Re: [rvm-research] Exception while executing Junit Test Cases while running pre-commit

Sir,

Please find below , the proposal which I would like to submit on GSoC
Webpage. I request to you give your suggestions and comments to improve it.

*2. Project description*
                     Habanero-Java (HJ) hopes to produce portable parallel
programming abstractions for future multi-core technologies. Currently, HJ
adds new keywords to the Java language for asynchronous task creation,
mutual exclusion in asynchronous tasks, barriers, and phasers. The goal of
this project is to modify and extend JPF to be able to verify HJ programs
for deadlock and user provided assertion violations. The goal achieved by
simulating the HJ runtime directly in JPF either through a specialized
verification version of the runtime or arbitrary state transformations on
the JPF state.


*3. What is the development methodology you propose to use for the project?
Please include information on possible deliverables and the tentative
schedule.*

               I would like to follow agile development methodology. I
would like to work on daily targets made by splitting the monthly targets.



*The deliverables are*

1.      1. Code Walk of existing Habanero runtime’s implementation.
Understanding the implementation.

2.      2. Study of research papers on Delay bounded scheduling, which will
be implemented in the developing the runtime.**

3.     3.  Developing the HJ runtime from scratch customized for
verification on JPF supporting the constructs “for all” and “finish”.**

4.      4. Evaluating the work done in step 3 by executing HJ programs.**

5.      5. Extending the runtime built in step 3 for other constructs.**

6.     6.  Evaluating the extended runtime building in step 5.**


*4. What are the goals that you hope to accomplish in your project?
*

  The goals I would like to accomplish are

1.       For the verification of HJ programs, we need to explore all
possible scheduling of Jobs(tasks) which are created by the programs. The
existing HJ runtime does not support for exploring all possible scheduling
since it uses  “work-sharing”, “work-stealing “ mechanisms to schedule the
jobs. The goal is to develop new runtime which uses “Delay Bound Scheduling
mechanisms” to explore all possible schedulings of the tasks.

2.       The new HJ runtime for JPF should   support the  verification of
HJ programs developed  using   the constructs “async”,”finish”,
“future”,”phasers”,”isolated” .

3.       Develop HJ programs which are used to verify the correctness of HJ
runtime built in step2.

Thank You

Pushparaj


On Tue, Apr 30, 2013 at 2:05 AM, Pushparaj Motamari
<[email protected]>wrote:

> Hi Erik,
>
> please find my queries inline.
>
> On Tue, Apr 30, 2013 at 1:24 AM, Erik Brangs <[email protected]> wrote:
>
>> Hi,
>>
>> On 29.04.2013 02:56, Pushparaj Motamari wrote:
>> > [...] please suggest next course of action. I am working on fix for the
>> > bug RVM-1017.
>>
>> I assume that you're basing your attempts to fix RVM-1017 (= Implicit
>> unlock in synchronized method relies on local variable 0) on an
>> unmodified Jikes RVM. Do you want to fix RVM-1017 for yourself or do you
>> plan to contribute a fix to mainline?
>>
>> In the latter case, it might be a good idea to discuss what the right
>> approach to fix the issue is (e.g. in a new thread on rvm-core). In the
>> former case, you're of course free to choose any approach you like.
>>
>> I'm not sure if it's a good idea to reserve an extra word for every
>> method. The reuse of the slot for the this pointer is only a problem
>> with respect to synchronized methods. Looking at your diff, I also don't
>> understand why you're excluding certain internal classes.
>>
> When I do include the internal classes, the precommit is failing. When I
> exclude , it is succesful. I do not know the reason for this. Please
> explain me if you know.
>
>>
>> My suggestion is the following: choose an approach for fixing the bug,
>> implement it in the baseline compiler and test the fix extensively.
>> After you're done with the baseline compiler, you can start to fix the
>> bug in the opt compiler. Not sure if that's helpful to you. Please try
>> to ask a more specific question if it isn't.
>>
>> Thanks& Regards
> Pushparaj
>
>

------------------------------------------------------------------------------
Introducing AppDynamics Lite, a free troubleshooting tool for Java/.NET
Get 100% visibility into your production application - at no cost.
Code-level diagnostics for performance bottlenecks with <2% overhead
Download for free and get started troubleshooting in minutes.
http://p.sf.net/sfu/appdyn_d2d_ap1

_______________________________________________
Jikesrvm-researchers mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/jikesrvm-researchers