|
#define | TILE_SIZE 512 |
|
#define | NUM_CORES 48 |
|
#define | MAX(a, b) ( ( ( a ) > ( b ) ) ? ( a ) : ( b ) ) |
|
#define | MIN(a, b) ( ( ( a ) < ( b ) ) ? ( a ) : ( b ) ) |
|
#define | FMULS_GEMM(m_, n_, k_) ( (m_) * (n_) * (k_) ) |
|
#define | FADDS_GEMM(m_, n_, k_) ( (m_) * (n_) * (k_) ) |
|
#define | FLOPS_DGEMM(m_, n_, k_) |
|
#define | FMULS_SYMM(side_, m_, n_) ( ( (side_) == Left ) ? FMULS_GEMM((m_), (m_), (n_)) : FMULS_GEMM((m_), (n_), (n_)) ) |
|
#define | FADDS_SYMM(side_, m_, n_) ( ( (side_) == Left ) ? FADDS_GEMM((m_), (m_), (n_)) : FADDS_GEMM((m_), (n_), (n_)) ) |
|
#define | FLOPS_DSYMM(side_, m_, n_) ( FMULS_SYMM(side_, (double)(m_), (double)(n_)) + FADDS_SYMM(side_, (double)(m_), (double)(n_)) ) |
|
#define | FMULS_TRSM_2(m_, n_) ( 0.5 * (n_) * (m_) * ( (m_) + 1 ) ) |
|
#define | FADDS_TRSM_2(m_, n_) ( 0.5 * (n_) * (m_) * ( (m_) - 1 ) ) |
|
#define | FMULS_TRSM(side_, m_, n_) ( ( (side_) == Left ) ? FMULS_TRSM_2((m_), (n_)) : FMULS_TRSM_2((n_), (m_)) ) |
|
#define | FADDS_TRSM(side_, m_, n_) ( ( (side_) == Left ) ? FADDS_TRSM_2((m_), (n_)) : FADDS_TRSM_2((n_), (m_)) ) |
|
#define | FLOPS_DTRSM(side_, m_, n_) ( FMULS_TRSM(side_, (double)(m_), (double)(n_)) + FADDS_TRSM(side_, (double)(m_), (double)(n_)) ) |
|
#define | FMULS_TRMM_2(m_, n_) ( 0.5 * (n_) * (m_) * ( (m_) + 1 ) ) |
|
#define | FADDS_TRMM_2(m_, n_) ( 0.5 * (n_) * (m_) * ( (m_) - 1 ) ) |
|
#define | FMULS_TRMM(side_, m_, n_) ( ( (side_) == Left ) ? FMULS_TRMM_2((m_), (n_)) : FMULS_TRMM_2((n_), (m_)) ) |
|
#define | FADDS_TRMM(side_, m_, n_) ( ( (side_) == Left ) ? FADDS_TRMM_2((m_), (n_)) : FADDS_TRMM_2((n_), (m_)) ) |
|
#define | FLOPS_DTRMM(side_, m_, n_) (FMULS_TRMM(side_, (double)(m_), (double)(n_)) + FADDS_TRMM(side_, (double)(m_), (double)(n_)) ) |
|
#define | FMULS_SYRK(k_, n_) ( 0.5 * (k_) * (n_) * ( (n_) + 1 ) ) |
|
#define | FADDS_SYRK(k_, n_) ( 0.5 * (k_) * (n_) * ( (n_)+1 ) ) |
|
#define | FLOPS_DSYRK(k_, n_) ( FMULS_SYRK( (double)(k_), (double)(n_) ) + FADDS_SYRK( (double)(k_), (double)(n_) ) ) |
|
#define | FMULS_SYR2K(k_, n_) ((k_) * (n_) * (n_)) |
|
#define | FADDS_SYR2K(k_, n_) ((k_) * (n_) * (n_) + (n_)) |
|
#define | FLOPS_DSYR2K(k_, n_) (FMULS_SYR2K((double)(k_), (double)(n_)) + FADDS_SYR2K((double)(k_), (double)(n_))) |
|
#define | FMULS_POTRF(n_) |
|
#define | FADDS_POTRF(n_) ( (1./6.) * (n_) * (n_) * (n_) - (1./6.) * (n_)) |
|
#define | FLOPS_DPOTRF(n_) ( FMULS_POTRF((double)(n_)) + FADDS_POTRF((double)(n_))) |
|
#define | FMULS_GETRF(m_, n_) ( ( (m_) >= (n_) ) ? 0.5 * (m_) * (n_) * (n_) - 1./6. * (n_) * (n_) * (n_) + 0.5 * (m_) * (n_) - 0.5 * (n_) * (n_) + 2./3. * (n_) : 0.5 * (n_) * (m_) * (m_) - 1./6. * (m_) * (m_) * (m_) + 0.5 * (n_) * (m_) - 0.5 * (m_) * (m_) + 2./3. * (m_) ) |
|
#define | FADDS_GETRF(m_, n_) ( ( (m_) >= (n_) ) ? 0.5 * (m_) * (n_) * (n_) - 1./6. * (n_) * (n_) * (n_) - 0.5 * (m_) * (n_) + 1./6. * (n_) : 0.5 * (n_) * (m_) * (m_) - 1./6. * (m_) * (m_) * (m_) - 0.5 * (n_) * (m_) + 1./6. * (m_) ) |
|
#define | FLOPS_DGETRF(m_, n_) ( FMULS_GETRF((double)(m_), (double)(n_)) + FADDS_GETRF((double)(m_), (double)(n_)) ) |
|
#define | DSSS_MACROS_H |
|
#define | FLOPS_DGTSV(n_) ( 8 * (n_) ) |
|
#define | FLOPS_DSPMV(nnz_) ( 2 * (nnz_) ) |
|
Macros definition.
LASs is a software package provided by: Barcelona Supercomputing Center - Centro Nacional de Supercomputacion
- Author
- Pedro Valero-Lara pedro.nosp@m..val.nosp@m.ero@b.nosp@m.sc.e.nosp@m.s
- Date
- 2017-01-02
Definition in file lass_macros.h.