gm2
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Gm2] Numerical Library : GM2 issue with multidimensional array para


From: Michael Riedl
Subject: Re: [Gm2] Numerical Library : GM2 issue with multidimensional array parameter
Date: Mon, 26 Feb 2018 18:25:52 +0100
User-agent: Mozilla/5.0 (X11; Linux i686; rv:52.0) Gecko/20100101 Thunderbird/52.6.0

Hallo Gaius,

thanks for taking care. The modules using the feature can now be compiled !!!

Another 6 modules in the numerical library still crash the compiler - will come back to that later.

I attached two files (one for 8 BYTE REAL - Eigen8.mod) and one for 12 BYTE LONGREAL (EigenLR.mod) as defined in GM2. In both cases you can change a  constant "UseLoc". If this is true a variant of the procedure "Rotier" (procedure name RotierLoc) is used which is not dependent on handing over single vectors out of a matrix - then the calculation is working fine. If the other variant is used I get an error

RTExceptions.mod:639:9:invalid address referenced in invalidloc

Perhaps you find some time to have a look on that.

Gruß / Regards

Michael


PS:

Just as an interesting point I also compared the timings of both the REAL64 and REAL96 variant (see third attachment). The tasks is to compute the eigensystem of a 256x256 matrix 100 times by a classical jacobi routine.

As expected in one of my former emails I see a huge time penalty for the 12 BYTE reals  (approx 1.8 times longer runtime if not optimized, 2.7 times longer runtime if optimized with -O2). Keeping in mind that using a linarised array for the input matrix for the Jacobi diagonalisation routine is memory but not cache optimal I expect an even more dramatic difference for the other routines (e.g. for the SVD routines not yet part of the routines I did sent you).


Am 19.02.2018 um 19:05 schrieb Gaius Mulley:
Michael Riedl <address@hidden> writes:

Hallo Gaius,


thanks for the update. Will recompile GM2 as soon as you think your
are ready

and start new tests.

Michael
Hi Michael,

I believe I've fixed the multi-dimensional array and high bugs.  I've
checked them all into the gcc-6.4.0 branch of gm2.  I will be checking
them all onto the gcc-svn branch soon and will be producing a gm2
release as well (based on gcc-6.4.0)

regards,
Gaius

Attachment: Eigen_time.txt
Description: Text document

Attachment: Eigen8.mod
Description: audio/mod

Attachment: EigenLR.mod
Description: audio/mod


reply via email to

[Prev in Thread] Current Thread [Next in Thread]