The DOE ACTS Collection Interfaces, Functionality and Interoperability New Methods for Developing Pe

1
The DOE ACTS CollectionInterfaces,
Functionality and InteroperabilityNew Methods
 for Developing Peta-Scalable CodesPittsburgh
Supercomputing Center, 2004

• Tony Drummond
• Lawrence Berkeley National Laboratory
• acts-support_at_nersc.gov
• LADrummond_at_lbl.gov

2
Motivation- Why do we need software libraries?
Large Scientific Codes A Common Programming
Practice
Algorithmic Implementations
I/O
Application Data Layout
Control
Tuned and machine Dependent modules
3
Motivation- Why do we need software libraries?
Algorithmic Implementations
I/O
Application Data Layout
Control
Tuned and machine Dependent modules
4
Motivation- Why do we need software libraries?
An Alternative Approach
USER's APPLICATION CODE (Main Control)
AVAILABLE
AVAILABLE
Algorithmic Implementations
AVAILABLE
I/O
Application Data Layout
LIBRARIES
LIBRARIES PACKAGES
LIBRARIES PACKAGES
Tuned and machine Dependent modules
5
Shortcomings?
"We need to move away from a coding style suited
for serial machines, where every macrostep of an
algorithm needs to be thought about and
explicitly coded, to a higher-level style, where
the compiler and library tools take care of the
details. And the remarkable thing is, if we
adopt this higher-level approach right now, even
on today's machines, we will see immediate
benefits in our productivity." W. H. Press and
S. A. Teukolsky, 1997 Numerical Recipes Does
This Paradigm Have a future?
6
Software Interfaces
CALL BLACS_GET( -1, 0, ICTXT ) CALL
BLACS_GRIDINIT( ICTXT, 'Row-major', NPROW, NPCOL
) CALL BLACS_GRIDINFO( ICTXT, NPROW, NPCOL,
MYROW, MYCOL ) CALL PDGESV( N, NRHS, A, IA,
JA, DESCA, IPIV, B, IB, JB, DESCB,
INFO )
Language Calls
Command lines
Problem Domain
7
ACTS Numerical Tools Functionality
8
ACTS Numerical Tools Functionality
9
Structure of PETSc
10
PETSc Example - Basic Linear Solver in C
SLES sles / linear solver
context / Mat A /
matrix / Vec x, b /
solution, RHS vectors / int n, its
/ problem dimension, number of
iterations / MatCreate(MPI_COMM_WORLD,PETSC_DEC
IDE,PETSC_DECIDE, n,n,A) / assemble
matrix / VecCreate(MPI_COMM_WORLD,PETSC_DECIDE,n,
x) VecDuplicate(x,b)
/ assemble RHS vector
/ SLESCreate(MPI_COMM_WORLD,sles)
SLESSetOperators(sles,A,A,DIFFERENT_NONZERO_PATTE
RN) SLESSetFromOptions(sles) SLESSolve(sles,b,x,
its) SLESDestroy(sles)
11
Hypre Conceptual Interfaces
12
Hypre Conceptual Interfaces
INTERFACE TO SOLVERS
List of Solvers and Preconditioners per
Conceptual Interface
13
ACTS Numerical Tools Functionality
14
ACTS Numerical Tools Functionality
15
ACTS Numerical Tools Functionality
16
ACTS Numerical Tools Functionality
17
ACTS Numerical Tools Functionality
18
TAO - Interface with PETSc
19
OPT Interfaces

• Four major classes of problems available
• NLF0(ndim, fcn, init_fcn, constraint)
• Basic nonlinear function, no derivative
  information available
• NLF1(ndim, fcn, init_fcn, constraint)
• Nonlinear function, first derivative information
  available
• FDNLF1(ndim, fcn, init_fcn, constraint)
• Nonlinear function, first derivative information
  approximated
• NLF2(ndim, fcn, init_fcn, constraint)
• Nonlinear function, first and second derivative
  information available

20
ACTS Numerical Tools Functionality
21
ACTS Numerical Tools Functionality
22
ACTS Numerical Tools Functionality
23
ACTS Numerical Tools Functionality
24

• Automatic Tuning
• PHiPAC www.icsi.berkeley.edu/bilmes/phipac
• ATLAS www.netlib.org/atlas
• XBLAS www.nersc.gov/xiaoye/XBLAS
• Sparsity www.cs.berkeley.edu/yelick/sparsity
• FFTW www.fftw.org

25
ACTS Numerical Tools Functionality
26
Tool Interoperability Tool-to-Tool
TOOL B
TOOL C
TOOL A
TOOL F
TOOL E
TOOL D
27
Component Technology!
Tool C
Tool A
Tool D
Tool B
ESI
CCA
28
PSE's and Frameworks
Ax b
PMatlab
View_field(T1)
User
PyACTS
29
PyACTS
30
ACTS Collection Tutorials
http//acts.nersc.gov/events/Workshop2003/
SIAM PP04 Short Course Workshop 2003
CD Request acts-support_at_nersc.gov - Its FREE!
ACTS Workshop 2004 http//acts.nersc.gov/events/Wo
rkshop2004