Skip to main content

Parallel Programming


If you are writing multi threaded application with .Net , there are various options to choose from. There is Thread,Background worker, delegate etc. 
.Net framework 4.5 has introduced another approach towards parallel programming "async and await". With async n await, code follows a logical structure which looks similar to synchronous code and provides benefits of asynchronous programming.




Above is an example of a program which downloads the historical prices from Yahoo finance and saves it as a csv file. 

 Here Act_Async is an asynchronous method which makes a call to GetStreamAsync and after that performs some operations (which are some print operation for ex.). Once those operation are complete and it requires the output of GetStreamAsync , it await on the return object (which is stream in this case). Once await is called, control is returned to the calling code (which is main method).

 In Main method, once the control comes back  it performs some operation (which are again some print operation for ex.) and the thread is never blocked. when it needs the output from Act_Async method, and if Act_Async method has not come back yet, it waits for it to come back. Once the results are back, it proceed with the result for further processing .

 Below is similar method written using Thread class.

Some of the key differences- 

  1.  You need to make use of a shared variable which thread updates since ThreadStart delegate can not point to a method which returns a value (difficult to use with functional programming) while async method can return an object.  
  2. The flow of the code is not as smooth as with async n await. code may look similar but code with async n await is much more easy to understand and follows a sequential structure. 
  3. The flexibility with using Thread is that You can manage it ( Do you really want to !!!) like setting up background/foreground , priority etc . 


 There may be some cases where you need certain notification on progress of parallel tasks (some of these are provided in background worker class) those are not available here but you can use async n await in almost any case where you are using any other threading approach for parallel processing.

 For more details , visit MSDN. 
Labels:

Comments

Post a Comment

Popular posts from this blog

The view or its master was not found or no view engine supports the searched locations!!!

On an MVC application, if you are getting error " The view or its master was not found or no view engine supports the searched locations " and you see that this view (.cshtml file)is already present under the correct folder structure and every thing is correct and you have also tried a lot of posts on SO on how to fix that and that hasn't helped,  just check the build action for this file and make sure that it is set to "content". I have seen the build action set to None (by mistake)and after you deploy/publish  the application, that view would be missing  and you get this error .
Post a Comment
Read more

Know where to patch!!!

Have you tried to patch a function call in your unit test and wondering why the hell patching is not working, value I'm returning from the patched function is not the value being used in the function call. Well, may be you are not patching at the right place. I saw this problem at work. I'll try to explain with a simple example. you have a module module_a which has a function greet. There is another module module_b which uses this function from module_a in another function - sayHi . When we write test case for module_b's sayHi, we want to patch function call to module_a's greet method. module_a.py def greet (): """ returns sum of two numbers""" return 'hello world' module_b.py from module_a import greet class myClass (): def sayHi ( self ): return greet() if __name__ == '__main__' : res = myClass().sayHi() print (res) test_mod_b.py import un...
Post a Comment
Read more

Optional Parameters in C#

C# 4.0 has a nice feature of creating methods with optional parameters which allows caller to pass only required parameters and not all. This is particular helpful in methods which needs a lot of parameters. Method with optional parameters : Out put :15 Here caller is only passing one parameter i.e. 5 and output is 15 because default value for parameter b is used which is 3. but, care should be taken when you create methods with optional parameters in a library which is used by third party application (where you have less control) because by just changing the default value in the library , if this new library is deployed (with every thing as it is),  client apps will NOT see the change. Client application also needs to be re-compiled with the updated library to see the changes because of the new default value.  in the above example , if OptionalParam.exe uses   updated library.dll (without recompiling the complete solution ) where default value of...
Post a Comment
Read more