PLASMA  2.8.0
PLASMA - Parallel Linear Algebra for Scalable Multi-core Architectures
int CORE_cparfb ( PLASMA_enum  side,
PLASMA_enum  trans,
PLASMA_enum  direct,
PLASMA_enum  storev,
int  M1,
int  N1,
int  M2,
int  N2,
int  K,
int  L,
PLASMA_Complex32_t *  A1,
int  LDA1,
PLASMA_Complex32_t *  A2,
int  LDA2,
const PLASMA_Complex32_t *  V,
int  LDV,
const PLASMA_Complex32_t *  T,
int  LDT,
PLASMA_Complex32_t *  WORK,
int  LDWORK 
)

CORE_cparfb applies a complex upper triangular block reflector H or its transpose H' to a complex rectangular matrix formed by coupling two tiles A1 and A2. Matrix V is: 

    COLUMNWISE                    ROWWISE

   |     K     |                 |      N2-L     |   L  |
__ _____________ __           __ _________________        __
   |    |      |                 |               | \
   |    |      |                 |               |   \    L

M2-L | | | K |_______________|_____\ __ | | | M2 | | __ |____| | | | K-L \ | | __ |______________________| __ L \ | | __ |______| __ | N2 |

| L | K-L |

Parameters
[in]side
  • PlasmaLeft : apply Q or Q**H from the Left;
  • PlasmaRight : apply Q or Q**H from the Right.
[in]trans
  • PlasmaNoTrans : No transpose, apply Q;
  • PlasmaConjTrans : ConjTranspose, apply Q**H.
[in]directIndicates how H is formed from a product of elementary reflectors
  • PlasmaForward : H = H(1) H(2) . . . H(k) (Forward)
  • PlasmaBackward : H = H(k) . . . H(2) H(1) (Backward)
[in]storevIndicates how the vectors which define the elementary reflectors are stored:
  • PlasmaColumnwise
  • PlasmaRowwise
[in]M1The number of columns of the tile A1. M1 >= 0.
[in]N1The number of rows of the tile A1. N1 >= 0.
[in]M2The number of columns of the tile A2. M2 >= 0.
[in]N2The number of rows of the tile A2. N2 >= 0.
[in]KThe order of the matrix T (= the number of elementary reflectors whose product defines the block reflector).
[in]LThe size of the triangular part of V
[in,out]A1On entry, the M1-by-N1 tile A1. On exit, A1 is overwritten by the application of Q.
[in]LDA1The leading dimension of the array A1. LDA1 >= max(1,N1).
[in,out]A2On entry, the M2-by-N2 tile A2. On exit, A2 is overwritten by the application of Q.
[in]LDA2The leading dimension of the tile A2. LDA2 >= max(1,N2).
[in]V(LDV,K) if STOREV = 'C' (LDV,M2) if STOREV = 'R' and SIDE = 'L' (LDV,N2) if STOREV = 'R' and SIDE = 'R' Matrix V.
[in]LDVThe leading dimension of the array V. If STOREV = 'C' and SIDE = 'L', LDV >= max(1,M2); if STOREV = 'C' and SIDE = 'R', LDV >= max(1,N2); if STOREV = 'R', LDV >= K.
[out]TThe triangular K-by-K matrix T in the representation of the block reflector. T is upper triangular by block (economic storage); The rest of the array is not referenced.
[in]LDTThe leading dimension of the array T. LDT >= K.
[in,out]WORK
[in]LDWORKThe dimension of the array WORK.
Returns
Return values
PLASMA_SUCCESSsuccessful exit
<0if -i, the i-th argument had an illegal value