muMAG MICROMAGNETICS WORKSHOP
Wednesday, November 8, 1995
John Oti (NIST) gave a brief overview of the activities of muMAG
since its inception. This workshop was the second in a series begun
with a meeting at the Intermag'95 conference in San Antonio,
Texas. Following the first workshop, two work groups were convened:
The Standard Problems and Verification, and the Public Portable Code
work groups. The aim of the second workshop was to report on the
activities of the work groups and provide a forum for the presentation
and discussion of topical and fundamental micromagnetic issues and the
discussion of future muMAG activities.
Larry Bennett (NIST) announced a forthcoming micromagnetic workshop at
the George Washington University in May, 1996. He asked those present
to watch out for future announcements about the workshop, and to
endeavor to actively participate in it. For more information about
this workshop, and to get on the mailing list, send e-mail to
Klaus Ramstöck (University of Erlangen-Nürnberg) presented
work on the computation of domain walls in a uniaxial medium of
infinite extent. He proposed using this problem as a benchmark for
measuring algorithm performance by utilizing different techniques in
solving it. This may permit isolating the influence of solution
methods on the calculated results and computational speed.
Jimmy Zhu (University of Minnesota) presented simulated magnetic
structures in rectangular and circular sub-micron thin metallic film
samples. These results were compared with experimental data.
Jim Blue (NIST) presented an update on the muMAG Public Portable
Code Project. He described key organizational and implementation
issues pertaining to problem domain definition, portability,
applicable algorithms, software engineering design issues, software
package development, distribution and support issues.
The following speakers presented solutions of the First Standard
Problem. This problem requires the calculation of magnetization
curves and magnetization structures of a 1 x 2 micron permalloy film.
The complete specification of the problem can be found at internet web
John Oti (NIST). Simulation results presented: Final zero field
magnetization structures beginning with uniform and random initial
magnetization states, and magnetization structures and curves
resulting from application of longitudinal and transverse magnetic
fields, beginning with uniform and random initial magnetization
Jay Hoinville (Maxtor). Simulation results presented: Final zero field
magnetization structures beginning with circularly oriented and canted
uniform initial magnetization states; magnetization structures and
curves as functions of applied longitudinal and transverse magnetic
fields; low field susceptibility curves. These results may not be a
valid solution to the standard problem, as Jay has noted an error in
his handling of the exchange stiffness parameter.
Mike Donahue (NIST). Simulation results presented: Magnetization
structures and hysteresis loops for applied longitudinal and
transverse magnetic fields, and zero field structures for a variety of
initial conditions, including random moments, uniform magnetization,
four-domain, vortex, and seven domain states.
The following points were made during the general discussion part of
Standard problems should be simple, and solutions should contain
detailed specification of input parameters and solution methods. The
problem sizes should be sufficiently small to run reasonably fast on
personal computers. Solutions to standard problems should be posted on
the internet in an agreed upon format. Also it would be helpful to
have individual energy components (e.g., exchange, demag and
anisotropy) plotted as a function of iteration count.
First Standard Problem should be made more specific with respect to
mesh size and parameters of dynamic equations and additional solutions
Proposal described above in part II of these minutes should be
adopted as a standard problem. There is indication that other workers
are involved in this kind of modeling.
Experimental data should be made available to verify solutions of
standard problems. Larry Bennett informed that Lakeshore Corp. has
offered to provide experimental permalloy samples to be used in
verifying the First Standard Problem.
- General Comments
Anybody interested in using standard problems is invited and
encouraged to participate in the problem development process. The
more participation, the better the results.
muMAG Nov '95 Workshop Report / NIST CTCMS /