extern "C" magma_int_t
magma_cgeqrf2_mgpu( magma_int_t num_gpus, magma_int_t m, magma_int_t n,
magmaFloatComplex **dlA, magma_int_t ldda,
magmaFloatComplex *tau,
magma_int_t *info )
{
/* -- MAGMA (version 1.4.1) --
Univ. of Tennessee, Knoxville
Univ. of California, Berkeley
Univ. of Colorado, Denver
December 2013
Purpose
=======
CGEQRF2_MGPU computes a QR factorization of a complex M-by-N matrix A:
A = Q * R. This is a GPU interface of the routine.
Arguments
=========
M (input) INTEGER
The number of rows of the matrix A. M >= 0.
N (input) INTEGER
The number of columns of the matrix A. N >= 0.
dA (input/output) COMPLEX array on the GPU, dimension (LDDA,N)
On entry, the M-by-N matrix dA.
On exit, the elements on and above the diagonal of the array
contain the min(M,N)-by-N upper trapezoidal matrix R (R is
upper triangular if m >= n); the elements below the diagonal,
with the array TAU, represent the orthogonal matrix Q as a
product of min(m,n) elementary reflectors (see Further
Details).
LDDA (input) INTEGER
The leading dimension of the array dA. LDDA >= max(1,M).
To benefit from coalescent memory accesses LDDA must be
dividable by 16.
TAU (output) COMPLEX array, dimension (min(M,N))
The scalar factors of the elementary reflectors (see Further
Details).
INFO (output) INTEGER
= 0: successful exit
< 0: if INFO = -i, the i-th argument had an illegal value
or another error occured, such as memory allocation failed.
Further Details
===============
The matrix Q is represented as a product of elementary reflectors
Q = H(1) H(2) . . . H(k), where k = min(m,n).
Each H(i) has the form
H(i) = I - tau * v * v'
where tau is a complex scalar, and v is a complex vector with
v(1:i-1) = 0 and v(i) = 1; v(i+1:m) is stored on exit in A(i+1:m,i),
and tau in TAU(i).
===================================================================== */
#define dlA(dev, i, j) (dlA[dev] + (i) + (j)*(ldda))
#define hpanel(i) (hpanel + (i))
// set to NULL to make cleanup easy: free(NULL) does nothing.
magmaFloatComplex *dwork[MagmaMaxGPUs]={NULL}, *dpanel[MagmaMaxGPUs]={NULL};
magmaFloatComplex *hwork=NULL, *hpanel=NULL;
magma_queue_t stream[MagmaMaxGPUs][2]={{NULL}};
magma_event_t panel_event[MagmaMaxGPUs]={NULL};
magma_int_t i, j, min_mn, dev, ldhpanel, lddwork, rows;
magma_int_t ib, nb;
magma_int_t lhwork, lwork;
magma_int_t panel_dev, i_local, i_nb_local, n_local[MagmaMaxGPUs], la_dev, dpanel_offset;
magma_queue_t cqueue;
magmablasGetKernelStream( &cqueue );
magma_device_t cdevice;
magma_getdevice( &cdevice );
*info = 0;
if (m < 0) {
*info = -1;
} else if (n < 0) {
*info = -2;
} else if (ldda < max(1,m)) {
*info = -4;
}
if (*info != 0) {
magma_xerbla( __func__, -(*info) );
return *info;
}
min_mn = min(m,n);
if (min_mn == 0)
return *info;
nb = magma_get_cgeqrf_nb( m );
/* dwork is (n*nb) --- for T (nb*nb) and clarfb work ((n-nb)*nb) ---
* + dpanel (ldda*nb), on each GPU.
* I think clarfb work could be smaller, max(n_local[:]).
* Oddly, T and clarfb work get stacked on top of each other, both with lddwork=n.
* on GPU that owns panel, set dpanel = dlA(dev,i,i_local).
* on other GPUs, set dpanel = dwork[dev] + dpanel_offset. */
lddwork = n;
dpanel_offset = lddwork*nb;
for( dev=0; dev < num_gpus; dev++ ) {
magma_setdevice( dev );
if ( MAGMA_SUCCESS != magma_cmalloc( &(dwork[dev]), (lddwork + ldda)*nb )) {
*info = MAGMA_ERR_DEVICE_ALLOC;
goto CLEANUP;
}
}
/* hwork is MAX( workspace for cgeqrf (n*nb), two copies of T (2*nb*nb) )
* + hpanel (m*nb).
* for last block, need 2*n*nb total. */
ldhpanel = m;
lhwork = max( n*nb, 2*nb*nb );
lwork = max( lhwork + ldhpanel*nb, 2*n*nb );
if ( MAGMA_SUCCESS != magma_cmalloc_pinned( &hwork, lwork )) {
*info = MAGMA_ERR_HOST_ALLOC;
goto CLEANUP;
}
hpanel = hwork + lhwork;
/* Set the number of local n for each GPU */
for( dev=0; dev < num_gpus; dev++ ) {
n_local[dev] = ((n/nb)/num_gpus)*nb;
if (dev < (n/nb) % num_gpus)
n_local[dev] += nb;
else if (dev == (n/nb) % num_gpus)
n_local[dev] += n % nb;
}
for( dev=0; dev < num_gpus; dev++ ) {
magma_setdevice( dev );
magma_queue_create( &stream[dev][0] );
magma_queue_create( &stream[dev][1] );
magma_event_create( &panel_event[dev] );
}
if ( nb < min_mn ) {
/* Use blocked code initially */
// Note: as written, ib cannot be < nb.
for( i = 0; i < min_mn-nb; i += nb ) {
/* Set the GPU number that holds the current panel */
panel_dev = (i/nb) % num_gpus;
/* Set the local index where the current panel is (j==i) */
i_local = i/(nb*num_gpus)*nb;
ib = min(min_mn-i, nb);
rows = m-i;
/* Send current panel to the CPU, after panel_event indicates it has been updated */
magma_setdevice( panel_dev );
magma_queue_wait_event( stream[panel_dev][1], panel_event[panel_dev] );
magma_cgetmatrix_async( rows, ib,
dlA(panel_dev, i, i_local), ldda,
hpanel(i), ldhpanel, stream[panel_dev][1] );
magma_queue_sync( stream[panel_dev][1] );
// Factor panel
lapackf77_cgeqrf( &rows, &ib, hpanel(i), &ldhpanel, tau+i,
hwork, &lhwork, info );
if ( *info != 0 ) {
fprintf( stderr, "error %d\n", (int) *info );
}
// Form the triangular factor of the block reflector
// H = H(i) H(i+1) . . . H(i+ib-1)
lapackf77_clarft( MagmaForwardStr, MagmaColumnwiseStr,
&rows, &ib,
hpanel(i), &ldhpanel, tau+i, hwork, &ib );
cpanel_to_q( MagmaUpper, ib, hpanel(i), ldhpanel, hwork + ib*ib );
// Send the current panel back to the GPUs
for( dev=0; dev < num_gpus; dev++ ) {
magma_setdevice( dev );
if (dev == panel_dev)
dpanel[dev] = dlA(dev, i, i_local);
else
dpanel[dev] = dwork[dev] + dpanel_offset;
magma_csetmatrix_async( rows, ib,
hpanel(i), ldhpanel,
dpanel[dev], ldda, stream[dev][0] );
}
for( dev=0; dev < num_gpus; dev++ ) {
magma_setdevice( dev );
magma_queue_sync( stream[dev][0] );
}
// TODO: if cpanel_to_q copied whole block, wouldn't need to restore
// -- just send the copy to the GPUs.
// TODO: also, could zero out the lower triangle and use Azzam's larfb w/ gemm.
/* Restore the panel */
cq_to_panel( MagmaUpper, ib, hpanel(i), ldhpanel, hwork + ib*ib );
if (i + ib < n) {
/* Send the T matrix to the GPU. */
for( dev=0; dev < num_gpus; dev++ ) {
magma_setdevice( dev );
magma_csetmatrix_async( ib, ib,
hwork, ib,
dwork[dev], lddwork, stream[dev][0] );
}
la_dev = (panel_dev+1) % num_gpus;
for( dev=0; dev < num_gpus; dev++ ) {
magma_setdevice( dev );
magmablasSetKernelStream( stream[dev][0] );
if (dev == la_dev && i+nb < min_mn-nb) {
// If not last panel,
// for look-ahead panel, apply H' to A(i:m,i+ib:i+2*ib)
i_nb_local = (i+nb)/(nb*num_gpus)*nb;
magma_clarfb_gpu( MagmaLeft, MagmaConjTrans, MagmaForward, MagmaColumnwise,
rows, ib, ib,
dpanel[dev], ldda, // V
dwork[dev], lddwork, // T
dlA(dev, i, i_nb_local), ldda, // C
dwork[dev]+ib, lddwork ); // work
magma_event_record( panel_event[dev], stream[dev][0] );
// for trailing matrix, apply H' to A(i:m,i+2*ib:n)
magma_clarfb_gpu( MagmaLeft, MagmaConjTrans, MagmaForward, MagmaColumnwise,
rows, n_local[dev]-(i_nb_local+ib), ib,
dpanel[dev], ldda, // V
dwork[dev], lddwork, // T
dlA(dev, i, i_nb_local+ib), ldda, // C
dwork[dev]+ib, lddwork ); // work
}
else {
// for trailing matrix, apply H' to A(i:m,i+ib:n)
i_nb_local = i_local;
if (dev <= panel_dev) {
i_nb_local += ib;
}
magma_clarfb_gpu( MagmaLeft, MagmaConjTrans, MagmaForward, MagmaColumnwise,
rows, n_local[dev]-i_nb_local, ib,
dpanel[dev], ldda, // V
dwork[dev], lddwork, // T
dlA(dev, i, i_nb_local), ldda, // C
dwork[dev]+ib, lddwork ); // work
}
}
// Restore top of panel (after larfb is done)
magma_setdevice( panel_dev );
magma_csetmatrix_async( ib, ib,
hpanel(i), ldhpanel,
dlA(panel_dev, i, i_local), ldda, stream[panel_dev][0] );
}
}
}
else {
i = 0;
}
/* Use unblocked code to factor the last or only block row. */
if (i < min_mn) {
rows = m-i;
for( j=i; j < n; j += nb ) {
panel_dev = (j/nb) % num_gpus;
i_local = j/(nb*num_gpus)*nb;
ib = min( n-j, nb );
magma_setdevice( panel_dev );
magma_cgetmatrix( rows, ib,
dlA(panel_dev, i, i_local), ldda,
hwork + (j-i)*rows, rows );
}
// needs lwork >= 2*n*nb:
// needs (m-i)*(n-i) for last block row, bounded by nb*n.
// needs (n-i)*nb for cgeqrf work, bounded by n*nb.
ib = n-i; // total columns in block row
lhwork = lwork - ib*rows;
lapackf77_cgeqrf( &rows, &ib, hwork, &rows, tau+i, hwork + ib*rows, &lhwork, info );
if ( *info != 0 ) {
fprintf( stderr, "error %d\n", (int) *info );
}
for( j=i; j < n; j += nb ) {
panel_dev = (j/nb) % num_gpus;
i_local = j/(nb*num_gpus)*nb;
ib = min( n-j, nb );
magma_setdevice( panel_dev );
magma_csetmatrix( rows, ib,
hwork + (j-i)*rows, rows,
dlA(panel_dev, i, i_local), ldda );
}
}
CLEANUP:
// free(NULL) does nothing.
// check that queues and events are non-zero before destroying them, though.
for( dev=0; dev < num_gpus; dev++ ) {
magma_setdevice( dev );
if ( stream[dev][0] ) { magma_queue_destroy( stream[dev][0] ); }
if ( stream[dev][1] ) { magma_queue_destroy( stream[dev][1] ); }
if ( panel_event[dev] ) { magma_event_destroy( panel_event[dev] ); }
magma_free( dwork[dev] );
}
magma_free_pinned( hwork );
magma_setdevice( cdevice );
magmablasSetKernelStream( cqueue );
return *info;
} /* magma_cgeqrf2_mgpu */
/***************************************************************************//**
Purpose
-------
CGEQRF computes a QR factorization of a complex M-by-N matrix A:
A = Q * R. This is a GPU interface of the routine.
Arguments
---------
@param[in]
ngpu INTEGER
Number of GPUs to use. ngpu > 0.
@param[in]
m INTEGER
The number of rows of the matrix A. M >= 0.
@param[in]
n INTEGER
The number of columns of the matrix A. N >= 0.
@param[in,out]
dlA COMPLEX array of pointers on the GPU, dimension (ngpu).
On entry, the M-by-N matrix A distributed over GPUs
(d_lA[d] points to the local matrix on d-th GPU).
It uses 1D block column cyclic format with the block size of nb,
and each local matrix is stored by column.
On exit, the elements on and above the diagonal of the array
contain the min(M,N)-by-N upper trapezoidal matrix R (R is
upper triangular if m >= n); the elements below the diagonal,
with the array TAU, represent the orthogonal matrix Q as a
product of min(m,n) elementary reflectors (see Further
Details).
@param[in]
ldda INTEGER
The leading dimension of the array dA. LDDA >= max(1,M).
To benefit from coalescent memory accesses LDDA must be
divisible by 16.
@param[out]
tau COMPLEX array, dimension (min(M,N))
The scalar factors of the elementary reflectors (see Further
Details).
@param[out]
info INTEGER
- = 0: successful exit
- < 0: if INFO = -i, the i-th argument had an illegal value
or another error occured, such as memory allocation failed.
Further Details
---------------
The matrix Q is represented as a product of elementary reflectors
Q = H(1) H(2) . . . H(k), where k = min(m,n).
Each H(i) has the form
H(i) = I - tau * v * v'
where tau is a complex scalar, and v is a complex vector with
v(1:i-1) = 0 and v(i) = 1; v(i+1:m) is stored on exit in A(i+1:m,i),
and tau in TAU(i).
@ingroup magma_geqrf
*******************************************************************************/
extern "C" magma_int_t
magma_cgeqrf2_mgpu(
magma_int_t ngpu,
magma_int_t m, magma_int_t n,
magmaFloatComplex_ptr dlA[], magma_int_t ldda,
magmaFloatComplex *tau,
magma_int_t *info )
{
#define dlA(dev, i, j) (dlA[dev] + (i) + (j)*(ldda))
#define hpanel(i) (hpanel + (i))
// set to NULL to make cleanup easy: free(NULL) does nothing.
magmaFloatComplex *dwork[MagmaMaxGPUs]={NULL}, *dpanel[MagmaMaxGPUs]={NULL};
magmaFloatComplex *hwork=NULL, *hpanel=NULL;
magma_queue_t queues[MagmaMaxGPUs][2]={{NULL}};
magma_event_t panel_event[MagmaMaxGPUs]={NULL};
magma_int_t i, j, min_mn, dev, ldhpanel, lddwork, rows;
magma_int_t ib, nb;
magma_int_t lhwork, lwork;
magma_int_t panel_dev, i_local, i_nb_local, n_local[MagmaMaxGPUs], la_dev, dpanel_offset;
*info = 0;
if (m < 0) {
*info = -1;
} else if (n < 0) {
*info = -2;
} else if (ldda < max(1,m)) {
*info = -4;
}
if (*info != 0) {
magma_xerbla( __func__, -(*info) );
return *info;
}
min_mn = min(m,n);
if (min_mn == 0)
return *info;
magma_device_t orig_dev;
magma_getdevice( &orig_dev );
nb = magma_get_cgeqrf_nb( m, n );
/* dwork is (n*nb) --- for T (nb*nb) and clarfb work ((n-nb)*nb) ---
* + dpanel (ldda*nb), on each GPU.
* I think clarfb work could be smaller, max(n_local[:]).
* Oddly, T and clarfb work get stacked on top of each other, both with lddwork=n.
* on GPU that owns panel, set dpanel = dlA(dev,i,i_local).
* on other GPUs, set dpanel = dwork[dev] + dpanel_offset. */
lddwork = n;
dpanel_offset = lddwork*nb;
for( dev=0; dev < ngpu; dev++ ) {
magma_setdevice( dev );
if ( MAGMA_SUCCESS != magma_cmalloc( &(dwork[dev]), (lddwork + ldda)*nb )) {
*info = MAGMA_ERR_DEVICE_ALLOC;
goto CLEANUP;
}
}
/* hwork is MAX( workspace for cgeqrf (n*nb), two copies of T (2*nb*nb) )
* + hpanel (m*nb).
* for last block, need 2*n*nb total. */
ldhpanel = m;
lhwork = max( n*nb, 2*nb*nb );
lwork = max( lhwork + ldhpanel*nb, 2*n*nb );
if ( MAGMA_SUCCESS != magma_cmalloc_pinned( &hwork, lwork )) {
*info = MAGMA_ERR_HOST_ALLOC;
goto CLEANUP;
}
hpanel = hwork + lhwork;
/* Set the number of local n for each GPU */
for( dev=0; dev < ngpu; dev++ ) {
n_local[dev] = ((n/nb)/ngpu)*nb;
if (dev < (n/nb) % ngpu)
n_local[dev] += nb;
else if (dev == (n/nb) % ngpu)
n_local[dev] += n % nb;
}
for( dev=0; dev < ngpu; dev++ ) {
magma_setdevice( dev );
magma_queue_create( dev, &queues[dev][0] );
magma_queue_create( dev, &queues[dev][1] );
magma_event_create( &panel_event[dev] );
}
if ( nb < min_mn ) {
/* Use blocked code initially */
// Note: as written, ib cannot be < nb.
for( i = 0; i < min_mn-nb; i += nb ) {
/* Set the GPU number that holds the current panel */
panel_dev = (i/nb) % ngpu;
/* Set the local index where the current panel is (j == i) */
i_local = i/(nb*ngpu)*nb;
ib = min(min_mn-i, nb);
rows = m-i;
/* Send current panel to the CPU, after panel_event indicates it has been updated */
magma_setdevice( panel_dev );
magma_queue_wait_event( queues[panel_dev][1], panel_event[panel_dev] );
magma_cgetmatrix_async( rows, ib,
dlA(panel_dev, i, i_local), ldda,
hpanel(i), ldhpanel,
queues[panel_dev][1] );
magma_queue_sync( queues[panel_dev][1] );
// Factor panel
lapackf77_cgeqrf( &rows, &ib, hpanel(i), &ldhpanel, tau+i,
hwork, &lhwork, info );
if ( *info != 0 ) {
fprintf( stderr, "error %lld\n", (long long) *info );
}
// Form the triangular factor of the block reflector
// H = H(i) H(i+1) . . . H(i+ib-1)
lapackf77_clarft( MagmaForwardStr, MagmaColumnwiseStr,
&rows, &ib,
hpanel(i), &ldhpanel, tau+i, hwork, &ib );
magma_cpanel_to_q( MagmaUpper, ib, hpanel(i), ldhpanel, hwork + ib*ib );
// Send the current panel back to the GPUs
for( dev=0; dev < ngpu; dev++ ) {
magma_setdevice( dev );
if (dev == panel_dev)
dpanel[dev] = dlA(dev, i, i_local);
else
dpanel[dev] = dwork[dev] + dpanel_offset;
magma_csetmatrix_async( rows, ib,
hpanel(i), ldhpanel,
dpanel[dev], ldda,
queues[dev][0] );
}
for( dev=0; dev < ngpu; dev++ ) {
magma_setdevice( dev );
magma_queue_sync( queues[dev][0] );
}
// TODO: if magma_cpanel_to_q copied whole block, wouldn't need to restore
// -- just send the copy to the GPUs.
// TODO: also, could zero out the lower triangle and use Azzam's larfb w/ gemm.
/* Restore the panel */
magma_cq_to_panel( MagmaUpper, ib, hpanel(i), ldhpanel, hwork + ib*ib );
if (i + ib < n) {
/* Send the T matrix to the GPU. */
for( dev=0; dev < ngpu; dev++ ) {
magma_setdevice( dev );
magma_csetmatrix_async( ib, ib,
hwork, ib,
dwork[dev], lddwork,
queues[dev][0] );
}
la_dev = (panel_dev+1) % ngpu;
for( dev=0; dev < ngpu; dev++ ) {
magma_setdevice( dev );
if (dev == la_dev && i+nb < min_mn-nb) {
// If not last panel,
// for look-ahead panel, apply H' to A(i:m,i+ib:i+2*ib)
i_nb_local = (i+nb)/(nb*ngpu)*nb;
magma_clarfb_gpu( MagmaLeft, MagmaConjTrans, MagmaForward, MagmaColumnwise,
rows, ib, ib,
dpanel[dev], ldda, // V
dwork[dev], lddwork, // T
dlA(dev, i, i_nb_local), ldda, // C
dwork[dev]+ib, lddwork, // work
queues[dev][0] );
magma_event_record( panel_event[dev], queues[dev][0] );
// for trailing matrix, apply H' to A(i:m,i+2*ib:n)
magma_clarfb_gpu( MagmaLeft, MagmaConjTrans, MagmaForward, MagmaColumnwise,
rows, n_local[dev]-(i_nb_local+ib), ib,
dpanel[dev], ldda, // V
dwork[dev], lddwork, // T
dlA(dev, i, i_nb_local+ib), ldda, // C
dwork[dev]+ib, lddwork, // work
queues[dev][0] );
}
else {
// for trailing matrix, apply H' to A(i:m,i+ib:n)
i_nb_local = i_local;
if (dev <= panel_dev) {
i_nb_local += ib;
}
magma_clarfb_gpu( MagmaLeft, MagmaConjTrans, MagmaForward, MagmaColumnwise,
rows, n_local[dev]-i_nb_local, ib,
dpanel[dev], ldda, // V
dwork[dev], lddwork, // T
dlA(dev, i, i_nb_local), ldda, // C
dwork[dev]+ib, lddwork, // work
queues[dev][0] );
}
}
// Restore top of panel (after larfb is done)
magma_setdevice( panel_dev );
magma_csetmatrix_async( ib, ib,
hpanel(i), ldhpanel,
dlA(panel_dev, i, i_local), ldda,
queues[panel_dev][0] );
}
}
}
else {
i = 0;
}
/* Use unblocked code to factor the last or only block row. */
if (i < min_mn) {
rows = m-i;
for( j=i; j < n; j += nb ) {
panel_dev = (j/nb) % ngpu;
i_local = j/(nb*ngpu)*nb;
ib = min( n-j, nb );
magma_setdevice( panel_dev );
magma_cgetmatrix( rows, ib,
dlA(panel_dev, i, i_local), ldda,
hwork + (j-i)*rows, rows,
queues[panel_dev][0] );
}
// needs lwork >= 2*n*nb:
// needs (m-i)*(n-i) for last block row, bounded by nb*n.
// needs (n-i)*nb for cgeqrf work, bounded by n*nb.
ib = n-i; // total columns in block row
lhwork = lwork - ib*rows;
lapackf77_cgeqrf( &rows, &ib, hwork, &rows, tau+i, hwork + ib*rows, &lhwork, info );
if ( *info != 0 ) {
fprintf( stderr, "error %lld\n", (long long) *info );
}
for( j=i; j < n; j += nb ) {
panel_dev = (j/nb) % ngpu;
i_local = j/(nb*ngpu)*nb;
ib = min( n-j, nb );
magma_setdevice( panel_dev );
magma_csetmatrix( rows, ib,
hwork + (j-i)*rows, rows,
dlA(panel_dev, i, i_local), ldda,
queues[panel_dev][0] );
}
}
CLEANUP:
// free(NULL) does nothing.
for( dev=0; dev < ngpu; dev++ ) {
magma_setdevice( dev );
magma_queue_destroy( queues[dev][0] );
magma_queue_destroy( queues[dev][1] );
magma_event_destroy( panel_event[dev] );
magma_free( dwork[dev] );
}
magma_free_pinned( hwork );
magma_setdevice( orig_dev );
return *info;
} /* magma_cgeqrf2_mgpu */