CP_DBCSR
Section can be repeated.
Benchmark and test the cp_dbcsr routines [Edit on GitHub]
Keywords
Keyword descriptions
- ALPHA: real = 1.00000000E+000
Usage: ALPHA 2.0
Multiplication factor [Edit on GitHub]
- ALWAYS_CHECKSUM: logical = F
Lone keyword:
T
Usage: ALWAYS_CHECKSUM
perform a checksum after each multiplication [Edit on GitHub]
- ASPARSITY: real = 0.00000000E+000
Usage: ASPARSITY 70
Sparsity of A matrix [Edit on GitHub]
- ATYPE: string = N
Usage: ATYPE N
Matrix A type [Edit on GitHub]
- BETA: real = 0.00000000E+000
Usage: BETA 1.0
Product premultiplication factor [Edit on GitHub]
- BSPARSITY: real = 0.00000000E+000
Usage: ASPARSITY 80
Sparsity of B matrix [Edit on GitHub]
- BS_K: integer = 1 13 2 15
Usage: BS_K 1 13 2 5
Block sizes of inner dimension [Edit on GitHub]
- BS_M: integer = 1 13 2 15
Usage: BS_M 1 13 2 5
Row block sizes of C [Edit on GitHub]
- BS_N: integer = 1 13 2 15
Usage: BS_N 1 13 2 5
Column block sizes of C [Edit on GitHub]
- BTYPE: string = N
Usage: BTYPE N
Matrix B type [Edit on GitHub]
- CSPARSITY: real = 0.00000000E+000
Usage: ASPARSITY 90
Sparsity of C matrix [Edit on GitHub]
- CTYPE: string = N
Usage: CTYPE N
Matrix C type [Edit on GitHub]
- DATA_TYPE: enum = REAL_8
Usage: DATA_TYPE real_8
Valid values:
REAL_8
Real (Double Precision)COMPLEX_8
Complex (Double Precision)
Data type of the matrices [Edit on GitHub]
- FILTER_EPS: real = -1.00000000E+000
Usage: FILTER_EPS 1.0
Threshold for on-the-fly and final filtering. [Edit on GitHub]
- K: integer = 256
Usage: A 1024
Inner dimension M [Edit on GitHub]
- KEEPSPARSE: logical = F
Lone keyword:
T
Usage: KEEPSPARSE
Keep product sparse [Edit on GitHub]
- M: integer = 256
Usage: A 1024
Dimension 1 of C [Edit on GitHub]
- N: integer = 256
Usage: A 1024
Dimension 2 of C [Edit on GitHub]
- NPROC: integer = 0
Usage: NPROC 128 16 1
Number of processors to test [Edit on GitHub]
- N_LOOP: integer = 10
Usage: N_LOOP 10
Number of operations being timed (useful for small matrices). [Edit on GitHub]
- TEST_TYPE: enum = MM
Usage: TEST_TYPE MM
Valid values:
MM
Run matrix multiplicationsBINARY_IO
Run binary IO checks
Which part of DBCSR is tested [Edit on GitHub]
- TRANSA: logical = F
Lone keyword:
T
Usage: TRANSA
Transpose matrix A [Edit on GitHub]
- TRANSB: logical = F
Lone keyword:
T
Usage: TRANSB
Transpose matrix B [Edit on GitHub]