Developer Reference for Intel® oneAPI Math Kernel Library - C

Developer Reference

  • 0.9
  • 09/09/2020
  • Public Content
Contents

Extended Eigensolver Input Parameters

The input parameters for Extended Eigensolver routines are contained in an
MKL_INT
 array named
fpm
. To call the Extended Eigensolver interfaces, this array should be initialized using the routine
feastinit
.
Parameter
Default
Description
fpm
[0]
0
Specifies whether Extended Eigensolver routines print runtime status.
fpm
[0]
=0
Extended Eigensolver routines do not generate runtime messages at all.
fpm
[0]
=1
Extended Eigensolver routines print runtime status to the screen.
fpm
[1]
8
The number of contour points
N
e
 = 8
 (see the description of FEAST algorithm). Must be one of {3,4,5,6,8,10,12,16,20,24,32,40,48}.
fpm
[2]
12
Error trace double precision stopping criteria
ε
 (
ε
 = 10
-
fpm
[2]
) .
fpm
[3]
20
Maximum number of Extended Eigensolver refinement loops allowed. If no convergence is reached within
fpm
[3]
 refinement loops, Extended Eigensolver routines return
info
=2.
fpm
[4]
0
User initial subspace. If
fpm
[4]
=0 then Extended Eigensolver routines generate initial subspace, if
fpm
[4]
=1 the user supplied initial subspace is used.
fpm
[5]
0
Extended Eigensolver stopping test.
fpm
[5]
=0
Extended Eigensolvers are stopped if this residual stopping test is satisfied:
  • generalized eigenvalue problem:
  • standard eigenvalue problem:
where
mode
 is the total number of eigenvalues found in the search interval and
ε
 = 10
-
fpm
[6]
 for real and complex or
ε
 = 10
-
fpm
[2]
 for double precision and double complex.
fpm
[5]
=1
Extended Eigensolvers are stopped if this trace stopping test is satisfied:
,
where
trace
j
 denotes the sum of all eigenvalues found in the search interval [
emin
,
emax
] at the
j
-th Extended Eigensolver iteration:
.
fpm
[6]
5
Error trace single precision stopping criteria (10
-
fpm
[6]
) .
fpm
[13]
0
fpm
[13]
=0
Standard use for Extended Eigensolver routines.
fpm
[13]
=1
Non-standard use for Extended Eigensolver routines: return the computed eigenvectors subspace after one single contour integration.
fpm
[26]
0
Specifies whether Extended Eigensolver routines check input matrices (applies to CSR format only).
fpm
[26]
=0
Extended Eigensolver routines do not check input matrices.
fpm
[26]
=1
Extended Eigensolver routines check input matrices.
fpm
[27]
0
Check if matrix
B
 is positive definite. Set
fpm
[27]
= 1
 to check if
B
 is positive definite.
fpm
[29]
 to
fpm
[62]
-
Reserved for future use.
fpm
[63]
0
Use the
Intel® oneAPI Math Kernel Library
 PARDISO solver with the user-defined PARDISO
iparm
 array settings.
This option can only be used by Extended Eigensolver Predefined Interfaces for Sparse Matrices.
fpm
[63]
=0
Extended Eigensolver routines use the
Intel® oneAPI Math Kernel Library
 PARDISO default
iparm
 settings defined by calling the
pardisoinit
 subroutine.
fpm
[63]
=1
The values from
fpm
[64]
 to
fpm
[127]
 correspond to
iparm
[0]
 to
iparm
[63]
 respectively according to the formula
fpm
[64 +
i
]
=
iparm
[
i
]
 for
i
 =
0, 1, ..., 63
.

Product and Performance Information

1

Intel's compilers may or may not optimize to the same degree for non-Intel microprocessors for optimizations that are not unique to Intel microprocessors. These optimizations include SSE2, SSE3, and SSSE3 instruction sets and other optimizations. Intel does not guarantee the availability, functionality, or effectiveness of any optimization on microprocessors not manufactured by Intel. Microprocessor-dependent optimizations in this product are intended for use with Intel microprocessors. Certain optimizations not specific to Intel microarchitecture are reserved for Intel microprocessors. Please refer to the applicable product User and Reference Guides for more information regarding the specific instruction sets covered by this notice.

Notice revision #20110804