0

I'm using java to make a multithreaded program that simulates a restaurant. I'm using the newest Eclipse Java Oxygen to do it. I've tried to run it on both windows 10 and linux mint, always using Eclipse, and it works well on windows but when I try the SAME project on linux it stops. Is there any difference between running a multithreaded program on windows and linux?

Thanks

Improve this question
6
  • 1
    Yes (obviously). Can we see some code? Commented May 6, 2018 at 1:01
  • 1
    What do you mean by "it stops"? Is there an error? Are you sure your issue is with concurrency, rather some dependencies that are windows specific? Commented May 6, 2018 at 1:03
  • By "it stops" I mean it gets stuck at one point, no errors though. Commented May 6, 2018 at 1:13
  • 1
    Does the code which causes the deadlock involve any IO such as reading from a file or sending data across machines? It'll be difficult to pinpoint even a possibility without more information. Please create an MCVE which allows us to reproduce the problem. Commented May 6, 2018 at 1:18
  • Yes it involves IO by registering on a logfile all the states of the problem, but in order to do it I'm using a library that my teacher provided, so I suppose it works well. Commented May 6, 2018 at 1:34

2 Answers 2

Reset to default
3

Because Java is built to be cross-os compatible, there should be no difference in your code no matter the OS. This means if you have an application that uses multiple threads in it, it will work on any Os as long as your actual application works as you want it to. The reason this is the case is that Java converts your Java code in a .java to byteCode in a .class which will be turned into the assembly code for each respective OS. Something that could be stopping your program is Java being outdated, or not installed altogether. If neither is the case then you can look into things like RAM availability and any kind of thread limitations per program in Linux Mint.

Improve this answer
Sign up to request clarification or add additional context in comments.

5 Comments

Thanks! Really appreciated your answer. That was just what I thought but I really did not figure out what's the problem I have on Linux Mint, have to check it out. So I can say I have the right code if it is working on windows (it is, perfectly), right?
Yes, that is correct. As far as I know as long as your application is not depending on anything Windows specific, you should be ok to run it on any OS.
@Kwright02 - You are also assuming that the OP's code doesn't have threading-related bugs in it. Otherwise known as heisenbugs. Such bugs often cause an application to behave differently on different platforms.
Yes, I am assuming that, and I stated such in my answer
I was commenting on your comment. Where you respond to the OPs question: "So I can say I have the right code if it is working on windows (it is, perfectly), right?". You said Yes. The correct answer is actually No.
2

Multithreading in Java relies on the OS to schedule native threads. The native thread schedulers are implemented in the kernel by the operating system itself. They are different pieces of code with different behavior. And they are also (possibly) tunable at the system level.

Yes, there are differences than can affect the way that a Java application runs, but it would be difficult to work out what they are ... let alone describe them simply.

It is also possible that the problem you are seeing is due to some other differences between the different execution platforms: there are lots of possibilities. Or it may be due to the way that your application talks to the file system or network or external applications, all of which have differences.

I would suggest you start by using the debugging tools available to you to characterize what is happening when "it stops". Have threads died? Are threads blocked on locks? Etcetera. Then look for specific causes for the specific behavior.

Note that there are two groups of "classic mistake" that people who are new to threads in Java make:

  1. Making unwarranted assumptions about how threads run; e.g. that threads are scheduled fairly, or that they will start and run in an intuitive order.

  2. Inadequate synchronization; e.g. when two or more threads access and/or modify a shared variable or data structure without adequate synchronization. (This can lead to behavior that is unpredictable and highly couter-intuitive.)

When all is said and done, you can write a multi-threaded Java program to work correctly on many platforms. But we can't help you with specific problems unless you describe the problem clearly and show us the relevant code. In a case like this, and MCVE is highly advisable.

Improve this answer

4 Comments

As long as the written code abides by the specification, wouldn't these inconsistencies be considered a bug in either the memory model or the used VM implementation?
In this case, the specification he needs to abide by includes a lot of stuff apart from JLS 17.4. For example all of the things you can't assume about threads, notify/wait, etcetera.
I only didn't post code because it's a really large assignment and even the relevant code is very long and complex. Anyway, I'll follow your advices. Thank you!
I advise you to reduce your application down to an MCVE (best!!) or maybe post the parts of the code that are relevant to multi-threading.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.