We have to speak frankly and clearly

[Ramine]

Hello,

We have to speak frankly and clearly, as is doing PhDs of IEEE
for example...


What i want to say this time, is that you have to know that parallel
programming is hard, cause you have to think about sequntial consistency
and about racing conditions and about Deadlocks and about Lock convoy
etc. and you have to know also that proving lockfree algorithms is hard
to do... so how can we proceed and walk confidently in this complex area
that
is parallel programming ? my way of doing is to implement a reduced
numbers of libraries that are stable and use them as base fondations
for other projects, this is my way of reducing the risk or eliminating
the risk, so what i have done by inventing my libraries and parallel
algorithms and synchronization algorithms is reducing the complexity of
parallel
programming by using those same parallel projects as building blocks for
other projects, i have taking a lot of care for verifying that
my parallel algotithms are free of bug and free of problems, and
i think i have succeeded in that regards by inventing and implementing
and bringing some stable parallel libraries and stable synchronization
algorithms, and as you have noticed i have not used Throttling
Concurrency in my Threadpool like they are doing in the following link:

http://msdn.microsoft.com/en-us/magazine/ff960958.aspx


Simply because my parallel projects dont contains waiting instructions
like sleep() or IO waiting or this kind of waiting functions, so
this makes my parallel algorithms efficient by using my Threadpool i
think.


So hope this will clear my intentions that were to bring to you
stable and fast parallel algorithms, and i think i have succeeded
in this regard.


You can download all my parallel projects from:

https://sites.google.com/site/aminer68/



Thank you,
Amine Moulay Ramdane.