MAGMA  1.6.1
Matrix Algebra for GPU and Multicore Architectures
 All Classes Files Functions Friends Groups Pages

Functions

magma_int_t magma_zbajac_csr (magma_int_t localiters, magma_z_sparse_matrix D, magma_z_sparse_matrix R, magma_z_vector b, magma_z_vector *x, magma_queue_t queue)
 This routine is a block-asynchronous Jacobi iteration performing s local Jacobi-updates within the block. More...
 
magma_int_t magma_zbcsrblockinfo5 (magma_int_t lustep, magma_int_t num_blocks, magma_int_t c_blocks, magma_int_t size_b, magma_index_t *blockinfo, magmaDoubleComplex_ptr val, magmaDoubleComplex_ptr *AII, magma_queue_t queue)
 For a Block-CSR ILU factorization, this routine copies the filled blocks from the original matrix A and initializes the blocks that will later be filled in the factorization process with zeros. More...
 
magma_int_t magma_zbcsrvalcpy (magma_int_t size_b, magma_int_t num_blocks, magma_int_t num_zblocks, magmaDoubleComplex_ptr *Aval, magmaDoubleComplex_ptr *Bval, magmaDoubleComplex_ptr *Bval2, magma_queue_t queue)
 For a Block-CSR ILU factorization, this routine copies the filled blocks from the original matrix A and initializes the blocks that will later be filled in the factorization process with zeros. More...
 
magma_int_t magma_zbcsrluegemm (magma_int_t size_b, magma_int_t num_brows, magma_int_t kblocks, magmaDoubleComplex_ptr *dA, magmaDoubleComplex_ptr *dB, magmaDoubleComplex_ptr *dC, magma_queue_t queue)
 For a Block-CSR ILU factorization, this routine updates all blocks in the trailing matrix. More...
 
magma_int_t magma_zbcsrlupivloc (magma_int_t size_b, magma_int_t kblocks, magmaDoubleComplex_ptr *dA, magmaInt_ptr ipiv, magma_queue_t queue)
 For a Block-CSR ILU factorization, this routine updates all blocks in the trailing matrix. More...
 
magma_int_t magma_zbcsrswp (magma_int_t r_blocks, magma_int_t size_b, magmaInt_ptr ipiv, magmaDoubleComplex_ptr x, magma_queue_t queue)
 For a Block-CSR ILU factorization, this routine swaps rows in the vector *x according to the pivoting in *ipiv. More...
 
magma_int_t magma_zbcsrtrsv (magma_uplo_t uplo, magma_int_t r_blocks, magma_int_t c_blocks, magma_int_t size_b, magmaDoubleComplex_ptr A, magma_index_t *blockinfo, magmaDoubleComplex_ptr x, magma_queue_t queue)
 For a Block-CSR ILU factorization, this routine performs the triangular solves. More...
 
void magma_zcompact (magma_int_t m, magma_int_t n, magmaDoubleComplex_ptr dA, magma_int_t ldda, magmaDouble_ptr dnorms, double tol, magmaInt_ptr active, magmaInt_ptr cBlock, magma_queue_t queue)
 ZCOMPACT takes a set of n vectors of size m (in dA) and their norms and compacts them into the cBlock size<=n vectors that have norms > tol. More...
 
magma_int_t magma_zjacobisetup_vector_gpu (int num_rows, magma_z_vector b, magma_z_vector d, magma_z_vector c, magma_z_vector *x, magma_queue_t queue)
 Prepares the Jacobi Iteration according to x^(k+1) = D^(-1) * b - D^(-1) * (L+U) * x^k x^(k+1) = c - M * x^k. More...
 
magma_int_t magma_zlobpcg_maxpy (magma_int_t num_rows, magma_int_t num_vecs, magmaDoubleComplex_ptr X, magmaDoubleComplex_ptr Y, magma_queue_t queue)
 This routine computes a axpy for a mxn matrix: More...
 
int magma_zbicgmerge1 (int n, magmaDoubleComplex_ptr skp, magmaDoubleComplex_ptr v, magmaDoubleComplex_ptr r, magmaDoubleComplex_ptr p)
 Mergels multiple operations into one kernel: More...
 
int magma_zbicgmerge2 (int n, magmaDoubleComplex_ptr skp, magmaDoubleComplex_ptr r, magmaDoubleComplex_ptr v, magmaDoubleComplex_ptr s)
 Mergels multiple operations into one kernel: More...
 
int magma_zbicgmerge3 (int n, magmaDoubleComplex_ptr skp, magmaDoubleComplex_ptr p, magmaDoubleComplex_ptr s, magmaDoubleComplex_ptr t, magmaDoubleComplex_ptr x, magmaDoubleComplex_ptr r)
 Mergels multiple operations into one kernel: More...
 
int magma_zbicgmerge4 (int type, magmaDoubleComplex_ptr skp)
 Performs some parameter operations for the BiCGSTAB with scalars on GPU. More...
 
magma_int_t magma_zbicgmerge_spmv1 (magma_z_sparse_matrix A, magmaDoubleComplex_ptr d1, magmaDoubleComplex_ptr d2, magmaDoubleComplex_ptr dp, magmaDoubleComplex_ptr dr, magmaDoubleComplex_ptr dv, magmaDoubleComplex_ptr skp, magma_queue_t queue)
 Merges the first SpmV using CSR with the dot product and the computation of alpha. More...
 
magma_int_t magma_zbicgmerge_spmv2 (magma_z_sparse_matrix A, magmaDoubleComplex_ptr d1, magmaDoubleComplex_ptr d2, magmaDoubleComplex_ptr ds, magmaDoubleComplex_ptr dt, magmaDoubleComplex_ptr skp, magma_queue_t queue)
 Merges the second SpmV using CSR with the dot product and the computation of omega. More...
 
magma_int_t magma_zbicgmerge_xrbeta (int n, magmaDoubleComplex_ptr d1, magmaDoubleComplex_ptr d2, magmaDoubleComplex_ptr rr, magmaDoubleComplex_ptr r, magmaDoubleComplex_ptr p, magmaDoubleComplex_ptr s, magmaDoubleComplex_ptr t, magmaDoubleComplex_ptr x, magmaDoubleComplex_ptr skp, magma_queue_t queue)
 Merges the second SpmV using CSR with the dot product and the computation of omega. More...
 
magma_int_t magma_zcgmerge_spmv1 (magma_z_sparse_matrix A, magmaDoubleComplex_ptr d1, magmaDoubleComplex_ptr d2, magmaDoubleComplex_ptr dd, magmaDoubleComplex_ptr dz, magmaDoubleComplex_ptr skp, magma_queue_t queue)
 Merges the first SpmV using different formats with the dot product and the computation of rho. More...
 

Detailed Description

Function Documentation

magma_int_t magma_zbajac_csr ( magma_int_t  localiters,
magma_z_sparse_matrix  D,
magma_z_sparse_matrix  R,
magma_z_vector  b,
magma_z_vector *  x,
magma_queue_t  queue 
)

This routine is a block-asynchronous Jacobi iteration performing s local Jacobi-updates within the block.

Input format is two CSR matrices, one containing the diagonal blocks, one containing the rest.

Parameters
[in]localitersmagma_int_t number of local Jacobi-like updates
[in]Dmagma_z_sparse_matrix input matrix with diagonal blocks
[in]Rmagma_z_sparse_matrix input matrix with non-diagonal parts
[in]bmagma_z_vector RHS
[in]xmagma_z_vector* iterate/solution
[in]queuemagma_queue_t Queue to execute in.
magma_int_t magma_zbcsrblockinfo5 ( magma_int_t  lustep,
magma_int_t  num_blocks,
magma_int_t  c_blocks,
magma_int_t  size_b,
magma_index_t *  blockinfo,
magmaDoubleComplex_ptr  val,
magmaDoubleComplex_ptr *  AII,
magma_queue_t  queue 
)

For a Block-CSR ILU factorization, this routine copies the filled blocks from the original matrix A and initializes the blocks that will later be filled in the factorization process with zeros.

Parameters
[in]lustepmagma_int_t lustep
[in]num_blocksmagma_int_t number of nonzero blocks
[in]c_blocksmagma_int_t number of column-blocks
[in]size_bmagma_int_t blocksize
[in]blockinfomagma_int_t* block filled? location?
[in]valmagmaDoubleComplex* pointers to the nonzero blocks in A
[in]AIImagmaDoubleComplex** pointers to the respective nonzero blocks in B
[in]queuemagma_queue_t Queue to execute in.
magma_int_t magma_zbcsrluegemm ( magma_int_t  size_b,
magma_int_t  num_brows,
magma_int_t  kblocks,
magmaDoubleComplex_ptr *  dA,
magmaDoubleComplex_ptr *  dB,
magmaDoubleComplex_ptr *  dC,
magma_queue_t  queue 
)

For a Block-CSR ILU factorization, this routine updates all blocks in the trailing matrix.

Parameters
[in]size_bmagma_int_t blocksize in BCSR
[in]num_browsmagma_int_t number of block rows
[in]kblocksmagma_int_t number of blocks in row
[in]dAmagmaDoubleComplex** input blocks of matrix A
[in]dBmagmaDoubleComplex** input blocks of matrix B
[in]dCmagmaDoubleComplex** output blocks of matrix C
[in]queuemagma_queue_t Queue to execute in.
magma_int_t magma_zbcsrlupivloc ( magma_int_t  size_b,
magma_int_t  kblocks,
magmaDoubleComplex_ptr *  dA,
magmaInt_ptr  ipiv,
magma_queue_t  queue 
)

For a Block-CSR ILU factorization, this routine updates all blocks in the trailing matrix.

Parameters
[in]size_bmagma_int_t blocksize in BCSR
[in]kblocksmagma_int_t number of blocks
[in]dAmagmaDoubleComplex_ptr * matrix in BCSR
[in]ipivmagmaInt_ptr array containing pivots
[in]queuemagma_queue_t Queue to execute in.
magma_int_t magma_zbcsrswp ( magma_int_t  r_blocks,
magma_int_t  size_b,
magmaInt_ptr  ipiv,
magmaDoubleComplex_ptr  x,
magma_queue_t  queue 
)

For a Block-CSR ILU factorization, this routine swaps rows in the vector *x according to the pivoting in *ipiv.

Parameters
[in]r_blocksmagma_int_t number of blocks
[in]size_bmagma_int_t blocksize in BCSR
[in]ipivmagma_int_t* array containing pivots
[in]xmagmaDoubleComplex_ptr input/output vector x
[in]queuemagma_queue_t Queue to execute in.
magma_int_t magma_zbcsrtrsv ( magma_uplo_t  uplo,
magma_int_t  r_blocks,
magma_int_t  c_blocks,
magma_int_t  size_b,
magmaDoubleComplex_ptr  A,
magma_index_t *  blockinfo,
magmaDoubleComplex_ptr  x,
magma_queue_t  queue 
)

For a Block-CSR ILU factorization, this routine performs the triangular solves.

Parameters
[in]uplomagma_uplo_t upper/lower fill structure
[in]r_blocksmagma_int_t number of blocks in row
[in]c_blocksmagma_int_t number of blocks in column
[in]size_bmagma_int_t blocksize in BCSR
[in]AmagmaDoubleComplex_ptr upper/lower factor
[in]blockinfomagma_int_t* array containing matrix information
[in]xmagmaDoubleComplex_ptr input/output vector x
[in]queuemagma_queue_t Queue to execute in.
magma_int_t magma_zbcsrvalcpy ( magma_int_t  size_b,
magma_int_t  num_blocks,
magma_int_t  num_zblocks,
magmaDoubleComplex_ptr *  Aval,
magmaDoubleComplex_ptr *  Bval,
magmaDoubleComplex_ptr *  Bval2,
magma_queue_t  queue 
)

For a Block-CSR ILU factorization, this routine copies the filled blocks from the original matrix A and initializes the blocks that will later be filled in the factorization process with zeros.

Parameters
[in]size_bmagma_int_t blocksize in BCSR
[in]num_blocksmagma_int_t number of nonzero blocks
[in]num_zblocksmagma_int_t number of zero-blocks (will later be filled)
[in]AvalmagmaDoubleComplex_ptr * pointers to the nonzero blocks in A
[in]BvalmagmaDoubleComplex_ptr * pointers to the nonzero blocks in B
[in]Bval2magmaDoubleComplex_ptr * pointers to the zero blocks in B
[in]queuemagma_queue_t Queue to execute in.
int magma_zbicgmerge1 ( int  n,
magmaDoubleComplex_ptr  skp,
magmaDoubleComplex_ptr  v,
magmaDoubleComplex_ptr  r,
magmaDoubleComplex_ptr  p 
)

Mergels multiple operations into one kernel:

p = beta*p p = p-omega*beta*v p = p+r

-> p = r + beta * ( p - omega * v )

Parameters
[in]nint dimension n
[in]skpmagmaDoubleComplex_ptr set of scalar parameters
[in]vmagmaDoubleComplex_ptr input v
[in]rmagmaDoubleComplex_ptr input r
in/out]p magmaDoubleComplex_ptr input/output p
[in]queuemagma_queue_t Queue to execute in.
int magma_zbicgmerge2 ( int  n,
magmaDoubleComplex_ptr  skp,
magmaDoubleComplex_ptr  r,
magmaDoubleComplex_ptr  v,
magmaDoubleComplex_ptr  s 
)

Mergels multiple operations into one kernel:

s=r s=s-alpha*v

-> s = r - alpha * v

Parameters
[in]nint dimension n
[in]skpmagmaDoubleComplex_ptr set of scalar parameters
[in]rmagmaDoubleComplex_ptr input r
[in]vmagmaDoubleComplex_ptr input v
s]s magmaDoubleComplex_ptr output s
[in]queuemagma_queue_t Queue to execute in.
int magma_zbicgmerge3 ( int  n,
magmaDoubleComplex_ptr  skp,
magmaDoubleComplex_ptr  p,
magmaDoubleComplex_ptr  s,
magmaDoubleComplex_ptr  t,
magmaDoubleComplex_ptr  x,
magmaDoubleComplex_ptr  r 
)

Mergels multiple operations into one kernel:

x=x+alpha*p x=x+omega*s r=s r=r-omega*t

-> x = x + alpha * p + omega * s -> r = s - omega * t

Parameters
[in]nint dimension n
[in]skpmagmaDoubleComplex_ptr set of scalar parameters
[in]pmagmaDoubleComplex_ptr input p
[in]smagmaDoubleComplex_ptr input s
[in]tmagmaDoubleComplex_ptr input t
in/out]x magmaDoubleComplex_ptr input/output x
in/out]r magmaDoubleComplex_ptr input/output r
[in]queuemagma_queue_t Queue to execute in.
int magma_zbicgmerge4 ( int  type,
magmaDoubleComplex_ptr  skp 
)

Performs some parameter operations for the BiCGSTAB with scalars on GPU.

Parameters
[in]typeint kernel type
in/out]skp magmaDoubleComplex_ptr vector with parameters
[in]queuemagma_queue_t Queue to execute in.
magma_int_t magma_zbicgmerge_spmv1 ( magma_z_sparse_matrix  A,
magmaDoubleComplex_ptr  d1,
magmaDoubleComplex_ptr  d2,
magmaDoubleComplex_ptr  dp,
magmaDoubleComplex_ptr  dr,
magmaDoubleComplex_ptr  dv,
magmaDoubleComplex_ptr  skp,
magma_queue_t  queue 
)

Merges the first SpmV using CSR with the dot product and the computation of alpha.

Parameters
[in]Amagma_z_sparse_matrix system matrix
[in]d1magmaDoubleComplex_ptr temporary vector
[in]d2magmaDoubleComplex_ptr temporary vector
[in]dpmagmaDoubleComplex_ptr input vector p
[in]drmagmaDoubleComplex_ptr input vector r
[in]dvmagmaDoubleComplex_ptr output vector v
in/out]skp magmaDoubleComplex_ptr array for parameters ( skp[0]=alpha )
[in]queuemagma_queue_t Queue to execute in.
magma_int_t magma_zbicgmerge_spmv2 ( magma_z_sparse_matrix  A,
magmaDoubleComplex_ptr  d1,
magmaDoubleComplex_ptr  d2,
magmaDoubleComplex_ptr  ds,
magmaDoubleComplex_ptr  dt,
magmaDoubleComplex_ptr  skp,
magma_queue_t  queue 
)

Merges the second SpmV using CSR with the dot product and the computation of omega.

Parameters
[in]Amagma_z_sparse_matrix input matrix
[in]d1magmaDoubleComplex_ptr temporary vector
[in]d2magmaDoubleComplex_ptr temporary vector
[in]dsmagmaDoubleComplex_ptr input vector s
[in]dtmagmaDoubleComplex_ptr output vector t
in/out]skp magmaDoubleComplex_ptr array for parameters
[in]queuemagma_queue_t Queue to execute in.
magma_int_t magma_zbicgmerge_xrbeta ( int  n,
magmaDoubleComplex_ptr  d1,
magmaDoubleComplex_ptr  d2,
magmaDoubleComplex_ptr  rr,
magmaDoubleComplex_ptr  r,
magmaDoubleComplex_ptr  p,
magmaDoubleComplex_ptr  s,
magmaDoubleComplex_ptr  t,
magmaDoubleComplex_ptr  x,
magmaDoubleComplex_ptr  skp,
magma_queue_t  queue 
)

Merges the second SpmV using CSR with the dot product and the computation of omega.

Parameters
[in]nint dimension n
[in]d1magmaDoubleComplex_ptr temporary vector
[in]d2magmaDoubleComplex_ptr temporary vector
[in]rrmagmaDoubleComplex_ptr input vector rr
[in]rmagmaDoubleComplex_ptr input/output vector r
[in]pmagmaDoubleComplex_ptr input vector p
[in]smagmaDoubleComplex_ptr input vector s
[in]tmagmaDoubleComplex_ptr input vector t
[out]xmagmaDoubleComplex_ptr output vector x
[in]skpmagmaDoubleComplex_ptr array for parameters
[in]queuemagma_queue_t Queue to execute in.
magma_int_t magma_zcgmerge_spmv1 ( magma_z_sparse_matrix  A,
magmaDoubleComplex_ptr  d1,
magmaDoubleComplex_ptr  d2,
magmaDoubleComplex_ptr  dd,
magmaDoubleComplex_ptr  dz,
magmaDoubleComplex_ptr  skp,
magma_queue_t  queue 
)

Merges the first SpmV using different formats with the dot product and the computation of rho.

Parameters
[in]Amagma_z_sparse_matrix input matrix
[in]d1magmaDoubleComplex_ptr temporary vector
[in]d2magmaDoubleComplex_ptr temporary vector
[in]ddmagmaDoubleComplex_ptr input vector d
[out]dzmagmaDoubleComplex_ptr input vector z
[out]skpmagmaDoubleComplex_ptr array for parameters ( skp[3]=rho )
[in]queuemagma_queue_t Queue to execute in.
void magma_zcompact ( magma_int_t  m,
magma_int_t  n,
magmaDoubleComplex_ptr  dA,
magma_int_t  ldda,
magmaDouble_ptr  dnorms,
double  tol,
magmaInt_ptr  active,
magmaInt_ptr  cBlock,
magma_queue_t  queue 
)

ZCOMPACT takes a set of n vectors of size m (in dA) and their norms and compacts them into the cBlock size<=n vectors that have norms > tol.

The active mask array has 1 or 0, showing if a vector remained or not in the compacted resulting set of vectors.

Parameters
[in]mINTEGER The number of rows of the matrix dA. M >= 0.
[in]nINTEGER The number of columns of the matrix dA. N >= 0.
[in]in,out]dA COMPLEX DOUBLE PRECISION array, dimension (LDDA,N) The m by n matrix dA.
[in]lddaINTEGER The leading dimension of the array dA. LDDA >= max(1,M).
[in]dnormsDOUBLE PRECISION array, dimension N The norms of the N vectors in dA
[in]tolDOUBLE PRECISON The tolerance value used in the criteria to compact or not.
[in]out]active INTEGER array, dimension N A mask of 1s and 0s showing if a vector remains or has been removed
[in]out]cBlock magmaInt_ptr The number of vectors that remain in dA (i.e., with norms > tol).
[in]queuemagma_queue_t Queue to execute in.
magma_int_t magma_zjacobisetup_vector_gpu ( int  num_rows,
magma_z_vector  b,
magma_z_vector  d,
magma_z_vector  c,
magma_z_vector *  x,
magma_queue_t  queue 
)

Prepares the Jacobi Iteration according to x^(k+1) = D^(-1) * b - D^(-1) * (L+U) * x^k x^(k+1) = c - M * x^k.

Returns the vector c. It calls a GPU kernel

Parameters
[in]num_rowsmagma_int_t number of rows
[in]bmagma_z_vector RHS b
[in]dmagma_z_vector vector with diagonal entries
[out]cmagma_z_vector* c = D^(-1) * b
[out]xmagma_z_vector* iteration vector
[in]queuemagma_queue_t Queue to execute in.
magma_int_t magma_zlobpcg_maxpy ( magma_int_t  num_rows,
magma_int_t  num_vecs,
magmaDoubleComplex_ptr  X,
magmaDoubleComplex_ptr  Y,
magma_queue_t  queue 
)

This routine computes a axpy for a mxn matrix:

Y = X + Y

It replaces: magma_zaxpy(m*n, c_one, Y, 1, X, 1);

/ x1[0] x2[0] x3[0] \
| x1[1] x2[1] x3[1] |

X = | x1[2] x2[2] x3[2] | = x1[0] x1[1] x1[2] x1[3] x1[4] x2[0] x2[1] . | x1[3] x2[3] x3[3] | \ x1[4] x2[4] x3[4] /

Parameters
[in]num_rowsmagma_int_t number of rows
[in]num_vecsmagma_int_t number of vectors
[in]XmagmaDoubleComplex_ptr input vector X
in/out]Y magmaDoubleComplex_ptr input/output vector Y
[in]queuemagma_queue_t Queue to execute in.