An important question ...

[aminer10]

Hello all, hello Dmitry Vyukov and others..


Now i am working on my Parallel compression library and finishing
it, it will support parallel LZMA, parallel BZIP , parallel Gzip and parallel LZ...

and since my parallel compression library is using the IO a lot, i am
overlapping the IO processing and CPU processing , this is
making my library more efficient and faster... and I will finish it soon..


Now my question is:

I am thinking about my next project, and i am thinking to construct/develop
a sort of compiler that 'parses' the parallel Object Pascal source code
and analyse it and construct a Petri Vet ...and after that you will be
able to visualize and analyse your Object Pascal parallel program
withtools like Tina etc.

Is it a good idea to construct such a parser etc. ?


Waiting for your answer ...

http://pages.videotron.com/aminer/



Regards
Amine Moulay Moulay.

[aminer10]

I wrote:

>Now my question is:

>I am thinking about my next project, and i am thinking to construct/develop
>a sort of compiler that 'parses' the parallel Object Pascal source code
>and analyse it and construct a Petri Vet ...and after that you will be

I mean: Construct a Petri Net.

>able to visualize and analyse your Object Pascal parallel program
>withtools like Tina etc.

>Is it a good idea to construct such a parser etc. ?


>Waiting for your answer ...

>http://pages.videotron.com/aminer

[aminer10]

I wrote:
>And you didn't answered my other post titled "An important question..."...


and prince Jedi Dmitriy Vyukov reponded

>I do not need it, I prefer to read code.

I am not so convinced with your kind of answer Dmitriy ..

Iwant also to automate the process of looking at the deadlocks and racing conditions etc.

So my compiler&parser will parse the Object Pascal source code and will construct
the Petri Net and 'automaticaly' visualise and analyse deadlocks and racing conditions
etc.

My compiler&parser will scan the Objects pascal source code and construct the Petri Net
and pass this Petri Net automaticaly to tools like Tina etc.

Andof course youcan visualize the Petri Net ..

I think that's still cool ..

I will try to give a demo of my compiler&parser that parses and construct
a Petri Net anddetects all kinds of dealocks , after that i will try to extend it
to racing conditions etc.


What others thing about my idea ?


Regards,
Amine Moulay Ramdane.

[aminer10]

I wrote:
>Hello all, hello Dmitry Vyukov and others..


>Now i am working on my Parallel compression library and finishing
>it, it will support parallel LZMA, parallel BZIP , parallel Gzip and parallel LZ...

>and since my parallel compression library is using the IO a lot, i am
>overlapping the IO processing and CPU processing , this is
>making my library more efficient and faster... and I will finish it soon..


Like in superscalar pipeline stages , i am overlapping the IO processing
and CPU processing and my Parallel compression library isgiving me
a very good performance in my benchmarks

I am using a lockfree_queue between the the worker threads producers
that compresses thedata and the worker thread that does IO processing
and copy the data to disk..

And of course since my Parallel library is 'more'IO bounded you have to have
a fasterhardsdisk (Raid etc..)


When i have compressed from a memory stream to memory stream it does
give a lighting speed(its veryfast on an in-memory compression..)


You will see myfull version soon that supports parallel LZMA, parallel BZIP ,
parallel Gzip and parallel LZ... and of course, it does uses my Threadpool engine
that uses work-stealing and a lock-free workerqueue for each worker thread
in the threadpool..


Regards,
Amine Moulay Ramdane.

[aminer10]

I wrote:
>I am not so convinced with your kind of answer Dmitriy ..
>Iwant also to automate the process of looking at the deadlocks and racing conditions etc.
>So my compiler&parser will parse the Object Pascal source code and will construct
>the Petri Net and 'automaticaly' visualize and analyse deadlocks and racing conditions
>etc.

>My compiler&parser will scan the Objects pascal source code and construct the Petri Net
>and pass this Petri Net automaticaly to tools like Tina etc.

>Andof course youcan visualize the Petri Net ..

>I think that's still cool ..

.I will try to give a demo of my compiler&parser that parses and construct
>a Petri Net anddetects all kinds of dealocks , after that i will try to extend it

I mean deadlocks ...

>to racing conditions etc.

>What others thing about my idea ?


>Regards,


Amine Moulay Ramdane.

[aminer10]

I wrote:

>What others thing about my idea ?

I mean what other think about my idea ?





Amine.

[aminer10]

Hello,

I have just looked at Tina for example..

http://homepages.laas.fr/bernard/tina/


Andi have wrote an article about Petri Nets

http://pages.videotron.com/aminer/PetriNet/formal.htm

If you look Dmitry at the ndr format ofthe toolTina

The NDR format of Tina looks like this:

p 77.999994 52.0 p1 1 n

p is a place and 77.999994 52.0 isa coordinate.


e t1 0 0 p1 0 0 1 n

and here you have for example a connection
between the transition t1 and the place p1 in
the Petri Net...


Now, i will design carefully my grammars and
construct the compiler&parser , i call it compiler
cause it will parses the Object Pascal source code
and compile itto the Petri Net modelin aTina format
or another format..

Andafter that you will be able to automate the process
of detecting deadlocks (even in complex programs...)

I will try to provide you with my first demo(after
i finish my Parallel compression library)that
compiles an Object Pascal program to the Petri Net
model and it will detects deadlocks...

After that iwill try to extendmy project to detect racing
conditions etc.


And you will be able to analyze and visualize also the Petri Net...

What do you think about my idea ?


Amine Moulay Ramdane.