallel performance. See why Intel Parallel Studio got high marks during beta</pre> Giacomo 2010-03-05T11:12:16 http://permalink.gmane.org/gmane.comp.graphics.qwt.general/2428 <pre>Thanks a lot. I do it with the item provided by the checked signal of the plot. item-&gt;setVisible(on); item-&gt;setItemAttribute(QwtPlotItem::AutoScale, on); Albano On Wed, Mar 3, 2010 at 7:54 AM, Uwe Rathmann &lt;Uwe.Rathmann&lt; at &gt;tigertal.de&gt; wrote: ------------------------------------------------------------------------------ Download Intel&amp;#174; Parallel Studio Eval Try the new software tools for yourself. Speed compiling, find bugs proactively, and fine-tune applications for parallel performance. See why Intel Parallel Studio got high marks during beta. http://p.sf.net/sfu/intel-sw-dev </pre> PaceyIV 2010-03-03T08:35:18 http://permalink.gmane.org/gmane.comp.graphics.qwt.general/2427 <pre> Also do "curve-&gt;setItemAttribute(QwtPlotItem::AutoScale, false);" to remove hidden items from autoscaling. Uwe ------------------------------------------------------------------------------ Download Intel&amp;#174; Parallel Studio Eval Try the new software tools for yourself. Speed compiling, find bugs proactively, and fine-tune applications for parallel performance. See why Intel Parallel Studio got high marks during beta. http://p.sf.net/sfu/intel-sw-dev </pre> Uwe Rathmann 2010-03-03T06:54:56 http://permalink.gmane.org/gmane.comp.graphics.qwt.general/2426 <pre>David, I did try this but still have a frame with properly scaled axes but no plot curve. thanks harold ------------------------------------------------------------------------ Subject: Re: no curve visible From: David Stranz &lt;David_Stranz&lt; at &gt;MassSpec.com&gt; Date: Fri, 26 Feb 2010 12:56:39 -0800 To: List for both Qwt users and developers &lt;qwt-interest&lt; at &gt;lists.sourceforge.net&gt; To: List for both Qwt users and developers &lt;qwt-interest&lt; at &gt;lists.sourceforge.net&gt; Harold, It looks OK. Try adding curve1-&gt;setStyle( QwtPlotCurve::Lines ); Could be the default is QwtPlotCurve::NoCurve. David _______________________________________________________________ David Stranz, Ph.D. david_stranz&lt; at &gt;MassSpec.com Sierra Analytics, Inc. 5815 Stoddard Road, Suite 601 Modesto, CA 95356 Tel: (209) 545-8508 http://www.massspec.com _______________________________________________________________ harold garbeil wrote: ------------------------------------------------------------------------------ Download Intel&amp;#174; Parallel Stu</pre> harold garbeil 2010-03-02T21:30:40 http://permalink.gmane.org/gmane.comp.graphics.qwt.general/2425 <pre>Hi, I've got a plot with some curves and I use the methods indicated in the cpu-plot example for hiding and show the curves when I click on the corresponding legend item. But in this way the axes do not scale according to the range of the visible data. How should I proceed? Thanks, Albano ------------------------------------------------------------------------------ Download Intel&amp;#174; Parallel Studio Eval Try the new software tools for yourself. Speed compiling, find bugs proactively, and fine-tune applications for parallel performance. See why Intel Parallel Studio got high marks during beta. http://p.sf.net/sfu/intel-sw-dev </pre> PaceyIV 2010-03-02T19:49:23 http://permalink.gmane.org/gmane.comp.graphics.qwt.general/2424 <pre> On Mar 1, 2010, at 9:25 PM, Charlie Martin wrote: I can verify that I saw this problem with Qt 4.5.2 but it is gone with Qt 4.6.2! So I hope you are allowed to upgrade Qt. Joey ------------------------------------------------------------------------------ Download Intel&amp;#174; Parallel Studio Eval Try the new software tools for yourself. Speed compiling, find bugs proactively, and fine-tune applications for parallel performance. See why Intel Parallel Studio got high marks during beta. http://p.sf.net/sfu/intel-sw-dev </pre> Joey Mukherjee 2010-03-02T15:17:14 http://permalink.gmane.org/gmane.comp.graphics.qwt.general/2423 <pre> On Feb 27, 2010, at 6:42, Uwe Rathmann wrote: Can anyone else verify if the rubberband zoom (in the bode plot) works on the Mac? I guess I'm a bit lost "on my own". However, if someone could give me some ideas of where to start, I could give it a try. Thanks, Charlie ------------------------------------------------------------------------------ Download Intel&amp;#174; Parallel Studio Eval Try the new software tools for yourself. Speed compiling, find bugs proactively, and fine-tune applications for parallel performance. See why Intel Parallel Studio got high marks during beta. http://p.sf.net/sfu/intel-sw-dev </pre> Charlie Martin 2010-03-02T03:25:28 http://permalink.gmane.org/gmane.comp.graphics.qwt.general/2422 <pre>------------------------------------------------------------------------------ Download Intel&amp;#174; Parallel Studio Eval Try the new software tools for yourself. Speed compiling, find bugs proactively, and fine-tune applications for parallel performance. See why Intel Parallel Studio got high marks during beta. http://p.sf.net/sfu/intel-sw-dev_______________________________________________ qwt-interest mailing list qwt-interest&lt; at &gt;lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/qwt-interest </pre> Vitali Anselm 2010-03-01T12:43:30 Search the mailing list at Gmane query http://search.gmane.org/?group=$group=gmane.comp.graphics.qwt.general