2006 Highlights

Week of 12/23/05

Theory Group

A paper entitled “Radial transport of fluctuation energy in two-field
model of drift-wave turbulence” has been reviewed for publication in
the Physics of Plasmas. This paper is co-authored by O. Gurcan and
P.H. Diamond at the University of California at San Diego, and T.S.
Hahm at PPPL. In this work, a theory of spatial propagation of
turbulence, referred to as “turbulence spreading”, has been developed
for the two-field model of drift-wave turbulence. Simplified closure
expressions are used to obtain two-field coupled reaction-diffusion
equations for the kinetic and internal energy. The efficacy of
various nonlinear interaction mechanisms for spreading has been
analyzed systematically. Spreading of the internal energy is
predicted to lead that of the kinetic energy. The important role of
zonal flow damping in spreading has been identified, but zonal flows
are shown not to be the dominant agents of turbulence spreading.

Professor Herb Berk visited the PPPL Theory Department to continue a
long-term collaboration with N.N. Gorelenkov and G.Y. Fu on energetic-
ion-driven Alfvenic instabilities in fusion plasmas. During his
visit, discussions were held on such issues as multiple instabilities
in DIII-D. In particular, a stability theory of thermal ion-driven
Alfven instabilities in reversed-shear plasmas has been advanced and
its application to ITER were discussed.

L.E. Zakharov gave a presentation entitled “The Li Wall divertor
solution for ITER versus the present high recycling divertor” at the
US Burning Plasma Workshop (Oak Ridge, TN, Dec. 7-9, 2005). The
proposed solution is consistent with the ITER baseline design,
tritium and helium control, and with safety limitations on the amount
of lithium. The LiWall divertor appears to be the only solution that
can provide, in addition, the ELM-free plasma regime and ignition in
ITER.

Week of 12/16/05

Theory Group

The Beam Equilibrium Stability and Transport (BEST) code has been
recently parallelized by Hong Qin in the transverse direction using a
particle decomposition method, in addition to the previously
implemented domain decomposition in the longitudinal direction. The
parallel scheme utilizes the advanced communicator functions provided
by the MPI protocol. For parallelization in the transverse direction,
the particle decomposition method has much better efficiency than the
commonly used domain decomposition. This is because the communication
load for transverse domain decomposition is very heavy due to the
frequent domain crossing of a single particle in the external
focusing lattice and the space-charge potential. The newly updated
BEST code has been used to study self-consistently the collective
effects of 3D bunched beams. For routine production runs with 1500
processors on IBM SP super computer at NERSC, the performance of the
BEST code scales up almost linearly with respect to the number of
processors.

Dr. G. Rewoldt gave a presentation on a “Comparison of
Microinstability Properties for Stellarator Magnetic Geometries” at a
National Stellarator Theory Teleconference on Thursday, December 15.
The teleconference was hosted by Dr. A. Reiman. The presentation
materials are available at http://ncsx.pppl.gov//Physics/NSTT/
Theory_Index.html. That site also contains an archive of
presentation materials from previous stellarator theory teleconferences.

Jay R. Johnson and Gwangson Choe attended the American Geophysical
Union (AGU) 2005 Fall Meeting held in San Francisco on Dec. 5-9. Jay
Johnson gave two invited talks entitled “An Information Theoretical
Approach to Solar Flare Occurrence,” and ”Plasma Heating and
Transport at the Magnetopause due to Nonlinear Interactions with
Kinetic ULF Waves” and another oral presentation entitled “An
Information-Theoretical Approach that Identifies a Solar Cycle
Dependence of Nonlinearity in Magnetospheric Activity.” Gwangson Choe
gave two presentations entitled “Generation of Erupting Flux Ropes by
Reconnection Among Smaller Scale Flux Tubes,” and “Waiting-time
Distributions of Solar Eruptions and the Evidence for Sympathetic
Flares.” Choe also organized the special session “Nonlinear Dynamical
Approaches to Solar-Terrestrial Phenomena” and served as chairman of
two sessions.

CPPG Group

The latest version of the TORIC RF full wave solver has been
installed and successfully tested in TRANSP. This version uses a
full axisymmetric tokamak numerical MHD equilibrium without
assumption of up-down symmetry. It has been validated on a number of
H-minority ICRF heating shots from the C-Mod tokamak.
In a collaboration on the M3D-C1 code with Dr. Andrew Bauer of the
Scientific Computation Research Center at the Rensselaer Polytechnic
Institute, a new technique was developed for evaluating the matrix
elements involving 3 nonlinear terms. The old method required
1.6x10^10 operations to initially define each fourth rank tensor
representing a non-linear operator which is then written to disk, and
then a read from disk and 1x10^5 operations each timestep to define
the corresponding matrix element. In the new method, there is no
initialization necessary, and there are 2.9x10^5 operations each time
step. However, since the new method requires no disk reads, it is
considerably faster than the old method. The fact that it does not
require a lengthy initialization now makes it feasible to introduce
adaptive meshing, where the unstructured triangular mesh can change
from timestep to timestep.
Stephane Ethier attended the NERSC Users' Group monthly phone
meeting. It was announced that the new IBM p575 Power 5, Bassi,
would soon be open to all users, although the system is still in
acceptance period so the users will not be charged for the cycles.
Bassi consists of 111 compute nodes with 8 processors on each node
sharing 32GB of memory. 888 processors are thus available for
applications. This system will help relieve the highly overloaded
Seaborg, NERSC's 6080-processor workhorse. Although much smaller
than Seaborg, Bassi has processors that are 6 to 7 times faster.
Dr. Phil Colella from Lawrence Berkeley Laboratory, PI of the SciDAC
APDEC center, visited PPPL for a discussion with CEMM members
regarding existing collaborations on the Adaptive Mesh Refinement
code, and also to discuss collaborations in SciDAC-II.

Week of 12/2/05

CPPG Group

D. McCune and S. Jardin participated in a PTRANSP conference call on
11/28/05 led by Prof. A. Kritz, LeHigh U. The PTRANSP project is to
develop a predictive version of the TRANSP code, with the option of a
free-boundary or a fixed boundary equilibrium solver. The free-
boundary solver will be based on the LLNL/Tech-X TEQ code. Solver
options will include the use of the GA supplied Globally Convergent
Newtons Method (GCNM) package, which is especially effective in
integrating the transport equations when the GLF23 transport module
is utilized. Present plans are to have a PTRANSP version available
for testing by April 1, 2006.

Week of 11/18/05

Theory Group

The Senior Manager of Advanced Server Hardware Systems, Dr. George
Chiu, and the BlueGene Chief Architect, Dr. Alan Gara, of IBM Watson
Research Center visited PPPL to discuss future computational
requirements of fusion codes with the SciDAC GPSC group at PPPL. Dr.
Wei-li Lee, the PI of GPSC, and Dr. Stephane Ethier, a member of
CPPG, presented an overview of the current work and future directions
of large-scale particle simulations on BlueGene-type computers. The
IBM team presented the Research Challenges and Future Requirements of
the BlueGene Supercomputers. Future collaborations between PPPL and
IBM include the benchmark runs for GTC on the 32K BlueGene and the
subsequent performance analysis.

CPPG Group

S. Ethier attended the annual Supercomputing (2005) meeting in
Seattle and presented a paper “Leading Computational Methods on
Scalar and Vector HEC Platforms.”

Daniel Reynolds (UC San Diego) and Carol Woodward (LLNL) visited PPPL
for two days to collaborate with Ravi Samtaney on the development of
a fully implicit 3D MHD code. This is part of a joint effort by the
CEMM, APDEC and TOPS SciDAC centers.

Week of 11/11/05

Theory Group

A Physical Review Letter paper by N.N. Gorelenkov has been accepted
for publication entitled “Double Gap Alfvén Eigenmodes: revisiting
eigenmodes interaction with the Alfvén continuum,” where a new type
of global shear Alfvén Eigenmode is reported to be found in tokamak
plasmas where the mode localization is in the region intersecting the
Alfvén continuum. The eigenmode is formed by the coupling of two
solutions from two adjacent gaps (akin to potential wells) in the
shear Alfvén continuum. For tokamak plasmas with reversed magnetic
shear it is shown that the toroidiciy-induced solution tunnels
through the continuum to match the ellipticity-induced Alfvén
eigenmode (TAE and EAE, respectively) so that the resulting solution
is continuous at the point of resonance with the continuum. The
existence of these Double Gap Alfven Eigenmodes (DGAEs) allows for
potentially new ways of coupling edge fields to the plasma core in
conditions where the core region is conventionally considered
inaccessible. Implications include new approaches to heating and
current drive in fusion plasmas as well as its possible use as core
diagnostic in burning plasmas.

Dr N. Loureiro (jointly at CMPD, University of Maryland and PPPL)
presented a Theory Seminar on Thursday, 10 November on the topic of
the “Nonlinear evolution of the tearing mode.” Dr Loureiro presented
recent numerical results on the nonlinear evolution of the strongly
and weakly driven resistive tearing mode. A high-resolution numerical
scan of the parameter space shows that, in general, the tearing mode
evolves through five stages: exponential growth, algebraic growth
(Rutherford stage), X-point collapse followed by current-sheet
exponential reconnection (Sweet--Parker stage), tearing instability
of the current sheet (generation of secondary islands), and
saturation. He discussed each evolutionary stage of the instability.

CPPG Group

S.Jardin and D.McCune, together with J.Breslau, M.Chance C.Kessel
and L.P.Ku attended the all-hands kickoff meeting for the Fusion
Simulation Project (FSP) on Simulation of Wave Interaction with MHD
(SWIM) held at ORNL. The three-day meeting was mostly devoted to
defining the overall structure and components of the Integrated
Plasma Simulator, which is central to the first of the two campaigns
that is focused primarily on the effect of RF and energetic particles
on the sawtooth and other MHD modes. Also discussed were the
development of closure relations for the second SWIM campaign on the
stabilization of the
neoclassical tearing mode with RF.

Week of 11/4/05

Theory Group

Dr. A. Reiman gave a talk entitled, "Effect of Magnetic Perturbations
Produced by a Divertor Control Coil on the Flux Surfaces in the W7AS
Stellarator" at the Second Error/Non-Axisymmetric Field Workshop on
October 28, in Denver. Collaborators on the work were M. Zarnstorff
and D. Monticello from PPPL, as well as A. Weller, J. Geiger and the
W7-AS Team from Greifswald, Germany. Calculations were presented
showing that a coil that influences the resonant harmonics near the
edge of the W7AS stellarator can have a significant effect on the
width of the stochastic region at the edge of the plasma.

Dr. Leonid E. Zakharov attended the International Workshop on
"Performance of KTM tokamak
and its research program" in Astana (Kazakhstan). Together with Dr.
I. Lublinskii from the "Red Star" Federal Enterprise he emphasized
the need for a reactor oriented, lithium based program for the
Kazakhstan tokamak. He also spent two weeks at JET and MAST working
on calibration of equilibrium reconstruction on these machines. While
MAST equilibrium reconstruction can be calibrated using the Cbc2e
code developed for CDX-U at PPPL, a special technique was proposed
for JET which has the iron core behaving non-linearly in the
magnetic field. (See http://w3.pppl.gov/~zakharov for relevant
presentations).

Dr H.C. Yee (NASA Ames Research Center) gave a Theory Seminar on
November 3 entitled "Adaptive Flow Sensor in High Order Methods for
Complex Multiscale Flows". Dr Yee described a high order, wavelet-
based nonlinear filter method for the control of numerical
dissipation. One of the unique features of the numerical method is
that it is suitable for nearly incompressible to highly compressible
gas dynamics and MHD flows. It is also stable and accurate for a wide
spectrum of flow physics ranging from long time wave propagation of
smooth flows to high-speed multiscale turbulent/combustion flows
including strong shock waves. Dr Yee had constructive discussions
with Drs G. Fu, W.W. Lee and J. Lewandowski and others on the
applicability of the nonlinear filter method to fusion problems.

CPPG Group

Dr. Jeffery Salzman, Head of Computing at Merck Research
Laboratories, presented a CPPG seminar titled "Nonlinear 3-D Magnetic
Resonance Image Registration Using Grid Generation Techniques and
Multigrid Solvers" In this talk he described a variational approach
for solving a magnetic resonance imaging problem. The resulting
Euler-Lagrange equations are three elliptic partial differential
equations. The equations are discretized and solved using a full
approximation scheme multigrid method.

Week of 10/28/05

Theory Group

In recognition of her outstanding achievement in plasma science
research, Elena Belova, a scientist at the U.S. Department of
Energy's Princeton Plasma Physics Laboratory (PPPL), received the
Katherine E. Weimer Award. It was given at the APS-DPP’s annual
meeting in Denver this month. Belova is being cited for pioneering
analytical and numerical contributions to the fundamental physics of
magnetically confined plasmas.

Dr. Harry Mynick presented an invited tutorial talk at the 2005
Meeting of the APS Division of Plasma Physics, entitled “Transport
Optimization in Stellarators.” Drs. Elena Belova, Jerome
Lewandowski, Ravi Samtaney and Weixing Wang gave invited talks. The
titles of these talks were: “Advances in the numerical modeling of
field reversed configurations” (Belova); “Global particle-in-cell
simulations of microturbulence with kinetic electrons” (Lewandowski);
“Tokamak Pellet Fueling Simulations using 3D Adaptive Mesh
Refinement” (Samtaney) and “Neoclassical and Turbulent Transport in
Shaped Toroidal plasmas” (Wang).

Drs. Daren Stotler, Stephane Ethier, Ravi Samtaney, Taik Soo Hahm,
Weixing Wang, and Wei-li Lee attended the Organizational Meeting of
the Center for Plasma Edge Simulation, October 23 and 24 in Denver.
The first day of the meeting focused on physics and applied math
aspects of the project; the second day was devoted to Computer Science.

Drs. Daren Stotler and Taik Soo Hahm participated in an Edge
Coordinating Committee meeting on verification and validation,
October 23 in Denver. In particular, members of the Committee
proposed solved physics problems that would be suitable for verifying
edge simulation codes.

CPPG Group

At the APS meeting last week in Denver, CO: Ravi Samtaney presented
an invited talk entitled “Tokamak Pellet Fueling using 3D Adaptive
Mesh Refinement.” D. McCune and S. Either made poster presentations,
and McCune also conducted a satellite TRANSP User Group meeting with
10 attendees representing 6 institutions.

S. Jardin attended a DOE meeting in Santa Cruz, CA, on “Frontiers of
Extreme Computing.” There he presented the keynote application
invited presentation on “Future Requirements for Fusion Computations.”

Week of 10/21/05

Theory Group

Dr. Wei-Li Lee gave a talk on “Plasma Turbulence Research on MPP
Platforms” at Fall Creek Falls Conference sponsored by the National
Center for Computational Sciences (NCCS) of ORNL (10/16-18/2005).
The talk is focused on our march toward ITER simulations using
leadership high-end computing facilities. This annual conference is
intended to bring together the researchers and vendors with the NCCS
staff from ORNL, and faculty and graduate students from the
university partners to discuss the latest developments in this
critical area of scientific and engineering research.

CPPG Group

The latest version of the GTC code, which contains recent vector
optimizations implemented by S. Ethier for the ORNL CRAY X1E, has
been run on the Japanese Earth Simulator computer by Dr. Jonathan
Carter of NERSC, currently visiting the ES Center. This work is part
of an on-going collaboration with the Future Technologies group at
LBL/NERSC, lead by Dr. Leonid Oliker. GTC was able to run with over
13 billion particles, and achieved an unprecedented speed of 7.2
Teraflops on 4096 processors with near perfect parallel scaling.

Using the NTCC XPLASMA software, TRANSP has been modified to create a
portable package of heating and fueling data for an external
predictive transport solver module. The package includes particle
source, ion heating, electron heating, and current drive profiles due
to neutral
beams, fusion alpha heating, and various types of RF heating and
current drive. This method will be tested in PTRANSP and is expected
to serve as a prototype for the Fusion Simulation Project.

Week of 10/14/05

Theory Group

The Theory Department presented its Five Year plan for a peer review
on October 10 and 11, to a distinguished panel, chaired by Dr. Bolton
of OFES. The panel included members of the OFES Theory team and
experts in theory and modeling. There were fourteen presentations,
covering topical science areas, advanced computing, support of PPPL
projects and off-site collaborations.

Three posters were presented by the theory group at the 15th
International Stellarator Workshop, held October 3-7 in Madrid: “SVD
Methods for Magnetic Diagnostics Design in NCSX,” by N. Pomphrey,
E.A. Lazarus, A. Brooks and M. Zarnstorff; “Derivatives of the Local
Ballooning Growth Rate with Respect to Surface Label, Field Line
Label and Ballooning Parameter,” by S. Hudson; “ Equilibrium Flux
Surface Calculations for W7AS and NCSX,” by A. Reiman, M. Zarnstorff,
A. Weller, J. Geiger, D. Monticello, S. Hudson, L. Ku and the W7-AS
Team. Two invited papers were given by international collaborators:
“Significance of MHD effects in stellarator confinement,” by A.
Weller et al, and “Properties of ballooning modes in the Heliotron
configurations,” by N. Nakajima, S. R. Hudson and C. C. Hegna.

Dr. Boozer from Columbia University presented a Theory Seminar on
Thursday, October 13th on the topic of “Magnetic reconnection in non-
toroidal plasmas.” It was demonstrated that the evolution of a
magnetic field with only point nulls is always locally ideal, which
limits the nature of reconnection in non-toroidal plasmas. Dr.
Boozer also showed that rapid magnetic reconnection occurs if the
length of the field lines is greater than about 20 times the Lyapunov
length.

CPPG Group

The biggest obstacle in porting the 3D Extended MHD code M3D to the
ORNL Cray X1E has been the performance of the elliptic solvers, which
consume about 90% of the running time of M3D. The existing PETSc
solvers were unable to utilize the vector registers on the Cray X1E,
and consequently were slower on the Cray X1E than on the NERSC IBM
SP3 computer
(Seaborg). Jin Chen and Josh Breslau (PPPL theory) have developed a
new solver based on the Dynamic Relaxation (DR) algorithm that
vectorizes well, and speeds up the typical M3D elliptic solver so
that it is now 3-6 times faster on the Cray X1e than on Seaborg. We
are in the process of working with the PETSc developers to make this
an option in future PETSc releases.

Dr. Ovsei Volberg of the Tech-X Corporation visited the CPPG staff
and presented a CPPG seminar on “Code Modernization and Coupling for
Future Integrated Modeling.” Discussions centered on utilizing some
of the concepts and accomplishments from the Tech-X SBIR on this
topic in the PTRANSP and Fusion Simulation Project activities.

Week of 10/7/05

CPPG Group

Stephane Ethier attended the NERSC Users Group meeting at LBNL and
made a presentation about the DOE Greenbook, which outlines the
present and future computational requirements for all the offices of
the DOE Office of Science. A copy of the Greenbook can be found at
http://www.pppl.gov/pub_report/2005//PPPL-4090.pdf.
PPPL Graduate Student Slava Lukin, presented a CPPG Seminar on “MHD
modeling on a logically rectangular curvilinear adaptive grid.”
Jae-Min Kwon, a post-doctorial research scientist from KBSI(Korea),
will be at PPPL working with D. McCune(CPPG) and CS Chang (NYU) for
the next two years, beginning on Oct 3, 2005. His research agenda
includes: extending the TRANSP neutral beam package NUBEAM to enable
Monte Carlo simulation of RF-resonant beam-injected and fusion
product fast ions, and also learning to use(FusionGrid) TRANSP for
future KSTAR applications.

Week of 9/30/05

Theory Group

G. Rewoldt, L. -P. Ku, and W. M. Tang have a paper entitled,
“Comparison of Microinstability Properties for Stellarator Magnetic
Geometries,” accepted for publication in Physics of Plasmas. There
they compared the linear properties of microinstabilities driven by
ion temperature gradients and trapped-electron dynamics in nine
distinct magnetic geometries corresponding to several variants of the
NCSX design, as well as to the existing or planned LHD, W7-X, HSX,
and QPS experiments. These studies provide useful insights into the
role of magnetic curvature and of the presence of trapped particles.

Igor Kaganovich visited University of Texas at Austin and gave two
talks; “Fast Evaluation of Ionization Cross Sections for Ion-Atom
Collisions in High Energy Ion Beams” to UT atomic physics group; and
“Interaction of Intense Ion Beams with Background Plasma:
Applications to Heavy Ion Inertial Fusion” at Institute of Fusion
Studies.

Leonid E. Zakharov gave a talk on “Heat removal from the plasma
facing Majeski-Kaita liquid lithium tray” to SOFE05 at Knoxville. He
presented a theory and numerical simulations explaining the
extraordinary heat transport from the small e-beam spot by the fluid
convection, which was generated in experiments on CDX-U by the
Marangoni effect due to the dependence of the surface tension on the
temperature.

CPPG Group

D. McCune, S. Jardin, and D. Stotler (Theory) participated in the
first EU-US Conference Call on Collaborative Activity between the
Integrated Fusion Modelling Initiatives. Among the items discussed
were to internationalize the development of a Module Library, revised
NTCC standards, machine static data files, and plans for an
international conference on integrated modeling.

A modification to TRANSP architecture was successfully tested, using
NUBEAM, to allow adapted NTCC modules to be called out as separate
server processes. This enables a serial TRANSP simulation to access
results of remotely run parallel calculations, and provides a method
for incremental MPI parallelization of TRANSP simulations.

S. Ethier presented a two-day mini-course on “Introduction to
Parallel Programming with MPI” on the main campus of Princeton
University as part of the PICASso program. About 30 graduate
students, from a number of different departments, attended the 4-hour
course.

Ravi Samtaney visited the Applied Numerical Algorithms Group at LBNL
from Sept 26-28. The purpose of the visit was to implement a non-
local electron heat flux model into the AMR MHD code used to study
pellet injection in tokamaks. This work is a part of an ongoing
collaboration between the APDEC and CEMM SciDAC centers.

Wall-to-wall carpet was applied to the PPPL High Resolution Display
Wall room in order to reduce acoustic echoing. Computer division (C.
Scimeca) ran “Access Grid” tests to GA and KBSI/Korea, which showed
marked improvement in the audio connectivity. The Display Wall room
is equipped to serve as a prototype remote control room for experiments.

Week of 9/23/05

Theory Group

Dr. Gwangson Choe attended the 11th European Solar Physics Meeting
held in Leuven, Belgium on 11-16 September 2005. In the Coronal Mass
Ejections and Space Weather Session of the meeting, Dr. Choe gave an
oral presentation entitled "Formation of Erupting Flux Ropes by
Reconnection Among Smaller Scale Flux Tubes." In this paper, he
showed that the more magnetic helicity is involved in a system, the
flux rope created by magnetic reconnection tends to have a longer
connection length. This study suggests that a large scale flux rope
in coronal mass ejection does not emerge in itself from below the
solar surface, but is created by magnetic reconnection in the pre-
eruption stage or during the eruption.

Dr. Phillip Lauber from IPP Garching (Germany) presented a Theory
Seminar on Tuesday 20th, 2005 on " Linear gyrokinetic calculations
on the kinetic properties of shear Alfven modes in tokamak plasmas".
Dr Lauber described the linear gyro-kinetic, electromagnetic code
LIGKA, which is based on an eigenvalue formulation and the drift-
kinetic HAGIS code for the accurate orbit calculations in general
geometry. He suggested that the newly developed code can be used to
study the properties of the kinetically modified TAE mode and its
coupling to the continuum spectrum. Dr Lauber will be visiting PPPL
until early December.

On Thursday, the Theory Department held its monthly Micro-seminar.
The speakers were Dr. Stuart Hudson, who spoke on, "Direct
calculation of ballooning growth rates with respect to pressure-
gradient, shear, and ballooning coordinates", and Dr. Harry Mynick,
who gave a talk entitled, "Topics in stellarator transport".

CPPG Group

Following the initial port of the GTC code to the ANL Blue Gene/L
computer by Stephane Ethier, IBM's BG expert Dr. Bob Walkup further
optimized GTC for this platform by introducing calls to the vector
library MASSV and other BG-specific changes. The optimizations
resulted in a 30-60% performance improvement compared to the original
timings. With access to IBM's own BG/L system, a weak scaling study
of GTC was carried out on up to 16,384 processors and showed perfect
scalability of the code, with peak speeds exceeding 2 Tflops.

S. Ethier and S. Jardin contributed to the gathering of information,
the writing, and the editing of the 2005 DOE Greenbook: Needs and
Directions in High Performance Computing for the Office of Science
(SC). This document, with a total of over 37 contributors from each
of the SC Program areas describes how computational science has been
used to advance each of the disciplines, and what advances might be
possible in the near future with increased computational resources.
S. Jardin presented this to a NERSC review team in June, and S.
Ethier will present it to the NERSC Users group next week. The
Greenbook is now available as a PPPL report, PPPL-4090 and can be
found on the PPPL web site.

Week of 9/16/05

Theory Group

Dr. Philip Lauber from IPP, Garching, Germany, has begun a three-
month long visit to the Theory Department (Sept. 8 to Dec. 8). During
the visit, he will collaborate with Dr. Guoyong Fu and others to
benchmark his recently developed gyrokinetic code for Alfven waves
and its applications to internal kink mode and Alfven eigenmodes in
tokamaks.

A new cluster named 'kite' has been started this week. It includes 24
SUN V20z servers, each with two 2.4 GHz AMD Opteron processors and
6GB of RAM (for a total of 48 CPUs and 144 GB of RAM). The servers
are connected with an Infiniband cluster interconnect (which provides
very high throughput--10Gb rated performance--and very low latency)
and also with a standard 1Gb Ethernet cluster interconnect. A special
MPI implementation has been created and tested which automatically
takes care of using Infiniband for message passing. The servers are
running the 64bit version of RedHat Enterprise Linux 3, Update 5, and
are performing very well on initial benchmark tests. A new queue
called 'kite' has been set up in the PBS scheduler.

CPPG Group

S. Jardin, together with V. Chan (GA), L. Lodestro (LLNL), D.
Batchelor (ORNL), C. S.Chang (NYU), and A. Kritz (Lehigh) formed the
U.S. Delegation to the US-Japan JIFT Workshop on Integrated Modeling
of Multi-Scale Physics in Fusion Plasmas with Participants from EU
and Korea, held at RIAM, Kyushu University, Fukuoka, Japan. He made
a presentation "Towards Integrated Modeling of Burning Plasmas".
Copies of this, and the other presentations at the workshop can be
found at the conference web site:
http://bpsi.nucleng.kyoto-u.ac.jp/bpsi/usjws3/agenda.html
One noteworthy feature of the conference was the widespread mention
by the other parties of modules downloaded from the NTCC modules
library, and incorporated into their integrated modeling codes.

Three new, modern dual processor dedicated TRANSP production servers
are validated and now in production, and are being utilized as needed
to meet PPPL and FusionGrid TRANSP user requests. This supports six
dedicated TRANSP job streams similar in performance to the multi-
purpose job streams on the recently acquired PPPL sunfire servers-all
approximately 2-3 times faster than PPPL Linux job servers in use a
year ago.

Week of 9/9/05

Theory Group

A paper on simulations of plasma turbulence in astrophysical
accretion disks has been submitted for publication by graduate
student P. Sharma. G.W. Hammett, E. Quataert (Berkeley), and J.M.
Stone (Princeton) are co-authors. Collisionless effects may be
important in low luminosity accretion flows, such as near the black
hole in the Milky Way's Galactic center, and were modelled by
extending the ZEUS MHD code to include an anisotropic pressure tensor
and parallel heat conduction. Cyclotron and mirror
microinstabilities set limits on the pressure anisotropy. The
anisotropic pressure tensor provides a qualitatively new mechanism
for transporting angular momentum in accretion flows and is found to
be comparable to the Maxwell stress, leading to moderate enhancements
in the rate of angular momentum transport over the ideal MHD limit.
("Shearing Box Simulations of the MRI in a Collisionless Plasma,"
http://www.arxiv.org/abs/astro-ph/0508502 or http://www.pppl.gov/
pub_report//2005/PPPL-4098-abs.html.)

Dr. Gwang-Son Choe visited the Geophysical Institute at the
University of Alaska, Fairbanks, AK from 29 August to 2 September
through PPPL-University Collaborations. Gwang-Son Choe and Jay R.
Johnson at PPPL have been collaborating with Prof. Antonius Otto at
UAF on dynamics of partially ionized plasmas in space. Specifically,
they are developing a three fluid (neutral-ion-electron) global
simulation code to be applied to the solar photosphere, chromosphere
and prominences as well as to the earth's ionosphere. During the
visit, Dr. Choe presented a seminar entitled "Opening of Magnetic
Fields in Coronal Mass Ejections."

A National Stellarator Theory Teleconference was hosted by Dr. A.
Reiman on Thursday, September 8. The teleconference focused on
ballooning modes, with a presentation by Dr. S. Hudson (PPPL) on
"Direct calculation of the ballooning eigenvalue derivatives with
respect to
surface label, field line label and ballooning parameter," and a
presentation by Dr. A. Ware (Univ. Montana) on "Ray tracing for
ballooning modes in quasi-symmetric devices." The presentation
materials are available at http://ncsx.pppl.gov//Physics/NSTT/
Theory_Index.html
This site also contains an archive of presentation materials from
previous stellarator theory teleconferences.

Dr. T.S. Hahm attended the 5th general scientific assembly of Asia
Plasma and Fusion Association on Fusion Science and Technology, which
was held in Jeju City, Korea on August 29-31, and presented an
invited talk on the status of modern nonlinear gyrokinetic theory.
He also gave lectures on nonlinear-gyrokinetic theory at the
International School on Plasma Transport and Turbulence, which was
held in Hangzhou, China, Aug 31-Sept 3.

Week of 9/2/05

Theory Group

Ron Davidson is presenting an invited paper on recent technical
progress in the U. S. heavy ion research program on high energy
density physics and fusion at the Inertial Fusion Sciences and
Applications (IFSA) conference in Biarritz, France, September 5 - 9.
The abstract for this paper is: "Key scientific results from
recent experiments, modeling tools, and heavy ion accelerator
research are summarized that explore ways to investigate the
properties of high energy density matter in heavy-ion-driven
targets, in particular, strongly-coupled plasmas at 0.01 to 0.1
times solid density for studies of warm dense matter, which is a
frontier area in high energy density physics. Pursuit of these
near-term objectives has resulted in many innovations that will
ultimately benefit heavy ion inertial fusion energy. These include:
neutralized ion beam compression and focusing, which hold the
promise of greatly improving the stage between the accelerator and
the target chamber in a fusion power plant; and the Pulse Line Ion
Accelerator (PLIA), which may lead to compact, low-cost modular
linac drivers." The conference proceedings are to be published
in the journal Fusion Science and Technology.

A classical trajectory method has been developed by Igor Kaganovich
for the calculation of charge-changing cross sections. National
Undergraduate Fellowship (NUF) student, Thomas Bender, has applied
this method to evaluate the cross sections for hydrogen in
collisions with fully stripped projectiles: H+, He+2, Li+3, C+6.
These simple systems were chosen to benchmark the method against
available experimental data, and the theoretical predictions are in
very good agreement. The method can be used for projectile velocity,
v, one to three times larger than the orbital velocity of hydrogen
vnl. At small projectile velocities, v<vnl,and at very large
projectile velocities, v>>vnl, quantum effects become
important. In this case, a "hybrid" method, developed
previously, can be applied.

CPPG Group

The national Fusion Collaboratory (FusionGrid) is deploying 4
dedicated dual Linux servers (8 processors) at PPPL.  These
will largely be used to support FusionGrid TRANSP run production in
the period leading up to APS, as well as providing dedicated
platforms for testing of new FusionGrid computational services. 
Based on past measurements on similar machines, these servers are
expected to have serial speeds about 3x faster than most of the
TRANSP/FusionGrid servers in use at PPPL a year ago.

The NSTX Display Wall is being upgraded with Fusion Collaboratory
software developed at the Princeton University Computer Science
Department.  The new software allows multiple
mouse/cursor/keyboard control streams to be shared through a single
display, thereby allowing multiple users to simultaneously and
independently interact with the system.  The new software will
be available for the last two weeks of the NSTX experimental
campaign.

PPPL researchers submitted 6 proposals to the National Energy
Research Supercomputer Center (NERSC) for computer time for FY06,
totaling over 5.3 million CPU-hours.  The requests were for
"Nonlinear Delta-f Particle Simulation of Collective Effects
for Heavy Ion Fusion Drivers and high Intensity Particle
Accelerators", 100,000; "Simulations of Field-Reversed
Configuration and Other Compact Tori Plasmas", 200,000;
Analysis and Design of Stellarator Configurations", 200,000; 3D
Extended MHD Simulation of Fusion Plasmas", 100,000; Tokamak
fueling via pellet and supersonic gas injection using AMR",
358,000; and "Center for Gyrokinetic Particle Simulations of
Turbulent Transport in Burning Plasmas and Multiscale
Mathematics", 3,500,000.  A total of 4 million NERSC
CPU-hours have already been used by PPPL in FY2005.

Prof. R. Bramley (U.Indiana) and Dr. D. Batchelor (ORNL) visite
PPPL for 2 days of planning meetings for the new Fusion Simulation
Project activity "Simulation of Wave Interaction with MHD
(SWIM)".  Initial teams have been formed in each of the
activity areas with the immediate tasks of defining software
interfaces between component modules. The entire SWIM team plans to
hold a meeting in November at ORNL.

Week of 8/26/05

CPPG Group

Dr. Andy Bauer, from the Scientific Computation Research Center at
the Rensselaer Polytechnic Institute, spent the week at PPPL working
with CPPG staff on modifying the M3D-C1 code to accept meshing
routines from the unstructured adaptive mesh refinement package in
Trellis. The adaptivity from Trellis, combined with the high-order
accuracy, high-continuity, and implicit time differencing in M3D-C1,
should lead to a very efficient package for analyzing global
macroscopic dynamics involving localized reconnection layers. This
visit was part of the collaboration between the two SciDAC centers
CEMM and TSTT.

Week of 8/19/05

Theory Group

Professor Hiroshi Naitou of Yamaguchi University is visiting PPPL
(8/9 - 8/26/05) under the sponsorship of US/Japan exchange on Fusion
Theory. He and his host, Dr. Wei-li Lee, will continue their
collaboration on gyrokinetic particle simulation of MHD modes, which
started nearly ten years ago. His visit is very timely in view of the
new multiscale gyrokinetic particle simulation project funded by
DoE. Prof. Naitou also gave a Theory seminar on "Gyro-Reduced-MHD
Simulation of Kinetic Internal Kink Modes." The abstract is as
follows: The simulation of kinetic internal kink modes using gyro-
reduced-MHD equations will be presented. The basic equations are the
moment equations obtained from the original gyrokinetic Vlasov-
Poisson-Ampere system. It is found that a three-field model can
accurately describe the linear and nonlinear evolution of these modes
with the stabilizing electron diamagnetic effect. However, this
stabilization is found to be incomplete and the residual unstable
modes can generate vortices due to the Kelvin-Helmholtz-like
secondary instablity. The strong coupling between the kinetic
internal kink modes and the resulting vortices may explain the
mysterious sawtooth related phenomena. The five-field model including
the Landau closure is developed to estimate the effects of the ion-
Landau-damping on these modes. The necessity of benchmarking between
the present code and gyrokinetic particle codes will be discussed.

CPPG Group

At the request of users, TRANSP has been upgraded to allow
"steering," i.e. restart of interrupted TRANSP simulations with
modifications to input data. An interrupted TRANSP run can be
"cloned" under a new name, allowing several "variant" runs to be
constructed off a common base run. The base simulation results are
copied, allowing a savings of run time. A single run can be
interrupted and steered multiple times. Merged input datasets for
steered runs are archived, allowing for accurate re-execution of
steered runs with future versions of the code, e.g. for regression
testing, or to take advantage of code improvements for important
simulations. The steering capability is expected to be useful in the
testing and development of the PTRANSP predictive simulation capability.

Week of 8/12/05

Theory Group

Dr. A. Reiman gave a talk on "An Overview of the PIES Code" for an
Integrated Modeling Teleconference organized by GA on Wednesday,
August 10. The PIES code is a 3D MHD equilibrium code, and the
meeting focused on issues related to the DIII-D I-Coil Perturbation
and error field experiments.

Dr. X. Tang visited PPPL from Aug 8-12. He gave a seminar entitled,
"Magnetic relaxation in laboratory and astrophysical plasmas." He
discussed, formation and sustainment of system-scale magnetic field
by relaxation, which is ubiquitous in laboratory (RFP, Spheromak,
Spherical Torus under CHI) and astrophysical (giant radio lobes)
plasmas. He presented the physical implications of Taylor's helicity-
conserving minimum energy state in laboratory and radio lobe plasmas
in terms of two classes of resonance phenomena, which was
demonstrated using Chandrasekhar-Kendall force-free eigenmodes. The
linear resonances were regularized by one of, plasma inertia, finite
pressure, and non-uniform normalized parallel current density. The
emergence of bifurcated regularized solutions was shown, along with
their physics interpretation and significance. Finally, the nonlinear
dynamics that led to magnetic relaxation was investigated using
initial value 3D MHD simulations. The relaxation of the driven plasma
was shown to follow a helical instability cascade.

CPPG Group

J. Chen, S. Ethier, and S. Jardin, together with J. Breslau and G. Fu
(PPPL Theory) attended a 3-day tutorial and workshop at ORNL on the
Crays X1E and XT3. The following talks were given by PPPL: "Running
M3D on Advanced computing architectures" (J. Chen), "Meshing and data
structure of M3D," (J. Breslau), "Vector performance issues with
GTC" (S. Ethier), and "M3D-Hybrid Option," (G. Fu). Plans for a
joint proposal for National Leadership Class Computer time for FY06
were also discussed, as well as issues related the kickoff of the
Simulation of Wave Interaction with MHD (SWIM) Fusion Simulation
Project.

Will Fisher, a summer intern in the Science Education Program,
presented his project: "Integrating Command Line Programs with Java
Clients." He completed this work with Eliot Feibush of CPPG as his
advisor. This work will be of immediate benefit to the TSCPLOT
program, which can now be invoked within the ElVis plotting program.

Week of 8/5/05

Theory Group

Drs. T. S. Hahm, G. Hammett, N. Fisch and W. W. Lee gave talks at the
Workshop on Nonlocal, Collisionless Electron Transport in Plasmas
held at PPPL, Aug. 2-4, 2005. Dr. T.S. Hahm gave a talk entitled
"Theory, simulation, and experimental test of turbulence spreading."
Dr. Hammett gave a talk entitled "Nonlocal fluid closure
approximations for collisionless dynamics with nonlinear extensions."
Dr. N. Fisch gave a talk entitled "Kinetic effects and current drive
in high temperature plasmas." Dr. W. W. Lee gave a talk entitled "How
to Reduce Noise in Particle Simulations."

JOINT PPPL/GA THEORY HIGHLIGHT:
In joint work with GA, further improvements to the 2D Green's
function calculation of the vacuum energy in delta-W stability codes
were made, enabling calculation for any practical value of the
toroidal mode number and distance of the plasma-wall separation.
There is considerable overlap with the applicability of previous
these techniques to assure cross-checks and accuracy. This work was
done by Dr. M. Chance of PPPL and Dr. A. Turnbull at GA.

CPPG Group

An initial conference call was held amongst the co-PIs (D. Batchelor,
ORNL; S. Jardin, PPPL; R. Bramley, U. Indiana; D. Keyes, Columbia U.)
of the newly funded proposal "Simulation of Wave Interaction with
MHD" (SWIM). This proposal was selected as part of the DOE Notice DE-
FG01-05ER05-11 (Scientific Discovery through Advanced Computing-
Fusion Simulation Prototype Centers). The funding was awarded jointly
by the Office of Fusion Energy Sciences (OFES) and the Office of
Advanced Scientific Computing Research (OASCR).

In connection with the P-TRANSP project, the TRANSP Data Module has
been extended. The module can be used to give predictive codes
access to the same experimental input data used by TRANSP, without
recourse to TRANSP internal data structures. In addition to
temperatures and densities, the data module now contains the time
evolution of beam powers and RF antenna
powers; sawtooth and pellet event times have been added as well.

Doug McCune attended the SciDAC-2 Workshop on Advanced Computing for
Experimental Science, held at Argonne National Lab. The needs of
ITER figured prominently in the discussion, both in terms of the need
for remote access to control rooms and collaborative experimental
operations, as well as the likely need for high performance,
distributed, near-real-time data analysis capability. The similarity
of the expected requirements of ITER and the Large Hadron Collider
(LHC) international collaborations were noted.

Week of 7/29/05

Theory Group

On July 28, Dr. Matthew Hole of Australian National University, ANU,
gave a talk entitled, 'Stepped pressure profile equilibria in
cylindrical plasmas via partial Taylor relaxation.' In this talk he
addressed magnetohydrodynamic (MHD) equilibrium states in three-
dimensions (3D) with smooth pressure profiles which have long
bedeviled containment theory due to the formation of magnetic islands
at rational surfaces, resulting in pressure flattening. In this work,
a new model was presented: the stepped-pressure profile equilibrium.
The system comprises multiple Taylor relaxed plasma regions, which
are separated by ideal MHD barriers. Such a model is well posed
mathematically, and follows rigorous existence proofs of MHD
equilibria with stepped pressure in weakly toroidally asymmetric
plasmas. In addition to a description of 3D equilibria, the model is
also motivated by observations of internal transport barrier
formation at irrational flux surfaces in magnetic confinement
experiments. He suggested that this formulation may provide a minimum
energy explanation for the existence of ITB's. In cylindrical
geometry, where analytic progress is possible, they have constructed
equilibrium solutions, and analyzed stability by a variational
formulation. We show the existence of tokamak-like equilibria, with
increasing smooth safety factor and stepped-pressure profiles. Unlike
reverse field pinch q profiles, only the plasma core necessarily has
reverse magnetic shear. Recent observations of ITB formation at
minima in the safety factor provide some preliminary guidance for
these calculations. This work is done as part of an ongoing PPPL-ANU
collaboration. Dr. Stuart Hudson is the PPPL collaborator

A National Stellarator Theory Teleconference was hosted by Dr. Allan
Reiman on Thursday, July 28. Dr. Michael Zarnstorff gave a
presentation on "Reconstructing the Equilibrium of W7-AS."

CPPG Group

Jin Chen participated in the NERSC User Group Conference call. It
was announced that the new 640-CPU Opteron cluster with a high-speed
InfiniBand network, Jacquard, will be in a production state effective
August 1. Every user of Seaborg will have an account on Jacquard.
It is roughly four times as fast as Seaborg, per node. The deadline
for applying for NERSC time for next year is September 13.

Week of 7/22/05

Theory Group

 L.E.Zakharov gave a talk "At the ground level of
integrated modeling" to Symposium for the Future of Integrated
Modeling. Two new developments, i.e., (a) free boundary equilibrium
reconstruction (using JET data) linked with transport simulations,
and (b) the "real time forecasting" mode (with the
calculation speed comparable to the simulated ITER scenario), were
demonstrated using existing predictive transport code ASTRA (working
in the "code talking" regime with ESC).



Dr. Matthew Hole of the Australian National University is visiting
PPPL for three weeks. During this time, he and Dr. Stuart Hudson of
the Theory Department will be discussing a new model of
three-dimensional MHD equilibrium construction. The equilibrium is
determined by a stepped pressure profile, and is motivated by recent
mathematical proofs that solutions to this model exist. Within each
constant pressure region, the magnetic field is assumed to have
relaxed to a force-free Beltrami field.

CPPG Group

 S. Jardin and D. McCune attended the Symposium on the
Future of Integrated Modeling.  Jardin made a presentation on
"An Integrated Modeling Framework for Fusion", and McCune
made a presentation on "Fortran-90 Based Modernization of
TRANSP".  These and other presentations will be available
on the project web site: http://www.mfescience.org/kritzfest
<http://www.mfescience.org/kritzfest>



Holger St. John, General Atomics, conducted a workshop at PPPL
discussing integration of GA's GCNM (Globally Convergent Newton
Method) into PPPL transport codes, PTRANSP in particular.  This
routine provides a powerful way to predicatively integrate transport
equations using the current generation of "stiff" critical
gradient predictive models for the transport coefficients, such as
GLF23, MMM95, and expected successors.

Week of 7/15/05

Theory Group

The electrostatic two-stream instability for a cold, longitudinally-
compressing charged particle beam propagating through a
background plasma has been investigated both analytically and
numerically. Small-signal coupled equations describing the evolution
of the amplitudes of the perturbation are derived and the asymptotic
solutions are obtained. The particle-in-cell delta-f code BEST has
been used to carry out detailed numerical studies of the instability,
and initial results are in reasonably good agreement with the theory.
We have found that the longitudinal beam compression strongly
modifies the space-time development of the instability. In
particular, the dynamic compression leads to a significant reduction
in the growth rate of the two-stream instability compared to the case
without an initial velocity tilt.

T.S. Hahm participated in the Festival de Theorie on "Turbulence
overshoot and resonant structures in fusion and astrophysical
plasmas" from July 4 to July 15, held in Aix-en-Provence, France.
This was the third meeting of the Festival de Theorie, which is held
every other year, organized jointly by CEA Cadarache and CNRS-
University of Provence. He gave a tutorial talk on the nonlinear
gyrokinetic turbulence, a lecture for students and postdocs on
turbulence spreading and shear flows, and led a discussion session on
turbulence overshoot in fusion plasmas and star convection zones.

CPPG Group

A significant milestone in the TSC/TRANSP coupling project has been
met. A TSC predictive calculation now has the option of obtaining
physics data from the TRANSP code in real time, as the calculation
proceeds. This has been used to import the experimental density
profile when performing a TSC simulation of an existing experimental
discharge. Future applications will be to import results from the
TRANSP RF, fusion particle, and Neutral beam calculations into a TSC
simulation to provide a more realistic, free boundary simulation
capability with the most advanced energy and density source models.
Results from the TSC/TRANSP package were reported last week in two
presentations at the IEA W60 Burning Plasma Physics meeting in
Tarragona, Spain.

Week of 7/8/05

Theory Group

Drs. N. N. Gorelenkov and W. W. Lee attended the EPS meeting and the
W60-Burning Plasma
workshop in Tarragona, Spain. At the EPS meeting Dr. Gorelenkov
presented recent findings on the instability of kinetic ballooning
modes in ITER, which are stable without fusion alphas, but are
excited due to the pressure gradient of fast ions. At the joint
session of the burning plasma workshop and ITPA expert group meeting
he reported on the predictive modeling of the AE linear stability and
quasilinear diffusion model for different ITER plasmas. It was shown
that the quasilinear diffusion model predicts AEs to be benign for
the nominal ITER plasmas. However rising plasma temperature can lead
to the strong losses of fusion alphas.

Dr. Wei-li Lee presented a poster on "Steady State Gyrokinetic
Particle Simulation of Ion Temperature Gradient Instabilities." He
also gave a talk at the workshop on Burning Plasma Physics and
Simulation on "Integrated Simulation of Fusion Plasmas."

CPPG Group

S. Jardin attended the workshop W60 on "Burning Plasma Physics and
Simulation" under the auspices of the IEA Large Tokamak Implementing
Agreement in Tarragona (Spain) and served as Chair and Discussion
leader of the 2-day session on "Integrated Modeling of Burning
Plasmas." It was concluded at the meeting that the international
community needs to continue to develop a hierarchy of codes with a
range of speed and physics. These include reliable validated
transport-timescale code packages with improved modules for all
processes and with reliable ranges of validity, improved fundamental
"first principles" nonlinear models, and coupled fundamental models
to examine strongly interacting physics issues. To obtain this
objective, each of the international partners should be encouraged to
continue their modeling initiatives, with increased emphasis on
verification and validation at all levels. A report on the workshop
will be issued within the next several months.

Week of 7/1/05

Theory and CPPG Groups

PPPL's Theory Department and CPPG were prominently involved at
the Scientific Discovery through Advanced Computing (SciDAC) 2005
Conference in San Francisco, CA from June 26-30.  In addition
to a well-received plenary session presentation by Dr. Stephane
Ethier on "Gyrokinetic Particle-in-Cell Simulations of Plasma
Microturbulence on Advanced Computing Platforms," invited
presentations from PPPL included contributions from  Drs. G.
Rewoldt (on behalf of Prof. Z. Lin of UCI) on "PIC Simulations
of Electron Transport from Plasma Turbulence," from S. Klasky
on "Data Management on the Fusion Computational Pipeline,"
from M. Adams (Columbia U. & PPPL) on "The Link between
Gyrokinetic PIC Simulations & Micro-FE Analysis in Orthopedic
Biomechanics,"  from W. Wang on "New Insights from
General Geometry Simulations of Plasma Transport," and from R.
Samtaney on "Adaptive Mesh Semi-Implicit Method for Resistive
MHD."

Week of 6/24/05

Theory Group

Dr. Bruce D. Scott from Institute for Plasma Physics, Garching, Germany
gave a talk on June 23, 2005, entitled, Global Gyrofluid Computations
of Electromagnetic Core Turbulence. He described computations using
the six-moment GEM code for turbulence using tokamak core parameters,
including the correct electron/deuterium mass ratio. Electrostatic
results were shown and compared to previous work. Electromagnetic
results from the same cases show some strong effects, narrowing the
spectrum and weakening the transport. Diagnostics illuminated scale
separation between turbulence and zonal flows, and the equilibrium
profiles and mean flows. The MHD equilibrium was dynamically, self
consistently computed within the S-alpha model. ETG results on domain
size and finite beta effects were also briefly mentioned.

CPPG Group

CPPG members Ravi Samtaney and K. Indireshkumar (Kumar) attended the
Computational Science Graduate Fellowship (CSGF) Conference in
Washington, D.C. from June 21 - 23. They presented a poster
highlighting the computational plasma physics research at PPPL.

Week of 6/17/05

Theory Group

Hong Qin received the Presidential Early Career Award for Scientists
and Engineers at a ceremony in Washington, D.C., Monday, June 13.
The Presidential award is the highest honor bestowed by the U.S.
Government on outstanding scientists and engineers who are beginning
their independent careers. Qin also received the DOE's Office of
Science Early Career Scientist and Engineer Award at a separate
ceremony June 13 at DOE Headquarters. He was among six from DOE
national laboratories to receive the Office of Science honors, as
well as the Presidential award. Both the Presidential and DOE awards
cite Qin for his contributions to the physics of high-intensity
particle beams, with application to ion-beam fusion energy, and for
his work on electromagnetic effects in magnetically confined plasmas,
with application to magnetic fusion energy.

A paper by Dr. J. Lewandowski describing a new method to implement
collision operators in particle in cell (PIC) codes has been
published in the May issue of Physics of Plasmas. The method is based
on the fact that relevant collision operators can be included
naturally in the
Lagrangian formulation, which exemplifies PIC codes, does not rely on
the generation of random numbers. It is shown that with the new
collision operator the simulated plasma has significantly better
momentum and energy conservation properties as compared with standard
collision operators based on random numbers.

Dr. Shuanghui Hu of University of California at Irvine completed a
two-week visit to collaborate with Dr Guoyong Fu in the area of
energetic particle-driven Alfven modes in high beta tokamak plasmas.
He had useful discussions with Drs Guoyong Fu, Stuart Hudson, Nikolai
Gorelenkov and Eric Fredrickson on generalization of his 1D s-alpha
model of alpha-TAE to 2D model with general equilibria as well as
experimental implications of alpha-TAE.

The Theory Department held its monthly Micro-Seminar on Thursday,
June 16. The Speakers were Dr. Hong Qin and Dr. Edward Startsev. Dr.
Qin introduced the on-going research on nonlinear delta-f particle
simulation of high intensity beams, with applications to heavy ion
inertial fusion, high energy density physics, and high intensity
accelerators. The Beam Equilibrium Stability and Transport (BEST)
code was described in detail with examples of simulation results for
the electron-ion two-stream instability and temperature anisotropy
Harris instability. Dr. Startsev discussed the generalization of the
analysis of the classical Harris instability to the case of a one-
component intense charged particle beam with anisotropic temperature.
He showed that the instability is kinetic in nature and is due to the
coupling of the particle's transverse betatron motion with the
longitudinal plasma oscillations excited by the perturbation. For a
long, coasting beam, the delta-f particle-in-cell code BEST has been
used to determine detailed 3D stability properties over a wide range
of temperature anisotropy and beam intensity. The BEST code has also
been used to study the nonlinear evolution and saturation of the
instability. The nonlinear saturation is governed by longitudinal
particle trapping by a spectrum (not just a single wave) of fast-
growing waves with a broad band of longitudinal wave numbers and zero
oscillation frequency. The presence of many waves leads to the
nesting and overlapping of particle resonances in longitudinal phase
space, and as a consequence, to the fast randomization of the trapped-
particle distribution and longitudinal heating.

CPPG Group

Ravi Samtaney visited the National Center for Computational Science
ORNL on June 10. Kenneth J. Roche, who is assimilating a data set
related to Department of Energy applications and missions that are
computationally bound or driven, was his host. The data set will be
primarily timing information retrieved during the execution of
science-based (including current production scale) benchmark runs on
current (and future) state of the art compute systems typically owned
and operated by the DOE SC. The Chombo based AMR MHD code was ported
to the SGI Altix at ORNL and will ported to other platforms such as
the CRAY X1, and benchmarked as part of this exercise.

Week of 6/10/05

Theory Group

Presentations were made at the PSACI PAC Meeting on June 2-3 at PPPL
by Drs. Wei-li Lee, PI of the SciDAC Center for Gyrokinetic
Simulation of Turbulent Transport, and Steve Jardin, the PI of the
SciDAC Center for Extended MHD Modeling.

CPPG Group

S.Klasky, D. McCune, and S.Jardin participated in the Joint ORNL/
Indiana University workshop on Computational Frameworks in Fusion.
Klasky presented a talk on the Center for Plasma Edge Simulation
(CPES) framework in which he discussed the three types of frameworks
necessary in a Fusion Simulation Project, and discussed the data
management framework, which the CPES will use. McCune described the
de facto framework, which the TRANSP code has developed and uses,
including the remote usage of TRANSP via the Fusion Collaboratory.
Jardin discussed a vision for a full Fusion Simulation Project
covering all aspects of simulation of a fusion device. Jardin also
met with the ORNL Cray X1 consulting team to discuss issues
related to the porting and optimization of the M3D and GTC codes on
the Cray X1.

Week of 6/3/05

Theory Group

Daren Stotler participated in an IAEA Technical Meeting on the
"Establishment of A&M Computer Code Network" last week in Vienna,
Austria. The proposed "network" will be anchored by a Web page
hosted by the IAEA Atomic & Molecular Data Unit. This page will
contain information on experts in the production and use of atomic
data, as well as on their codes. A list of A&M data used in fusion
and their origin will be produced and maintained so that data
producers can get some recognition for their efforts. A forum for
information exchange between the A&M experts and the rest of the
fusion community will be incorporated into the Web page. Finally,
the Web page will be linked to the existing GENIE system, an A&M meta-
database capable of sending data requests to the various A&M data
centers around the world and collating their replies for the user.

Gwang-Son Choe and Jay R. Johnson attended the 2005 Joint Assembly of
AGU, SEG, NABS and SPD/AAS held in New Orleans, LA on May 23 - 27.
Gwang-Son Choe gave two oral presentations entitled "Formation and
Expulsion of Large Scale Flux Ropes in Major Solar Eruptions by
Reconnection among Smaller Scale Flux Tubes" and "Roll Effect in
Prominence Eruption Explained by Anisotropic Electrical
Conductivity." Frank Cheng was the co-author of these papers.

JOINT PPPL-GA THEORY:
Drs. Morrell Chance of PPPL and Alan Turnbull of GA have made
significant progress in computing the vacuum contribution for modes
with high toroidal mode number, n, in the ideal delta-W formulation.
The Green's function calculation for the magnetic scalar potential
used in the vacuum codes currently employs a recursion relation to
generate the functions at finit toroidal mode number, n. The
recursion is initiated from the complete elliptic integrals of the first
and second kind. It was recently found that the loss of numerical
precision can be serious even for n about 5 and above for normal
tokamak configurations. In view of this we have crafted a new method
to calculate the Green's function by directly integrating the
relevant integral representation. A judicious treatment of the
singular behavior of the function, together with transformation of the
independent variable, and an accurate quadrature scheme enable a
precision that's much greater than before, even approaching machine
accuracy.

CPPG Group

There were two CPPG seminars this week. Professor Mladen A. Vouk,
Computer
Science, and Associate Vice-Provost for Information Technology, North
Carolina State University presented a seminar on "Next Generation
Scientific Workflows: A case study". This was followed Bertram
Ludaescher (UC Davis) & Ilkay Altintas (San Diego Supercomputing
Center) who spoke on "Overview of Scientific Workflow Automation with
Kepler/SPA". The speakers spent the entire day at PPPL discussing
data management and workflow issues with members of the GPS, CEMM,
and FSP Scidac teams. The SPA team will be automating the workflows
for the AMRMHD code, the GTC code, and the M3D code.

Week of 5/27/05

Theory Group

On May 26, Dr. A. Reiman hosted a National Stellarator Theory
Teleconference. The agenda consisted of a presentation by Dr. J.
Hanson of Auburn University on "Equilibrium Reconstruction in
Stellarators", and a presentation by Dr. Reiman on "The Effect of
Ambipolar Plasma Flow on the Penetration of Resonant Magnetic
Perturbations in a Quasi-Axisymmetric Stellarator".

On May 26, Hong Qin gave the Theory Department seminar, which was
entitled "A footnote on the adiabatic invariants". His discussion
covered the idea that the adiabatic invariant for the time-dependent
oscillator equation is just an asymptotic approximation to an exact
invariant, the Courant-Snyder invariant. A thorough study of the
symmetry and invariance of the related dynamics reveals many deeper,
interesting structures that have important implications. Dr. Qin
pointed out that, we can show that the adiabatic invariant is actually
a stronger invariant than that proved by Arnold (1978), and a more
general one than that proved by Kulsrud (1957) and Kruskal (1961).

CPPG Group

NUBEAM, the TRANSP Monte Carlo fast ion model, has been updated for
improved accounting of momentum losses associated with orbits that
leave the plasma and strike a wall or limiter. The change affects beam
driven JxB torques to the target plasma in the edge region, and
improves the self-consistency of treatment of JxB torques in NUBEAM.
The need for the improvement was discovered through careful comparisons
with other models carried out by the MAST physics team. The model
improvement has been placed in production and is now available to all
TRANSP Fusion Grid users.

Dan Martin from the Applied Numerical Algorithms Group at LBNL visited
PPPL on Monday 05/23/05 and Tuesday 05/24/04, hosted by CPPG member
Ravi Samtaney. They completed a first release of the single fluid
resistive AMR MHD code as part of the Chombo distribution. This code
uses an unsplit upwind formulation and implicit treatment of the
diffusion terms in the resistive MHD equations. Chombo is the adaptive
mesh refinement framework developed at LBNL. This is a collaborative
effort between the APDEC and CEMM SciDAC programs.

Jin Chen attended International Conference on Computational Science
2005 (ICCS 2005) in Atlanta and presented a paper "Solving Anisotropic
Transport Equations on Misaligned Grids". She also had many discussions
with participants regarding optimization on the Cray-X1.

S.C. Jardin attended the International Conference on Parallel
Computational Fluid Dynamics 2005, held at the University of Maryland.
He presented an invited talk "Finite Element Calculations of the
Magnetohydrodynamics of Magnetic Fusion Devices and Magnetic
Reconnection."

John Kichury, Principal Systems Engineer, of Silicon Graphics, Inc.
visited on May 24 to discuss the new SGI Prism computer installed at
PPPL to drive the Display Wall and do complex analysis required for
high-end visualization. He presented a CPPG seminar "SGI Visualization
- Prism Architecture and Performance."

Week of 5/20/05

Theory Group

Greg Hammett gave a talk on May 19 at the Kavli Institute for
Theoretical Physics (KITP) at UC Santa Barbara, on "Current Status of
Fusion Energy Research and Related Plasma Physics Studies." Hammett is
a participant in the KITP spring research program on the Physics of
Astrophysical Outflows and Accretion Disks.

Leonid Zakharov presented the theory of the extraordinary heat
propagation from the e-beam hot spot in the Majesky-Kaita Liquid
Lithium (MKLiLi) tray, recently discovered on CDX-U, at a Boundary
Science Focus Group meeting. For the first time, he addressed the
Marangoni surface tension effect into the MHD modeling of liquid
lithium, which generates the convective flow and efficiently mixes the
fluid on a scale comparable with the size of the tray. The relevant
equation, physical picture, and scaling laws have been outlined with
and without the presence of a strong magnetic field. The results of 3-D
simulations, with a new code 'Cbebm', confirm the conclusion that
MKLiLi is not sensitive to the peaks in the power deposition, which is
spread over the entire tray. These properties of MKLiLi offer new
research opporutunities for divertor studies and spherical tokamaks.

CPPG Group

S. Jardin attended a review of the National Energy Research
Supercomputer Center (NERSC) and made a presentation on "Computational
Challenges and Requirements in the Office of Science" to the review
committee. This summarized material gathered from computational
scientists representing all the program offices within the DOE Office
of Science.

Week of 5/13/05

Theory Group

Prof. Scott Parker from University of Colorado, Boulder gave a theory
seminar on "Gyrokinetic simulation of the collisionless and
semi-collisional tearing mode instability" on May 10, where the
nonlinear evolution of the collisionless and semi-collisional tearing
mode instability is studied using an electromagnetic gyrokinetic
particle-in-cell simulation. He showed that collisionless nonlinear
saturation compares well with existing theory in terms of saturation
level and electron bounce oscillations. Electron-ion collisions were
included to study the semi-collisional regime. The algebraic growth
stage was observed and compared favorably with theory. Nonlinear island
saturation was found to depend on collisionality.

Dr. Harry Mynick presented a talk "Tutorial on Stellarator Transport-3:
Transport Optimization" at the Theory Department and Transport &
Turbulence SFG Joint Seminar on May 12. The third of three talks in a
seminar series on stellarator transport, the talk discussed the physics
underlying the various approaches to reducing neoclassical and
turbulent transport in stellarators, some of which are being
implemented in the new generation of stellarator experiments like HSX,
NCSX, QPS, and W7-X.

Prof. Herbert Berk from Institute for Fusion Studies visited PPPL from
May 9 to May 13. Part of his visit included collaboration with Drs.
N.N.Gorelenkov and G.Y.Fu on energetic particle driven instabilities.
In particular recent DIII-D experiments of multi Alfven mode excitation
were discussed in which high-n modes with n up to 40 have been
observed.

CPPG Group

Doug McCune visited C-Mod for three days, meeting with Drs. Marco
Brambilla and Roberto Bilato (IPP/Garching) and Drs. Paul Bonoli and
John Wright (MIT) for collaboration on the further development of the
ICRF module, TORIC. A cvs repository was established at MIT containing
a single TORIC source that will build both the serial and the parallel
versions of the code and aid shared development in the future.
PPPL/TRANSP access to the repository was confirmed. Progress was also
made on the installation of the new TORIC in TRANSP; changes in model
behavior relative to the old version were verified to be based on
physics improvements.

An initial version of a TRANSP "data module" was committed into the
TRANSP cvs repository. This is a step for the P-TRANSP project, as it
will enable access to and sharing of data between the traditional
"transport analysis" components of TRANSP, and new predictive
components, without having to couple the predictive components to
TRANSP internal data structures.

S. Jardin and S. Klasky participated in the May 11th panel review of
the proposals submitted for the DOE SciDAC Fusion Simulation Project
call for proposals. They were part of the proposal teams organized by
D. Batchelor in RF/MHD (Jardin) and by C. S. Chang in the Edge Physics
area (Klasky).

Week of 5/6/2004

Theory Group

Drs. Martha Redi and Greg Rewoldt have found NSTX to be near the
critical threshold for kinetic ballooning on a chosen magnetic surface,
when modeling an NSTX plasma with the gyrokinetic code, GS2, using an
equilibrium based on experimental data. The linear microstability of a
low-density NSTX L-Mode pulse, #112996, was studied at the radius
r/a=0.55. This radius was selected as a promising starting point; a
radial survey is planned. At this radius, where the local beta is 9%,
the ITG/TEM modes are unstable, and the KBM was near marginal
stability. A self-consistent beta scan, based on a Miller model MHD
equilibrium, indicates that the critical beta is 11% for strongly
growing kinetic ballooning modes. The toroidal mode numbers and the
real frequency are of the order of those expected for the KBM, with the
growth rates peaking when the mode number is in the range 6 to 9.

Dr. Stefan Gerhardt presented a talk entitled, "Overview of
Measurements and Theory of Flows in Stellarators" as the second of
three talks in the Theory Dept. and Transport & Turbulence SFG Joint
Seminar Series on stellarator transport on May 3. In this talk, he
discussed flows in
stellarators and introduced some of the transport-optimized stellarator
concepts which are now being implemented in new generation experiments
in the US and abroad.

Drs. Russell Kulsrud and Masaaki Yamada presented a talk on the "Study
of Two-Fluid MHD Physics of Magnetic Reconnection in Laboratory and
Space Plasmas" at the Theory department's weekly seminar. They
highlighted the most recent findings of the MRX laboratory experiments,
which address the two-fluids MHD physics of magnetic reconnection. The
results were compared with recent space observations. The experimental
operation regime has moved from the collisional to the collision-free
regime, and two-fluid effects have become more evident. The recent
development from the one-fluid MHD to the two-fluid MHD formulations
was presented to illuminate the physics of the Hall MHD in a
collision-free reconnection layer. A clear experimental verification of
an out-of-plane Hall quadrupole field has been made in a Harris-like
neutral sheet, with the width comparable to the ion skin depth during
magnetic reconnection. High frequency fluctuations observed in the
reconnection layer also exhibit two fluid effects, demonstrating
different kinematics for electrons and ions. The recently developed
theory that investigates the casual relationship between the observed
fast reconnection rate, magnetic turbulence and the Hall quadrupole
fields were discussed in this talk.

Week of 4/29/05

Theory Group

Dr. Anthony Mezzacappa, Group Leader for Theoretical Astrophysics in
the Physics Division of Oak Ridge National Laboratory and Principal
Investigator of DOE's Terascale Supernova Initiative visited PPPL on
Monday, April 25. After presenting a talk on "The Computational
Challenge of Modeling the Explosion of Massive Stars," at the Princeton
Institute for Computational Science and Engineering (PICSciE)
Colloquium, Dr. Mezzacappa came to PPPL to discuss computational
science issues of mutual interest with Drs. W. Lee, S. Klasky, M.
Adams, and W. Tang.

On April 28, Dr. Harry Mynick gave a tutorial talk on the stellarator
transport in a joint seminar between the Theory Department and
Transport & Turbulence SFG. An introductory presentation on stellarator
neoclassical transport was given with a discussion on the various
transport mechanisms, ambipolarity constraint, ion & electron roots,
etc. A recent paper "The effect on stellarator neoclassical transport
of a fluctuating electrostatic spectrum", by Harry Mynick and Allen
Boozer (Columbia) has been accepted for publication in the Physics of
Plasmas.

CPPG Group

Stephane Ethier attended the "Blue Gene Applications Workshop" at
Argonne National Laboratory. During the hands-on session, the GTC code
was ported to Blue Gene and scaling tests were run on up to 1024
processors. Although each processor delivered a performance similar to
the IBM Power3, Blue Gene showed remarkable scalability.

It was announced Wednesday that CPPGs Stephane Ethier is the newly
elected
vice-chair of the NERSC Users Group, where he will work with the new
Chair, David Dean, a Nuclear Physics scientist from ORNL. According to
NERSC Director, Horst Simon: "Both [Stephane and David] are long term
users of NERSC with an excellent understanding of the issues facing
NERSC users."

Week of 4/22/05

Theory Group

Dr. Harry Mynick presented a poster paper at the 2005 Sherwood Theory
Conference in Stateline, NV, in collaboration with Allen Boozer and
Thomas Pedersen of Columbia, entitled "Transport Studies in the
Columbia Non-neutral Torus."

On April 18, Michael Hesse of Goddard Space Flight Center/NASA gave a
theory seminar on "mechanisms of electron demagnetization in
collisionless magnetic reconnection." In this seminar he presented the
mechanisms that foster electron demagnetization in magnetic
reconnection with and without a guide field based on 2.5D and 3D
particle simulations and an analytic theory. From this study, he tried
to determine demagnetization scales for electrons.

On Friday, April 22, Dr. Joshua Breslau and Dr. Leonid E. Zakharov were
the presenters at the monthly Theory Department Micro-seminar. Dr.
Breslau's talk was entitled "Numerical Modeling of the Sawtooth Crash
in a Small Tokamak." He reported on results from a validation study of
the M3D code in which the sawtooth cycle in the CDX-U tokamak was
successfully reproduced using a resistive MHD model. Plans to refine
the model to more accurately reflect the device physics were also
discussed.

Dr. Zakharov presented the theory and technology (based on a new
numerical code Cbc2e) of calibrating tokamaks for purposes of
equilibrium reconstruction. The response functions were introduced for
elimination of the effect of eddy currents (essentially unpredictable)
on readings of magnetic sensors. A set of Equilibrium Reconstruction
Equations (ERE), relying on response functions and the time history of
signals, was presented.

CPPG Group

The TRANSP FusionGrid service is now available on the upgraded PPPL
Linux Cluster. While retaining access to older servers, TRANSP now
shares access to newer machines with factors of 2 to 3 improvement in
speed. TRANSP job requests are sent preferentially to the fast servers
when these are available.

Week of 4/15/05

Theory Group

A number of physicists from the Theory Department contributed to the
2005 Transport Task Force Meeting in Napa, California, and the Sherwood
meeting in Stateline, Nevada.

Daren Stotler gave an invited review talk entitled “Why Should I
Believe My Code - The Quest for Verification and Validation” at the
TTF.  The presentation outlined the efforts to verify and validate the
Monte Carlo neutral transport code DEGAS 2 and contained suggestions
for further discussion.  Two subsequent parallel working group sections
devoted to verification and validation featured many lively exchanges.

On April 11, Ernest Valeo presented a description of “PPPL's Topical
Computational Facility Experiences,” during a special session at the
Sherwood Fusion Theory Conference devoted to Cluster Architecture.

Dr. A. Reiman et al  had a paper on the “Effect of Ambipolar Plasma
Flow on the Penetration of Resonant Magnetic Perturbations in a
Quasi-Axisymmetric Stellarator” accepted for publication in the journal
Nuclear Fusion.  The coauthors are M. Zarnstorff, D. Mikkelsen, L.
Owen(ORNL), H. Mynick, S. Hudson, and D. Monticello.  A reference
equilibrium for the US National Compact Stellarator Experiment is
predicted to be sufficiently close to quasi-symmetry to allow the
plasma to flow in the toroidal direction with little viscous damping,
yet to have sufficiently large deviations from quasi-symmetry that
non-ambipolarity significantly affects the physics of the shielding of
resonant magnetic perturbations by plasma flow.  The unperturbed
velocity profile is modified by the presence of an ambipolar potential,
which produces a broad velocity profile.  In the presence of a resonant
magnetic field perturbation, non-ambipolar transport produces a radial
current, and the resulting jxB force resists departures from the
ambipolar velocity and enhances the shielding.

A joint TTF US-Japan Workshop on Energetic Particle Physics was held in
Napa, California, April 6-9, 2005 (organizers: G.Y. Fu and N.N.
Gorelenkov of PPPL, Y. Todo of NIFS, and B. Breizman of IFS. The
workshop had 21 contributed papers, five from Japan, one from Europe,
and 15 from the US. PPPL attendees included R.V. Budny, G.Y. Fu, N.N.
Gorelenkov, G. Kramer and K.L. Wong.  The presentations and discussions
covered a broad variety of energetic particle physics in
stellarators/helical devices, tokamaks, spherical tokamaks, and burning
plasmas. The participants discussed energetic particle confinement,
Alfven eigenmodes, energetic particle modes, fishbone modes,
axisymmetric energetic-particle driven oscillations, kinetic Alfven
wave, and improvement of burning plasma performance. This workshop also
provided a forum on important issues to be addressed before the burning
plasma experiments.

Daren Stotler gave an invited review talk entitled “Why Should I
Believe My Code - The Quest for Verification and Validation” at the
TTF.  The presentation outlined the efforts to verify and validate the
Monte Carlo neutral transport code DEGAS 2 and contained suggestions
for further discussion.  Two subsequent parallel working group sections
devoted to verification and validation featured many lively exchanges.

Dr. Fu presented simulation results for fishbone instability using the
M3D hybrid code with both MHD nonlinearity and particle nonlinearity.
It was found that flattening of distribution function of energetic ions
causes strong frequency chirping while MHD nonlinearity reduces
saturation level.

Dr. Gorelenkov showed that ITER reversed and normal shear plasmas are
unstable to TAE if NBI ion drive is included. It was also found that
the thermal ion Landau eta_i drive can excite a “high-n sea” of Alfven
modes in DIII-D.

On April 15, Dr. Dirk Van Eester from Laboratory for Plasma Physics,
Association EURATOM, Belgium presented a theory seminar on “a simple
method to account for drift orbit effects when modeling radio frequency
heating in tokamaks.” In this seminar, a semi-analytical method is
proposed to evaluate the dielectric response of a plasma to
electromagnetic waves in the ion cyclotron domain of frequencies
accounting for drift orbit effects. The method relies on subdividing
the orbit into elementary segments in which the integrations can be
performed analytically or by tabulation, and it hinges on the local
bookkeeping of the relation between the variables defining an orbit and
those describing the magnetic geometry.

CPPG Group

Ravi Samtaney visited the Applied Numerical Algorithms Group at LBNL
from Monday, April 11 thru Thursday, April 14 2005 to continue the
ongoing implementation of the AMR MHD code. The ideal MHD AMR code is
now part of the Chombo distribution. On Friday April 15, he visited
CASC, LLNL to continue the development of the Jacobian Free
Newton-Krylov implicit MHD code. These projects are ongoing
collaborations between the CEMM, APDEC and TOPS SciDAC Centers.

Week of 4/8/05

Theory Group

On April 7, Igor Kaganovich presented a theory seminar “Dynamics of Ion
Beam Interaction with Background Plasma” In this seminar, he discussed
the ion beam transport in a background plasma as pertains to heavy ion
inertial fusion. Theory and simulations of the plasma response to the
propagation of an intense ion beam pulse have been reviewed.
Visualization of the electron dynamics reveals the complex nature of
the physical processes. Particular attention was paid to an analysis of
common misconceptions and difficulties encountered in studies of
collective phenomena in ion beam-plasma interactions.

CPPG Group

Triangular high order finite elements have been available for some time
now in the M3D code. However, due to the associated longer running
times, they have not been used in the majority of production runs. In
order to take advantage of the improved accuracy of higher order
elements and at the same time improve the running times, Jin Chen has
implemented “lumped mass” high order elements in M3D. These finite
elements have the Lagrange properties at the nodal points, which now
coincide with the collocation points. Mathematically, they are similar
to spectral Lagrange elements in that the mass matrix M remains
diagonal. In numerical experiments, it has been shown that the 2nd
order lumped mass high order element code has the same or improved
accuracy as the old higher order code (improvement is due to an extra
3rd order term), and also executes nearly as fast as the linear element
version.

Week of 4/1/05

Theory Group

On Friday, April 1, Dr. Harry Mynick gave the Plasma Physics Colloquium
at Columbia University, entitled “The effect of fluctuations on
neoclassical transport in stellarators and tokamaks.” He reported on
work done recently in collaboration with Dr. Allen Boozer of Columbia
University.

On March 31, Dr. John Krommes gave a pedagogical introduction to the
classical Fluctuation--Dissipation Theorem (FDT) in a theory seminar.
He reviewed the intuition behind the FDT, derivation for an
unmagnetized plasma, the role of normal modes, extension to
gyrokinetics, application to the delta-f simulation algorithm, and the
relationship to steady-state turbulence.

CPPG Group

Daren Stotler presented the CPPG seminar on “Why Should I Believe My
Code? The quest for Verification and Validation.” The seminar
addressed, by way of example, some of the techniques that have been
used to verify and validate the DEGAS 2 Monte Carlo neutral transport
code. These include documenting, version control, analytic benchmarks,
convergence tests, cross code benchmarks, and comparison with
experiments.

Josh Breslau has produced a utility making it possible to copy M3D
checkpoint files across platforms so that a run that was begun on one
platform can be completed on another. The utility reads the restart
file (which is in native binary format) converts it to the machine
independent HDF5 format, transfers it to the new machine, and then
converts it into the native binary format for that machine. The source
code for the utility is: /p/m3d/jbreslau/puncture/bin2hdf5.c. It
should be compiled on each system with a local version of the serial
HDF5 library, version 1.6.1 or later. Usage is then: bin2hdf5 -v
checkpoint.r48.

Week of 3/25/05

Theory Group

T.S. Hahm, Wei-li Lee, and Ernie Valeo attended the US-Japan Microwave
Imaging Workshop held at PPPL from March 21 to 22, 2005. At this they each
presented invited talks, Hahm’s talk was titled, "A theorist's perspective on
microturbulence and its measurements", Lee’s title was "Steady state
gyrokinetic particle simulation of microturbulence", and Valeo’s talk was on
the, "Status of development of a 3-D multi-region code for the simulation of
correlation reflectometry in fusion plasmas".

Daren Stotler and Charles Skinner made presentations at the "New Directions
for Computer Simulations and Experiments in Plasma-Surface Interactions for
Fusion" workshop this week in Oak Ridge. The titles of their talks were
"Plasma-Surface Interaction Processes in a Neutral Transport Code - What's
Not in the Model" and "Is Carbon a Realistic Choice for ITER's Divertor?”
respectively. The workshop drew 61 participants, including many
international and ITER-affiliated experts. Highlights included exciting
advances in molecular dynamics codes that revealed plasma surface
interactions in atomic detail and the latest experimental data on plasma
surface interactions issues for ITER. The end result was a clear picture of
the current state of computational approaches to plasma-surface interactions
and a list of suggested future applications of those codes.

A national Stellarator Theory Teleconference was hosted by Allan Reiman on
Thursday, March 24. An overview of stellarator theory research at Columbia
University was presented by Allen Boozer.

CPPG Group

The main internal communications mechanism of the TRANSP code has been
modernized. Using a semi-automated procedure aided by Python script code, the
old "fortran-77 COMMON block INCLUDE file" has been replaced with a set of
fortran-90 modules. The change affected 1283 subroutines in slightly over
1000 source files. The use of the fortran-90 module allows dynamic memory
allocation, with the dual benefit that (a) larger simulations can be run than
was previously possible, and (b) smaller simulations run more efficiently
(measured to be 5% to 12% faster depending on hardware), because arrays of
exactly the right size are allocated; rather than using overly large arrays
in a static build designed to support both small and large problems. A
specific physics modeling benefit is that fast ion distributions can be
calculated with finer radial resolution. The f90 upgrade of TRANSP is
undergoing final tests and will be available for production use (Fusion Grid)
shortly.

In collaboration with GA, tests were successfully completed for an "advanced
reservation" service in the Fusion Collaboratory. An authorized client is
able to reserve a fast machine "for priority use" over a pre-determined time
slot. If the client actually submits jobs in this time window, these receive
priority. Other jobs continue, but at reduced priority. If the time window
expires, the client's job reverts to normal priority. The advanced
reservation system is envisioned as a possible tool for "between shots"
deployment of TRANSP or other Fusion Grid computational services.

Week of 3/18/05

Theory Group

T.S. Hahm, G. Rewoldt, and S. Ethier of PPPL, along with P.H. Diamond
and O. Gurcan of the University of California at San Diego and Z. Lin
of the University of California at Irvine, have submitted a paper
titled “On the dynamics of edge-core coupling” for a “Special Topics”
section of Physics of Plasmas for the Second IAEA Technical Meeting on
the Theory of Plasma Instabilities: Transport, Stability and their
Interaction. One of the nagging, unresolved questions in fusion theory
is concerned with defining the extent of the edge. Gyrokinetic particle
simulations of toroidal ion temperature gradient (ITG) turbulence
spreading using the Gyrokinetic Toroidal Code (GTC) and its related
dynamical model have been extended to a system with a radially varying
ion temperature gradient, in order to study the inward spreading of
edge turbulence toward the core plasma. Due to such spreading, the
turbulence intensity in the core region is significantly enhanced over
the value obtained from simulations of the core region only, as a
consequence the precise boundary of the edge region is blurred. Even
when the core gradient is within the Dimits shift regime (i.e.,
dominated by self-generated zonal flows which reduce the transport to a
negligible value), a significant level of turbulence can penetrate to
the core due to spreading from the edge. The scaling of the turbulent
front propagation speed is closer to the prediction from a nonlinear
diffusion model than from one based on linear toroidal coupling.

In the monthly theory department microseminar on March 18, Gwang-Son
Choe gave a talk entitled “Solar Eruptive Processes Studied by Static
and Dynamic Simulations.” He first presented force-free field solutions
having more magnetic energy than open fields in relation to coronal
mass ejections, and he briefly discussed the energetic possibility of
field line opening. In the second part of the talk, he presented a
theory on the roll effect in solar prominence eruption. He argued that
the anisotropy of electric conductivity tensor in partially ionized
plasma breaks geometrical symmetry and causes the observed rolling
motions of erupting prominences. Jay Johnson described work in the
space physics group modeling auroral phenomena. He described an
iterative model developed to describe ion outflows resulting from ion
cyclotron heating in the topside ionosphere. An ion cyclotron wave
propagation code and Monte Carlo simulation are iterated until they
converge to a stable state. He also described simulations with Hideo
Okuda to understand the relationship between thin ionization layers and
the enhanced aurora.

CPPG Group

There has been progress in the implementation of an MPI-parallelized
version of NUBEAM, the TRANSP fast ion Monte Carlo module. The orbit
following section of the code has been parallelized and scales linearly
in tests so far on the PPPL Linux Cluster and SGI Altix environments.
With the same random number generator initialization, the serial
version and the parallel version of the same calculation produce the
same results to within machine round off error.

Week of 3/11/05

Theory Group

Drs. T.S. Hahm, C.Z. Cheng, and G. Rewoldt attended the 2nd IAEA
Technical meeting on the Theory of Plasma Instabilities: Transport,
Stability, and their Interaction, held at the International Center for
Theoretical Physics in Trieste, Italy from March 2 to 4, 2005.  At this
workshop, Hahm presented an invited talk, “Theory, Simulation, and
Experimental Test of Turbulence Spreading,” Cheng presented an invited
talk “Energetic Particle Physics in Tokomak Burning Plasmas,” and
Rewoldt presented an invited talk “Recent Progress in Gyrokinetic
Particle Simulations of Turbulent Plasmas”.  Hahm also served as
Chairman of the International Advisory Committee and presented a
summary talk on turbulence.

Dr. Vladmir Yankov of VyOptics presented a theory seminar entitled
“Improvement of Confinement in Tokamaks by Weakening of Poloidal
Magnetic Field Near Boundary” on March 10. Theory of turbulent
equipartition and experiment indicate that density, pressure, and
temperature profiles follow the poloidal magnetic field profile. The
TFTR current ramp-down experiments were cited as an example. Based on
this model he proposed effecting a change in the magnetic geometry
between the core and the boundary by toroidal conductors and/or the
plasma current. As a result density and temperature gradients would
become steeper, and stored energy will be higher with low boundary
plasma parameters. He suggested that this new mode of confinement might
simplify the task of achieving ignition. Applications to stellarators
were also discussed.

Dr. Wei-Li Lee gave a seminar at the Department of Applied Physics and
Applied Mathematics at Columbia University entitled “Steady State
Simulation of Microturbulence.” The talk focused 1) on the important
role played by the velocity-space nonlinearity in achieving the
turbulent steady state, and also 2) on the validity of using the
fluctuation-dissipation theorem (FDT) for studying numerical noise in
the turbulent steady state. The velocity space nonlinearity, which
provides another channel of nonlinear interaction in the simulation,
has been routinely ignored in most of the recent gyrokinetic
simulations. As for the noise issue, since FDT is based on the linear
dispersion relation of a quiescent system with damped modes, one cannot
arbitrarily extend FDT's regime of validity to a driven system with
growing modes. This is based simply on the fact that the contour
integration of 1/\omega Im D in the complex frequency (\omega) plane