Performance issue of WPF application when .NET 4.5 is installed - wpf

We upgraded our development machines and with it, we installed .NET 4.5
Right after that, we noticed our application, which is developed with .NET 4.0 became noticeably slower. This is without touching the source code or what version of .NET the application is compiled against. The problem is, that the application itself is pretty complex and separating possible cause is impossible. The possible causes might include:
General WPF slowness (not probable)
Telerik library (same as above)
WriteableBitmap, which is used heavily even when user doesn't make any actions (yes, SetDirty is used properly)
WCF used to receive "large" amount (~5kB/s) on constant basis through TCP
I'm asking if there is any relevant information about this issue. I was trying to google, but I only get performance tips and no issue that would fit mine. Right now, the workaround is to uninstall 4.5 and install 4.0, which fixes the performance problem.

So, I figured it out.
In the current code, Dispatcher.Invoke is used to synchronize changes from the network queue to the UI. This is called many times per second. To fix this, I used CompositionTarget.Render event with double-buffer to push the network queue to the UI and it fixed the performance problems.
It seems that .NET 4.5 did some changes to the Dispatcher, so it became a bottleneck in our use case.

Related

Mcafee update blocked silverlight and the future of the Silverlight Technology

Today, while I'm running a Silverlight project in the Internet Explorer by pressing F5 in Visual Studio 2012 in my Windows 8 machine, I found that McAfee started to block Silverlight XAPs (Which is loaded by Prism).
This leads me to think again about the future of Silverlight. I'm at the beginning of my LOB Application. Should I stop what I like to work in Silverlight and return back to WPF. I which that I can continue to develop in Silverlight until Windows 8 becomes rich like Silverlight. That is why I limit the Model and MVVM to PCL to be easier to be ported to WinRT in the future. Using await async and so...
Please advise me which is better for a LOB application that will run in three countries throw the Internet. Should I continue in Silverlight for zero deployment, or work in WPF or even Windows Forms and use clickonce?
The fact that McAffee blocked something has NOTHING to do with it's long term future. McAfee is nobody, they don't determine whether or not a specific technology can be used or not, or will be used in the future.
Silverlight seems to have reached a dead end, due to Microsoft realizing that if they create a multi-platform application environment, people might just stop using Windows.
WPF also seems to have been abandoned in favor or WinRT XAML, but of course WinRT XAML is not an option for us developers at the moment, simply because it is Windows 8-only, and our customers don't have Windows 8 or greater.
Besides, technically, WinRT-XAML seems to be really inferior to WPF XAML, and lacks many important features.
Of course winforms is completely useless and is not an option, unless you need to run your applications in my grandma's 80386 computer with an Hercules monochrome monitor
(exaggeration).
Seriously, that dead technology is not the answer to any of today's challenges. Things that you can do easily in any of the XAML-based technologies are either impossible or require a bunch of horrible hacks in winforms.
I suppose the definitive answer depends a lot on the usage scenario, for example:
Silverlight makes more sense if you have to publish your application in a web site and have anyone download it and use it.
WinRT XAML makes sense only if you target Windows 8, or want to create 'metro style' apps.
WPF makes sense if you want to create Windows Desktop applications, and have them deployed via ClickOnce or Windows Installer to a more limited and controlled set of users (because it needs installation of the .Net Framework, which Click-Once can deal with anyways).
winforms makes no sense whatsoever because it's a completely useless dinosaur technology that doesn't support anything.
Thank you very much for your answer and sharing your experience with me. I always try keeping WinRT as a strategy planning for each line of code I write. As I wish to migrate my code easily to WinRT in the future. The future is translated to me as may be within two years I may find myself in a situation that we will be targeted to migrate to something like Windows 9.
The best successful migration scenario - as I wish, it could be by increasing the usage of PCL at the client side as much as possible and test in a small piece of WinRT module.
We hope that Microsoft succeed in Windows 9.
I am sadly decided to shift to WPF instead of Silverlight in case that Silverlight depends on browsers that may not be supported in the future. As we here that Chrome, Safari and others are stopping supporting Silverlight. Why should I insist in relaying with such great technology that is dead before it finishes.
The main difficulties I could face in the future migration could be tied to two patterns:
- Prism.Regions: Windows Store has better than that.
- Prism.Modularity: Windows Store has no migration strategy to that.
At least at the moment I'm writing my thoughts Prism for Windows Store are far from implementing Regions and Modularity.
This is not a final answer but a clue of my what I should and should not to do.

Silverlight 4 seems like starving of memory

I have been playing a bit with Silverlight and try to port my Silverlight 3.0 application to Silverlight 4.0.
My application loads different XAP files and upon a user request create an instance of a Xaml user control and adds it to the main container, in a sort of MEF approach in order I can have an extensible and pluggable application.
The application is pretty huge and to keep acceptable the performances and the initial loading I have built up some helper classes to load in the background all pages and user controls that might be used later on.
On Silverlight 3.0 everything was running smoothly without any problem so far.
Switching to SL 4.0 I have noticed that when the process approaches to create the instances of the user controls the layout freezes unexpectedly for a minute and sometimes for more. Looking at the task manager the memory usage of IE jumps from 50MB to 400MB and sometimes up to 1.5 GB.
If the process won't take that much the layout is rendered properly even though the memory usage is still extremely high. Otherwise everything crashes due to out of memory exception.
Running the same application compiled in SL3, the memory used is about 200MB when all the usercontrols are loaded. Time spent to load the application in SL3 is about 10 seconds, while it takes up to 3 mins in SL4 There are no transparencies, no opacities set, no effects and animations in the layout.
User controls are instantied on the fly and added or removed in the visual tree on purpose when the user switches from one screen to another. The resources are all cleaned properly when a usercontrol is removed from the visual tree to allow the GC to operate in the background.
I may do something wrong but I could not figure out where exactly nail out the source of this problem. As far as I know there is no memory profiler in SL4 that can help me out to find where to look at. But again I could not be updated on new debugging tools available.
UPDATE:
Silverlight 4 Service Release to fix Memory Leaks: http://timheuer.com/blog/archive/2010/09/01/silverlight-service-release-september-2010-gdr1.aspx
Silverlight 4 has known memory leaks and the fix is currently being tested.
Here is a Microsoft Thread about it:
The user "heuertk" is a Microsoft Silverlight Developer....he explains the issues and the status of the fix...
http://forums.silverlight.net/forums/t/171739.aspx
To be honest despite your assertion "The resources are all cleaned properly when a usercontrol is removed from the visual tree" this where I'd start looking. It really does smell of elements of UserControls not being released properly.
This typically occurs when there are long lived (for example static) objects that raise events that more transient objects attach to. If these event handlers aren't detached it leaves these objects hanging around in memory.
I am keep testing and would like to share what I discovered so far.
It might be helpful to understand the behaviors of SL4.
Since it sounded pretty weird the UIThread can take so long to render a bunch of graphic components and considering Microsoft has improved the render pipeline, I have reverted my solution to SL3 but I have kept SL4 installed on my localhost.
The application uses RIA Services and moving back and forth to SL4 means that I have to do some changes in the code as per released documentation.
The application runs extremely smooth more than it was if it is tested with Visual Studio 2008. The memory usage is lower than it was before when SL4 wasn't installed.
As soon as I switch to VS2010 it is a completely different scenario. Memory grows undefinitely, the layout is slow to catch the user interaction and sometimes it freezes as before explained.
I have disabled RIA Services, by using standard Rest service and the process hasn't changed in quality.
In conclusion, considering I will keep testing in order I can finally understand what is really preventing to get the application running in an acceptable mode, I believe the memory issue is due to the debug process of VS 2010 or a combination of VS 2010 and SL4
Just for your info: there is a memory leak when using Silverlight Toolkit Charts in Silverlight 4.0. This is a known bug and Microsoft is working on it.
Silverlight 4 is Leaking everywhere. It Doesn't release memory properly.

Are WPF and Silverlight on a collision course?

It seems like these two technologies, already similar, are on a path to merge into a single technology. There are a lot more WPF-like controls in the Silverlight toolbox, and WPF now has Silverlight's VisaulStateManager. At this point, it's probably fair to say that Silverlight has even surpassed WPF in terms of the number of themes available.
How long until these two technologies become one? How long until the difference between a rich client app and a rich browser app is a simple compile-time setting?
EDIT
Let me clarify my question. I realize that any browser application needs to run in a "sandbox" for security reasons, and I also understand the desire to keep the browser plugin as small as possible, but there are several minor differences between the two technologies that could probably be massaged out without compromising either of these goals. For example, there could be a lot more overlap between UI controls and themes. Today, you can't just use a Silverlight theme in a WPF app, but how much of a leap would it be for Microsoft to make this possible?
I don't think they'll ever merge into one product. Microsoft has intentionally left a lot out of Silverlight to keep its footprint small. And then there's a plethora of security issues Silverlight must abide by when running in a browser. And of course they've designed it so it'll run on a PC or a Mac (unfortunately the same can't be said for the .NET Framework).
I am happy to see resources shared between WPF and Silverlight though. They were supposed to be similar from the beginning. As a result, it's relatively easy to port a Silverlight project into WPF. On the flip side, it's not quite as easy going from WPF to Silverlight simply because WPF has always had more features, but that's just the nature of the beast.
UPDATE:
So your revised question is interesting. It would be cool if Microsoft could make it possible for you to basically flip a switch to change the behavior of your app between Silverlight-like functionality and WPF. They would be facing a great deal of challenges though, not only with security but with the fundamental behaviors of some of the lightweight Silverlight controls vs the feature-rich WPF controls. These differences could potentially complicate things for the developer even further.
For example, in WPF there's a built-in multiple undo & redo system in the textboxes. In Silverlight there is no such thing so I actually had to write my own. In order for the developer to account for things like that they'd have to do build a lot of feature-checks into the application.
With all that said, I suppose a compile-time switch as you described might be feasible. But I still think it's unlikely Microsoft will create this kind of capability any time soon.
XAML and databinding may become closer between the two
but the rest of the framework will probably never be the same.
For once, you can not automate an Office application using Silverlight.
And that may never happen, unless MS decides to open a bridge of some sort
between the plug-in and the .NET Framework.
Security, consistence, and great UI are main driving forces in Silverlight.
If you can do something in Windows that you cannot do in Mac than
consistence is lost.
Silverlight installations has its own libraries built for specific operating systems. We as developers use what microsoft gives us that can run on those systems. WPF is full trust to the windows operating system which uses specific windows api calls.. So i'd say never will they merge.
Yes we will see that WPF and Silverlight becomes more and more alike, if the acutely will be merge? maybe It's not impossible but what we will see is just that what you stated that they will be more alike. So in the future you will not implement WPF OR Silverlight you will just implement XMAL.

Switching to WPF. Is it time?

I'm considering switching from MFC to WPF.
My first concern is that there are too many users who don't have .NET with WPF installed yet. Can anybody point to a source containing the WPF penetration numbers?
My second concern is speed.
Any other considerations?
I've been banging away at WPF for a while now. It is brilliant, but it still has (occasional) holes you've to plug yourself. However all indications are .NET 4.0 will be a significant step forward.
I would say start now. The WPF learning curve is REALLY steep, and it'll be a while before you'll be releasing software to users, believe me. Also do yourself a favour and get the WPF Unleashed book. It's superior.
Speed isn't a consideration. The power WPF gives is well worth any drawbacks with speed, which - coming from Windows Forms - I haven't noticed to be honest.
What kind of application are you developing? If it's a wide-distribution desktop app that you want your grandmother to install, your concern about .NET 3.0/3.5 adoption is valid. So far from what I've seen, performance is less of a concern.
WPF penetration
First of all, Windows Vista and Windows 7 both have WPF preinstalled, which accounts for 35% of the market automatically. Windows XP has had it as it had .NET Framework 3.0 as an option in Windows Update for over three years, and many applications ship with it, so it is likely to also be installed on a high percentage of Windows XP machines. StatOwl indicates that about 80% of NET Framework installations are version 3 or above.
If you're shipping on CD it is no big deal to include the latest .NET Framework on the CD and have it install automatically. If users are downloading your application, it can contact Microsoft's web server to download and install the latest .NET Framework. Online ClickOnce deployment also has this capability if you want people to be able to start their application directly from the web browser without installing it.
So the bottom line is, you probably don't need to worry about whether people will have WPF installed on their machines or not unless your target market consists primarily of dial up customers on Windows XP who don't run much third-party software (i.e., they just run Windows and your application).
Speed
Not an issue. I have a 200 MHz Pentium Pro with 384 MB RAM from 1998 that I test my software on, and my WPF applications have comparable performance with equivalent MFC applications. If your WPF application uses lots of fancy graphics and animation it will run slowly on ancient CPUs and graphics cards, but so would an ordinary MFC application with the same features.
Don't even bother trying to use WPF if you are sticking with Visual Studio 2008 for the next year or two. The experience will be way too painful. I'm talking about "my IDE crashed again" type of pain.
If you are going to use VS 2010 in the near future, then WPF is a blast. Download the beta, a couple of themes off CodePlex, are start playing. Once you get past the (freaking huge) learning curve I think you will find it to be quite enjoyable.
IMHO, you should wait for Visual Studio 2010 and WPF 4.0 to make the actual migration. They will close some very annoying gaps in the product.
Meanwhile, you can try it out. In terms of coding/readability -- it's going to be WAAAY better than with MFC =)
As for the performance and platform -- it shouldn't be a problem unless you have any very special circumstances (like if you can't require users to install .NET).
Also see this related question on switching to WPF from Windows Forms.
If you are thinking about a larger, modular, appliation I recommend checking out Prism. It's a bit of a beast itself, but you should be able to tackle it after coming to grips with C#, Dependency Properties and XAML. Plus, learning Prism gave me a much better understanding of WPF/Silverlight, at least from the development/binding side.
Mike Taulty posted an excellent 10 part video series on Prism. It's a great way to get your head around the platform.
I'd also recommend the pages linked to from the Getting Started page on codeplex. After all that, you're probably ready to tackle the Reference Implementation which comes with the download.
A previous answer of mine might also help clear up any remaining confusion around Controllers/Presenters in the framework that you might have (I did).

WPF vs Silverlight 3.0

Silverlight 3.0 beta has just been announced at Microsofts Mix Conference in Las Vegas.
Two features of the new beta are 3D-graphics and the ability to run applications outside of the browser, which to me seemed to be two of the major features that WPF (Windows Presentation Foundation) previously offered over silverlight.
I am currently evaluating WPF and Silverlight for possible use in our companies future development activity, and this announcement has left me confused as to the intended direction of these two UI technologies and why I would choose one over the other.
Has anyone implemented a new application using WPF recently, and if so, what drove you to that decision? Given the announced changes to silverlight, Would your decision have changed had you made it now, and if not, why?
Any advice would be appreciated.
The biggest difference I find is the asynchronous model you have to adopt
in your Silverlight application.
It does seems like an advantage (and it can be), but it does make
life very difficult sometimes.
There are also some limitations that can be a real challenge like the absence
of print support.
I would recommend Silverlight over WPF when:
- There is no need for best possible performance (graphics included)
- Can get around the absence of print support (it will come, we just don't know when)
- Camera/Microphone support is not needed
- Can tolerate the assync app/development model
- Can tolerate limitations on WCF (no support for WS-Security at this point)
- There is no need to store huge amount of data on the client.
- There is no need to direct integration with client side applications like Office.
- Has a server to host your application
I would say the main difference is that WPF requires the client to have the .Net 3.0+ framework. Silverlight only requires the runtime. Now that being said, WPF is geared more for controlled environments such as an intranet. Silverlight is meant for the public web. Another difference is that Silverlight is cross platform (Windows, Mac, Linux in the future & Cross browser). WPF is meant for Windows only.
The .Net framework can be a huge download for some users. Silverlight is only 4-5MBs. This is a big difference to run your app on the web, but not a big issue if its an internal application at your company.
Silverlight is Sandboxed which is meant for web use. So if your app requires more permissions you will need WPF.
There are also some differences between Silverlight code and WPF. But from what I've heard, the ultimate goal is to get a Silverlight to run inside of WPF with minimal code changes. But they aren't there just yet.
I have just worked on a WPF project that in hindsight we feel we might have chosen SilverLight for. It is probably more important to know the differences and select the one that is most appropriate for what you're doing.
Here's my starter for ten on some of the important differences - there were originally some differences in the available controls, but that has largely been smoothed out now.
Silverlight
Runs entirely on the client with AJAX
calls to the server for data
Can run on any server, including Windows and Linux / Apache
Uses COMPACT .NET framework
WPF
Runs on the client... usually calls services for data
Runs on Windows XP / Vista with .NET 3.5
Utilises the entire .NET framework
Silverlight is basically a stripped down version of WPF in order to make the runtime libary download as small as possible.
As a result, WPF simply has a lot more functionality available in it and tasks that are simple in WPF often become not so simple in Silverlight.
If running as a web app is not a requirement then the decision is a no-brainer - WPF all the way.
Has anyone implemented a new application using WPF recently, and if so, what drove you to that decision: Well since WPF was desktop only (or browser based using XBAPS - but that was more a deployment system than a real system) that was a good reason to it.
"Would your decision have changed had you made it now, and if not, why?" - No Silverlight, even on the desktop in v3, is still highly sandboxed and so certain functions are going to be hard/impossible to do due to the sandbox. Also the ability to use DirectX parts in WPF will still give another optimisation area which Silverlight and it's 3d won't be able to use.
It's worth noting that Silverlight's 3D is not the full 3D support of WPF, but only projection of 2D into 3D - i.e. take the 2D plane and allow rotation in X, Y & Z directions. WPF has full 3D modelling with materials, view ports, lighting and camera positional support etc.
I'm well along in the development of our first WPF app for release. Silverlight 3 looks great, but for this application I would still have chosen WPF. The application centers around presenting and manipulating very large sets of images hosted on a central server on our clients' networks. Additionally, the software update/change rate will be minimal. Mass import of new images from a local drive, no Internet connectivity requirements, performance concerns, etc. make this a project well suited for WPF.
One of our upcoming projects, however, will require many remote users to access a single data store on our network. The data they work with requires significant validation and error handling, so running that code locally is ideal. They will need the ability to work both on and offline and remain in synch (probably with SQL Data Services). SLOOB (Silverlight Out Of the Browser) will most likely be our choice for that one so they can have all the Silverlight advantages but use it like a regularly installed application, even without an Internet connection.
Both formats have their place: the trick will be to avoid using Silverlight for everything - we have more tools than just 1 hammer. :-)
Another difference is that with SL you only have one 'window', you can't have dialogs (they can be simulated but their size is limited to the main window) and you can't add multi monitor support.
If you have to interact with existing business applications (e.g. open a document in the archive viewer) you need to use WPF.
I recently have built several internal tool using wpf, and I chose it simply because It was easier for me coming from win32 work. I don't really think that the differences are major, and really... everything i have seen/heard indicates that porting between wpf and silverlight is quite easy.
Storage: You only have 25MB of isolated storage out-of-browser. If I remember correctly from some mix09 video, this limit is lower if your app is in-browser.
http://bliny.net/blog/post/Out-of-Browser-with-Silverlight-3.aspx
No FlowDocument: So there are limitations there too.

Resources