Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

My Scalable Parallel C++ Conjugate Gradient Linear System Solver Library was updated to version 1.60

26 views
Skip to first unread message

Ramine

unread,
Feb 16, 2017, 4:14:39 PM2/16/17
to
Hello......


My Scalable Parallel C++ Conjugate Gradient Linear System Solver Library
was updated to version 1.60, and it has become really powerful.

I have just eliminated a contention on memory, and now it has become
fully scalable on NUMA architecture, but you have to set the second
boolean parameter of the PCG_DENSE_Solver() constructor to true , and it
will support processor groups on windows and it will allow you to go and
scale beyond 64 logical processors and it will be NUMA efficient.

You can download my new Scalable Parallel C++ Conjugate Gradient Linear
System Solver Library version 1.60 from:


https://sites.google.com/site/aminer68/scalable-parallel-c-conjugate-gradient-linear-system-solver-library


Thank you,
Amine Moulay Ramdane.

Rick C. Hodgin

unread,
Feb 16, 2017, 6:38:42 PM2/16/17
to
What does it do? What's an example of its use? Pseudo-code will be
fine for an example.

Since it's really powerful, It would be wrong to pass it up due to my
own ignorance.

Thank you,
Rick C. Hodgin

Ramine

unread,
Feb 17, 2017, 11:24:04 AM2/17/17
to
It's used also in finite element method:

In this paper, we introduce a special way to store only the nonzero
elements of the stiffness matrix to obtain an efficient parallel
algorithm to solve partial differential equations with finite element
method. This storage method is obtained by considering the structure of
the mesh and the form of the stiffness matrix. The size of the matrix is
“the number of unknowns” by “a constant” instead of “the number of
unknowns” by “the number of unknowns”. Our method and algorithm are
efficient in both time and memory. Experimental results are presented.

http://www.sciencedirect.com/science/article/pii/S0096300303002790



And read about Conjugate gradient method:

https://en.wikipedia.org/wiki/Conjugate_gradient_method

Rick C. Hodgin

unread,
Feb 17, 2017, 11:33:40 AM2/17/17
to
It's interesting. I would suggest posting it in scientific, math, or
engineering forums where they could use your tool as a tool, rather
than looking at the underlying code (as we might do in a C++ forum).

luisc

unread,
Feb 17, 2017, 12:50:45 PM2/17/17
to
Congrats for your work. I look forward to have some time to implement your solver in our FEA package (https://github.com/lcpt/xc). Maybe the people of OpenSees, Calculix, Kratos, Code-Aster,... will be happy to hear about your work.

Ramine

unread,
Feb 17, 2017, 1:07:31 PM2/17/17
to
Thank you.

You can add it as an option, because my Dense and the Sparse solvers are
scalable on NUMA architecture.



Thank you again,
Amine Moulay Ramdane.





Ramine

unread,
Feb 17, 2017, 1:09:27 PM2/17/17
to
On 2/17/2017 12:50 PM, luisc wrote:
Hello,

And don't forget to read the readme files inside the zip file, i have
explained how to use all the objects and all the methods.


Thank you,
Amine Moulay Ramdane


Ramine

unread,
Feb 17, 2017, 1:12:49 PM2/17/17
to
On 2/17/2017 12:50 PM, luisc wrote:
If you want to send me an email please send it to:

aminer68 at gmail.com

Chris M. Thomasson

unread,
Feb 17, 2017, 4:58:37 PM2/17/17
to
On 2/16/2017 3:38 PM, Rick C. Hodgin wrote:
> What does it do? What's an example of its use? Pseudo-code will be
> fine for an example.
>
> Since it's really powerful, It would be wrong to pass it up due to my
> own ignorance.

Actually, Ramine is "sometimes" not all that bad wrt some of the
algorithms, and even analysis on such. He, afaict, has tried out and
critiqued some of my algorithms logic basins. Afaict, the hyper followup
postings wrt correcting errors can be collated. This person is excited,
and should try to go ahead and just "calm down", from time to time...

:^)

0 new messages