Coffeehouse Thread

25 posts

.Net VS JVM

Back to Forum: Coffeehouse
  • User profile image
    Kerberos Mansour

    Just wondering....

    Is there a performance test out there for the java virutal machine Vs .net platform?

    I mentioned it in an xna thread but all I could find online was a written comparison between java and c# structure.

    What I was wondering about was a performance test of the same algorithm written for a .net enviroment and Java to see who comes out on top? I asume J# and java have the most similar syntax and most likely to be used.

    The main java program I use regularly is Azureus and it is a resouce hog (just like most java programs) but from what I have read it is somewhat subjective since it, and .net for that matter do bot call the OS for the relocation and thus register them differently than unmanaged code.....

    can some one straighten this out, for atleast shed more light on this subject?

  • User profile image
    Cannot​Resolve​Symbol

    Kerberos Mansour wrote:
    Just wondering....

    Is there a performance test out there for the java virutal machine Vs .net platform?

    I mentioned it in an xna thread but all I could find online was a written comparison between java and c# structure.

    What I was wondering about was a performance test of the same algorithm written for a .net enviroment and Java to see who comes out on top? I asume J# and java have the most similar syntax and most likely to be used.

    The main java program I use regularly is Azureus and it is a resouce hog (just like most java programs) but from what I have read it is somewhat subjective since it, and .net for that matter do bot call the OS for the relocation and thus register them differently than unmanaged code.....

    can some one straighten this out, for atleast shed more light on this subject?


    Hmmm...  I suppose one could do something like sorting in both Java and J# or C#...  might have to try that someday (brush off my old APCS notes as I haven't been in a class since where I had to write a sort).

    I'd suspect that C# would be slightly faster because it's more tightly integrated with the OS (and its IL is more closely correlated to x86 machine code).  Probably would be pretty close, though, especially with the JIT compiler for Java.

  • User profile image
    Kerberos Mansour

    CannotResolveSymbol wrote:


    I'd suspect that C# would be slightly faster because it's more tightly integrated with the OS (and its IL is more closely correlated to x86 machine code).  Probably would be pretty close, though, especially with the JIT compiler for Java.


    here is the thing though...is it faster because it uses a MS platform or because of its close correlation to x86 machine code? how much of that is a factor...lets say we test the same code on MONO for example to get a more objective answer

    I am not a developer by profession but I do enjoy coding in c# than in c++ it just seems simpler. But what a 12 year vet developer told me: if it has something to do with Borland (Anders Hejlsberg C# ) then I trust them, these guys know what they are doing and I respect them. It should also be noted that Borland has stoped supporting Delphi somewhat and moved on to C#

    anyways c# is an open standard java isn't then why isn't it that platform independent (mono support is limited) is it because java came out at the right time and its been around longer...or?

  • User profile image
    Cannot​Resolve​Symbol

    Kerberos Mansour wrote:
    CannotResolveSymbol wrote:

    I'd suspect that C# would be slightly faster because it's more tightly integrated with the OS (and its IL is more closely correlated to x86 machine code).  Probably would be pretty close, though, especially with the JIT compiler for Java.


    here is the thing though...is it faster because it uses a MS platform or because of its close correlation to x86 machine code? how much of that is a factor...lets say we test the same code on MONO for example to get a more objective answer

    I am not a developer by profession but I do enjoy coding in c# than in c++ it just seems simpler. But what a 12 year vet developer told me: if it has something to do with Borland (Anders Hejlsberg C# ) then I trust them, these guys know what they are doing and I respect them. It should also be noted that Borland has stoped supporting Delphi somewhat and moved on to C#

    anyways c# is an open standard java isn't then why isn't it that platform independent (mono support is limited) is it because java came out at the right time and its been around longer...or?


    Testing Mono too sounds like a good idea.  The reason Java is more platform independent is (1) it has more first-party support for platform independence (Sun makes a Unix-like operating system) and (2) it is older.  Mono will get better with time; the project hasn't really been around for all that long.

  • User profile image
    Kerberos Mansour

    ok but what about first party support for none MS platfrms? Tongue Out

    what about mac .net framework or for that matter none x86-x64bit support...
     speaking of which... does Itanium support .net (sounds like a no brainer to promote the damn fledging thing hehe!)

     

    on a side not the xbox360 is a none X86 platform so.... (oh come one PS3 its not like you'll support java anytime soon Tongue Out)

  • User profile image
    El Bruno

    here, is an interesting comparative between some different programming languages

    Wink

    Bye from Spain

     

     

     

     

  • User profile image
    Ion Todirel

    Kerberos Mansour wrote:
    But what a 12 year net developer told me: if it has something to do with Borland (Anders Hejlsberg C# ) then I trust them, these guys know what they are doing and I respect them. It should also be noted that Borland has stoped supporting Delphi somewhat and moved on to C#
    Borland stoped to support their IDE line because they want to sell it. (C++ Builder, Delphi/Kylix, JBuilder)

  • User profile image
    Sven Groot

    Kerberos Mansour wrote:
    what about mac .net framework or for that matter none x86-x64bit support...
     speaking of which... does Itanium support .net (sounds like a no brainer to promote the damn fledging thing hehe!)

    .Net 2.0 is available for IA64. Besides the Itanium .Net is also officially supported on several other non-x86 platforms such as ARM by the .Net Compact Framework, and a port for the PowerPC (the XBox360) is underway.

  • User profile image
    Kerberos Mansour

    Sven Groot wrote:

    .Net 2.0 is available for IA64. Besides the Itanium .Net is also officially supported on several other non-x86 platforms such as ARM by the .Net Compact Framework, and a port for the PowerPC (the XBox360) is underway.


    Assumed as much with the IA64 and Xbox360(XNA)...any word on SPARC?

    oh and Ion Todirel I said vet (veteran) developer of 12 years, but anyways...how do you sell an IDE by not supporting it?

  • User profile image
    wkempf

    CannotResolveSymbol wrote:
    Testing Mono too sounds like a good idea.  The reason Java is more platform independent is (1) it has more first-party support for platform independence (Sun makes a Unix-like operating system) and (2) it is older.  Mono will get better with time; the project hasn't really been around for all that long.


    1) Mono has been around as long as .NET has, which is less than Java but significant enough time not to be considered "new".

    2) Mono is nearly "complete".  The compiler has been complete for some time.  So has the BCL.  Their packages for MS specific APIs are nearing completion (ASP.NET is nearly 2.0 complete, and WinForms now can run many existing applications).

    In regards to this thread, Mono has also been doing a lot of optimization work lately.

  • User profile image
    Cannot​Resolve​Symbol

    I'm doing some informal performance testing using a bubble sort right now.  Results and code to come when it finishes Wink

    If anyone has ideas for other performance measures or any better testing methods, tell me.  I'll provide the code if you want to run it on Mono, but you'd have to run all the tests over again yourself to get results you can compare it with...

  • User profile image
    blowdart

    CannotResolveSymbol wrote:
    I'm doing some informal performance testing using a bubble sort right now.  Results and code to come when it finishes .


    Remember the .net EULA has strict rules around publishing benchmarks Perplexed

    My favourite bit is the "I know you are, but what am I" sentence;

    Notwithstanding any other agreement you may have with Microsoft, if you disclose such benchmark test results, Microsoft shall have the right to disclose the results of benchmark tests it conducts of your products that compete with the .NET Component, provided it complies with the same conditions above.


  • User profile image
    Kerberos Mansour

    CannotResolveSymbol Best of luck on that test man Smiley My guess would be to perform a mono Vs JVM on linux Then repeat the same proccess on XP using .Net VS JVM I assume the closest comparison would be J# to Java, although c# is the logical alternative.... what do you guys recommend?

  • User profile image
    Cannot​Resolve​Symbol

    Kerberos Mansour wrote:
    CannotResolveSymbol Best of luck on that test man My guess would be to perform a mono Vs JVM on linux Then repeat the same proccess on XP using .Net VS JVM I assume the closest comparison would be J# to Java, although c# is the logical alternative.... what do you guys recommend?


    As I've stated, I can't test Mono on this machine.  I'm using C# on .net, the code is identical except for a couple places where I was forced to use the Java and .Net APIs.

    Results and code to come; I'm going to provide both so I'll comply with the EULA.

  • User profile image
    BryanF

    Kerberos Mansour wrote:
    on a side not the xbox360 is a none X86 platform so.... (oh come one PS3 its not like you'll support java anytime soon )
    Actually, the Blue Ray format that is to be used by the PS3 does include Java support, so it's not entirely out of the question. (Source)

  • User profile image
    Kerberos Mansour

    Actually, the Blue Ray format that is to be used by the PS3 does include Java support, so it's not entirely out of the question. That is very intresting....and since PS3 ships with a linux installed HD...then yu can have MONO and if thats the case then the Open GL in C# API is the way to go then? In anycase intresting Times are upon us

  • User profile image
    Cannot​Resolve​Symbol

    Okey dokey--  I've finished testing and put up my results.  Note that this reflects my experience, and may not reflect yours or even the performance of either in general, since my test doesn't use all of either platforms' capabilities.  Just a quick experiment to see if they're close or not.

    http://cannotresolvesymbol.googlepages.com/javaspeedtesting

  • User profile image
    Kerberos Mansour

    CannotResolveSymbol thanks man anyone intrested in testing this on mono?

Comments closed

Comments have been closed since this content was published more than 30 days ago, but if you'd like to continue the conversation, please create a new thread in our Forums, or Contact Us and let us know.