Parallel version solver
based on iterative solution method
Super
Matrix Solver P-ICCG
Special features
of Super Matrix Solver P-ICCG
- Proven performance applicable to
diverse fields of analysis such as electromagnetic and structural
analyses.
- Wide range of usage from 1 CPU, 2CPU's to full-scale parallel processing.
- Readily installed with easy-to-understand product manuals.
- No need for special mathematical knowledge-easy integration of modules
provided in executable format (such as DLL).
- Supports various types of platforms.
- Complete with support system for evaluation and use.
What is ICCG
Method?
- ICCG is an iterative solution method
based on CG (Conjugate Gradient) method. In ICCG, the calculation
speed of CG method is enhanced with pre-processing technology (Incomplete
Cholesky Factorization). Compared with CG method that has no pre-processing,
ICCG method is faster and more stable.
- ICCG is an iterative method with many actual performance results in diverse
analysis fields such as structural, electromagnetic and computational fluid
dynamic analyses.
P-ICCG Summary
Specifications
Items
Descriptions
Notes
Target Analysis Fields
Structural Analyses and Electromagnetic, etc.
Target Coefficient Matrix
Sparse matrices that are generated from discretization methods
such as finite element, finite volume and differential methods.
Problems with Zero Diagonal Elements
Capable of calculating.
Not all problems with zero diagonal elements
can be calculated.
Types of Unknowns
Real (double) and Complex Numbers
Symmetry of Problems
Symmetric problems only. Unable to calculate asymmetric problems.
Parallelization Method
Supports shared memory type (SMP)
Maximum Number of CPU's
Technically unlimited; however, 1 to 8 CPU's are recommended.
Users must purchase the same number of licenses
as the number of CPU's used.
Input Data
Coefficient matrix, right-hand side vector, target convergence,
maximum number of iteration, etc.
The number of CPU's to be used in calculation
may be assigned.
Output Data
Solution vector, achieved relative residual, actual number
of iteration, etc.
Indication of Error Messages
Warnings and error messages returned as return value (calculation
information, system information, etc).
Method of Provision
DLL format for Windows; Static library format for Linux and
UNIX.
Source code will not be disclosed.
Attached Materials
Product manual (with explanations about data format, parameters,
integration procedures, etc.), sample data, sample program
for integrating SMS-AMG (C and FORTRAN).
P-ICCG's Parallel
Calculation Performance
Actual Calculation Time (sec. )
Types
of Problems
Number
of Unknowns
Convergence
Time (sec.); target convergence: norm＜1.0e-10 )
1-CPU
2-CPU
4-CPU
8-CPU
Magnetic field analysis
220K
87.0
43.6
26.7
13.4
Fluid analysis
500K
78.3
54.2
32.7
19.3
Structural analysis
250K
1227.5
744.7
390.4
224.1
Fluid analysis
1000K
178.5
141.7
82.7
48.5
Comparison of Calculation Performance
(1 CPU calculation speed as 1)
Types of
Problems
Number of
Unknowns
Convergence
Time (sec.); target convergence: norm＜1.0e-10 )
NOTE: Materials provided on this web site does not guarantee functions
and performance of the product.
Product specifications may change without notice.
