root / man / man3 / HPL_indxg2lp.3
Historique | Voir | Annoter | Télécharger (1,94 ko)
| 1 | 1 | equemene | .TH HPL_indxg2lp 3 "September 10, 2008" "HPL 2.0" "HPL Library Functions" |
|---|---|---|---|
| 2 | 1 | equemene | .SH NAME |
| 3 | 1 | equemene | HPL_indxg2lp \- Map a local index into a global one. |
| 4 | 1 | equemene | .SH SYNOPSIS |
| 5 | 1 | equemene | \fB\&#include "hpl.h"\fR |
| 6 | 1 | equemene | |
| 7 | 1 | equemene | \fB\&void\fR |
| 8 | 1 | equemene | \fB\&HPL_indxg2lp(\fR |
| 9 | 1 | equemene | \fB\&int *\fR |
| 10 | 1 | equemene | \fI\&IL\fR, |
| 11 | 1 | equemene | \fB\&int *\fR |
| 12 | 1 | equemene | \fI\&PROC\fR, |
| 13 | 1 | equemene | \fB\&const int\fR |
| 14 | 1 | equemene | \fI\&IG\fR, |
| 15 | 1 | equemene | \fB\&const int\fR |
| 16 | 1 | equemene | \fI\&INB\fR, |
| 17 | 1 | equemene | \fB\&const int\fR |
| 18 | 1 | equemene | \fI\&NB\fR, |
| 19 | 1 | equemene | \fB\&const int\fR |
| 20 | 1 | equemene | \fI\&SRCPROC\fR, |
| 21 | 1 | equemene | \fB\&const int\fR |
| 22 | 1 | equemene | \fI\&NPROCS\fR |
| 23 | 1 | equemene | \fB\&);\fR |
| 24 | 1 | equemene | .SH DESCRIPTION |
| 25 | 1 | equemene | \fB\&HPL_indxg2lp\fR |
| 26 | 1 | equemene | computes the local index of a matrix entry pointed to by |
| 27 | 1 | equemene | the global index IG as well as the process coordinate which posseses |
| 28 | 1 | equemene | this entry. The local returned index is the same in all processes. |
| 29 | 1 | equemene | .SH ARGUMENTS |
| 30 | 1 | equemene | .TP 8 |
| 31 | 1 | equemene | IL (output) int * |
| 32 | 1 | equemene | On exit, IL specifies the local index corresponding to IG. IL |
| 33 | 1 | equemene | is at least zero. |
| 34 | 1 | equemene | .TP 8 |
| 35 | 1 | equemene | PROC (output) int * |
| 36 | 1 | equemene | On exit, PROC is the coordinate of the process owning the |
| 37 | 1 | equemene | entry specified by the global index IG. PROC is at least zero |
| 38 | 1 | equemene | and less than NPROCS. |
| 39 | 1 | equemene | .TP 8 |
| 40 | 1 | equemene | IG (input) const int |
| 41 | 1 | equemene | On entry, IG specifies the global index of the matrix entry. |
| 42 | 1 | equemene | IG must be at least zero. |
| 43 | 1 | equemene | .TP 8 |
| 44 | 1 | equemene | INB (input) const int |
| 45 | 1 | equemene | On entry, INB specifies the size of the first block of the |
| 46 | 1 | equemene | global matrix. INB must be at least one. |
| 47 | 1 | equemene | .TP 8 |
| 48 | 1 | equemene | NB (input) const int |
| 49 | 1 | equemene | On entry, NB specifies the blocking factor used to partition |
| 50 | 1 | equemene | and distribute the matrix A. NB must be larger than one. |
| 51 | 1 | equemene | .TP 8 |
| 52 | 1 | equemene | SRCPROC (input) const int |
| 53 | 1 | equemene | On entry, if SRCPROC = -1, the data is not distributed but |
| 54 | 1 | equemene | replicated, in which case this routine returns IG in all |
| 55 | 1 | equemene | processes. Otherwise, the value of SRCPROC is ignored. |
| 56 | 1 | equemene | .TP 8 |
| 57 | 1 | equemene | NPROCS (input) const int |
| 58 | 1 | equemene | On entry, NPROCS specifies the total number of process rows |
| 59 | 1 | equemene | or columns over which the matrix is distributed. NPROCS must |
| 60 | 1 | equemene | be at least one. |
| 61 | 1 | equemene | .SH SEE ALSO |
| 62 | 1 | equemene | .BR HPL_indxg2l \ (3), |
| 63 | 1 | equemene | .BR HPL_indxg2p \ (3), |
| 64 | 1 | equemene | .BR HPL_indxl2g \ (3), |
| 65 | 1 | equemene | .BR HPL_numroc \ (3), |
| 66 | 1 | equemene | .BR HPL_numrocI \ (3). |