Advertisement
If you have a new account but are having problems posting or verifying your account, please email us on hello@boards.ie for help. Thanks :)
Hello all! Please ensure that you are posting a new thread or question in the appropriate forum. The Feedback forum is overwhelmed with questions that are having to be moved elsewhere. If you need help to verify your account contact hello@boards.ie

Language Benchmarks

Options
124»

Comments

  • Closed Accounts Posts: 1,567 ✭✭✭Martyr


    marco_polo wrote:
    As for whether or not there is room to optimize during runtime, the JIT compiler will decide that, not you

    The cpu will decide what code runs faster..thats something JIT or you can't argue about :D
    marco_polo wrote:
    As benchmark programs are trivial, there will obviously not be much optimizing that can be done by any compiler, it is really comes into its own for much larger programs where you cannot be sure that your code is 'perfect'.

    Assembly, the representation of machine code..the native language of the processor, is far more appropriate to use in benchmarking than anything else. :D
    hobbes wrote:
    *sigh* There is nothing wrong with the mac. Exceptionally good machines. If there was any complaint it would be their price.

    I just meant that intel would have a bigger share of processor market now that Mac's don't use PowerPC cpu anymore.(but they always did anyway)


  • Registered Users Posts: 3,945 ✭✭✭Anima


    Assembly, the representation of machine code..the native language of the processor, is far more appropriate to use in benchmarking than anything else.

    What does that even mean?


  • Closed Accounts Posts: 1,567 ✭✭✭Martyr


    Anima wrote:
    What does that even mean?

    It means that CPU's don't run on Java bytecode.

    Got any results yet, Anima?


  • Registered Users Posts: 21,264 ✭✭✭✭Hobbes


    Anima wrote: »
    What does that even mean?

    Nothing.
    It means that CPU's don't run on Java bytecode.

    They don't run on C++ OBJ code, or .NET bytecode either.


  • Closed Accounts Posts: 1,567 ✭✭✭Martyr


    Good observation there, hobbes.

    You're finally getting the bigger picture about how a cpu actually works. :P


  • Advertisement
  • Closed Accounts Posts: 1,567 ✭✭✭Martyr


    i suppose hobbes and Anima require a basic tutorial in computer architecture?

    there is a good enough description of assembly on wikipedia here

    But according to the best programmer in ireland, it means "nothing"

    i won't be doing your homework this weekend either, Anima..do it yourself.


  • Registered Users Posts: 1,916 ✭✭✭ronivek


    I think you need a basic tutorial in manners.
    Assembly, the representation of machine code..the native language of the processor, is far more appropriate to use in benchmarking than anything else.

    Your statement makes absolutely no sense; and attempting to belittle the people who think so is not going to magically make your statements appear to be anything other than faff.


  • Closed Accounts Posts: 1,567 ✭✭✭Martyr


    I think you need a basic tutorial in manners.

    I don't like being patronised.
    Your statement makes absolutely no sense; and attempting to belittle the people who think so is not going to magically make your statements appear to be anything other than faff.

    OK, let me help you out.

    An assembly language is a low-level language for programming computers. It implements a symbolic representation of the numeric machine codes and other constants needed to program a particular CPU architecture.

    read here

    clear enough for you?


  • Registered Users Posts: 5,618 ✭✭✭Civilian_Target


    This is a really interesting thread.

    I'm a serious Java programmer, and have got my hands dirty in both Java bytecode and Java performance analysis in the past.
    I also have written C++, and work in a company that uses C++ and Java (and TPF assembler :(), and is widely believed to have written some of the finest C++ code in the world, that you have all used.

    I think you have to go a lot further than ask what is faster. For what? On a mobile phone, Java apps can't hold a candle to any C app, but few people persist in writing distributed web apps in C++ -- machine time is so much cheaper than programmer time that people will gladly take Java and throw more hardware at the problem, and JEE servers tend to scale better anyway.

    Similarly, Java flys on SPARC boxes :D With computer power so cheap, its much nicer to just use the right tools for the job, and not worry about performance. Anyone who says it's not up to the job, take a look at the new Real Time VMs with pre-deterministic garbage collection http://www.ibm.com/developerworks/java/library/j-rtj4/index.html and tell me it isn't good enough for autopilot on a plane


  • Closed Accounts Posts: 1,567 ✭✭✭Martyr


    out of 100 posts Civilian_Target, yours was the best! :D


  • Advertisement
  • Registered Users Posts: 3,945 ✭✭✭Anima


    Martyr, you are making absolutely no sense what so ever. For the majority of the thread you are saying C++ is king bastard, then Civilian_Target says his bit and you thank him for his effort which is clearly pro-Java. What was the point of your previous posts?


  • Registered Users Posts: 9,579 ✭✭✭Webmonkey


    Anima wrote: »
    Martyr, you are making absolutely no sense what so ever. For the majority of the thread you are saying C++ is king bastard, then Civilian_Target says his bit and you thank him for his effort which is clearly pro-Java. What was the point of your previous posts?
    I agree, I don't know what is going on.


  • Registered Users Posts: 3,945 ✭✭✭Anima


    i won't be doing your homework this weekend either, Anima..do it yourself.

    That is completely unnecessary, I merely asked for reliable benchmarks and opinions. I am and was always willing to do the leg work to find the most accurate and satisfactory answer. You seem to have some kind of chip on your shoulder mate and to be quite honest, you are a bit of an asshole. Better yet you don't seem to have any formal education in programming. I know it's very easy to slate you from here on in for that but you seem to go out of your way to prove that you know things when clearly you don't.

    In short, I don't care what you think. Thnx k bye.


  • Registered Users Posts: 21,264 ✭✭✭✭Hobbes


    Anima wrote: »
    I merely asked for reliable benchmarks and opinions.

    I still stand by the "It depends". However if your planning to do your paper well I recommned just googling for similar tests and see what methods they used.

    Then use those methods as well on your own system you set up. The point about the processor to use (if has optimizations for Java or not) is a point to bring up.

    Lastly as Civilian Target points out try to take performance a bit further. So for example which scales better, which is better performance in related areas (eg. Mobile development, Web Services, Routers, Gaming, etc).

    Also which offers the better time to market and software maintaince which can be huge factors in maintaining performance.

    I only mention the last bit because simple "my app is 12 nanoseconds faster then your app" type papers have been done to death. :) These days 90% of all apps don't need to be bleeding edge fast.


  • Registered Users Posts: 5,618 ✭✭✭Civilian_Target


    Anima wrote: »
    Martyr, you are making absolutely no sense what so ever. For the majority of the thread you are saying C++ is king bastard, then Civilian_Target says his bit and you thank him for his effort which is clearly pro-Java. What was the point of your previous posts?

    Woah, I'm not pro anything.
    The first and most important thing for any task is to pick the right tools. I wouldn't dream of writing a site in either Java or C++ when PHP or Ruby on Rails is going to make my life much easier.


  • Closed Accounts Posts: 218 ✭✭Void


    Pick common compression algorithm like Huffman, C code here -> http://bcl.comli.eu/home-en.html. Both Java and C# have implementations of these algorithms in their standard libraries. Algorithms are old so assume implementations are optimal? Use this instead of "do nothing" in benchmark timing loop.


Advertisement