change integer type of permr[]/perm_c[] from int_t to int, always 32bit.

PYTHON/pdbridge.c

@@ -747,7 +747,7 @@ int dcreate_matrix_from_csc(SuperMatrix *A,
 }
 
 
-int count_swaps(int_t perm[], int n) {
+int count_swaps(int perm[], int n) {
     bool visited[n];
     for (int i = 0; i < n; i++) {
         visited[i] = false;

PYTHON/pdbridge.h

@@ -51,4 +51,4 @@ extern void pdbridge_logdet(void ** pyobj, int * sign, double * logdet);
 #endif 
 
 int dcreate_matrix_from_csc(SuperMatrix *A, int_t m, int_t n, int_t nnz, int_t *rowind0, int_t *colptr0, double *nzval0, gridinfo_t *grid);
-int count_swaps(int_t perm[], int n);
+int count_swaps(int perm[], int n);

SRC/CMakeLists.txt

@@ -269,11 +269,7 @@ if (TPL_ENABLE_CUDALIB)
     list(APPEND sources cuda/psgstrs_lsum_cuda.cu cuda/ssuperlu_gpu.cu)
 endif()
 if (HAVE_COMBBLAS)
-   if (enable_double)
-       list(APPEND sources double/d_c2cpp_GetHWPM.cpp)
-   else()
        list(APPEND sources single/s_c2cpp_GetHWPM.cpp double/dHWPM_CombBLAS.hpp)
-   endif()
 endif()
 
 endif() ########## enable single
@@ -465,4 +461,4 @@ install(TARGETS ${targets}
 install(FILES ${headers}
 # DESTINATION ${CMAKE_INSTALL_PREFIX}/include)
   DESTINATION ${CMAKE_INSTALL_INCLUDEDIR}
-)
+)

SRC/complex16/pzdistribute.c

@@ -69,8 +69,8 @@ zReDistribute_A(SuperMatrix *A, zScalePermstruct_t *ScalePermstruct,
                 doublecomplex *a[])
 {
     NRformat_loc *Astore;
-    int_t  *perm_r; /* row permutation vector */
-    int_t  *perm_c; /* column permutation vector */
+    int  *perm_r; /* row permutation vector */
+    int  *perm_c; /* column permutation vector */
     int_t  i, irow, fst_row, j, jcol, k, gbi, gbj, n, m_loc, jsize,nnz_tot;
     int_t  nnz_loc;    /* number of local nonzeros */
     int_t  SendCnt; /* number of remote nonzeros to be sent */

SRC/complex16/pzgssvx.c

@@ -530,8 +530,8 @@ pzgssvx(superlu_dist_options_t *options, SuperMatrix *A,
     fact_t  Fact;
     doublecomplex *a;
     int_t   *colptr, *rowind;
-    int_t   *perm_r; /* row permutations from partial pivoting */
-    int_t   *perm_c; /* column permutation vector */
+    int   *perm_r; /* row permutations from partial pivoting */
+    int   *perm_c; /* column permutation vector */
     int_t   *etree;  /* elimination tree */
     int_t   *rowptr, *colind;  /* Local A in NR*/
     int_t   nnz_loc, nnz;
@@ -812,7 +812,8 @@ pzgssvx(superlu_dist_options_t *options, SuperMatrix *A,
 
                         MPI_Bcast( &iinfo, 1, MPI_INT, 0, grid->comm );
 		        if ( iinfo == 0 ) {
-		            MPI_Bcast( perm_r, m, mpi_int_t, 0, grid->comm );
+			    //MPI_Bcast( perm_r, m, mpi_int_t, 0, grid->comm );
+			    MPI_Bcast( perm_r, m, MPI_INT, 0, grid->comm );
 		            if ( job == 5 && Equil ) {
 		                MPI_Bcast( R1, m, MPI_DOUBLE, 0, grid->comm );
 		                MPI_Bcast( C1, n, MPI_DOUBLE, 0, grid->comm );
@@ -821,7 +822,8 @@ pzgssvx(superlu_dist_options_t *options, SuperMatrix *A,
 	            } else {
 		        MPI_Bcast( &iinfo, 1, MPI_INT, 0, grid->comm );
 			if ( iinfo == 0 ) {
-		            MPI_Bcast( perm_r, m, mpi_int_t, 0, grid->comm );
+		            //MPI_Bcast( perm_r, m, mpi_int_t, 0, grid->comm );
+		            MPI_Bcast( perm_r, m, MPI_INT, 0, grid->comm );
 		            if ( job == 5 && Equil ) {
 		                MPI_Bcast( R1, m, MPI_DOUBLE, 0, grid->comm );
 		                MPI_Bcast( C1, n, MPI_DOUBLE, 0, grid->comm );
@@ -935,7 +937,7 @@ pzgssvx(superlu_dist_options_t *options, SuperMatrix *A,
         }
 
 #if ( DEBUGlevel>=2 )
-        if ( !iam ) PrintInt10("perm_r",  m, perm_r);
+        if ( !iam ) PrintInt32("perm_r",  m, perm_r);
 #endif
     } /* end if (!factored) */
 
@@ -1365,7 +1367,7 @@ pzgssvx(superlu_dist_options_t *options, SuperMatrix *A,
 	    int buffer_peak_rank = global_struct.rank;
 	    float buffer_peak = global_struct.val*1e-6;
 
-            if ( iam==0 ) {
+            if ( iam == 0 ) {
                 printf("** Total highmark (MB):\n"
 		       "    Sum-of-all : %8.2f | Avg : %8.2f  | Max : %8.2f\n",
 		       avg * 1e-6,

SRC/complex16/pzgssvx3d.c

@@ -543,8 +543,8 @@ void pzgssvx3d(superlu_dist_options_t *options, SuperMatrix *A,
     fact_t Fact;
     doublecomplex *a;
     int_t *colptr, *rowind;
-    int_t *perm_r;			/* row permutations from partial pivoting */
-    int_t *perm_c;			/* column permutation vector */
+    int *perm_r;			/* row permutations from partial pivoting */
+    int *perm_c;			/* column permutation vector */
     int_t *etree;			/* elimination tree */
     int_t *rowptr, *colind; /* Local A in NR */
     int colequ, Equil, factored, job, notran, rowequ, need_value;

SRC/complex16/pzgssvx3d_csc_batch.c

@@ -409,7 +409,7 @@ pzgssvx3d_csc_batch(
 #if (PRNTlevel >= 1)
     printf("\tBIG system: berr[0] %e\n", berr[0]);
     printf("after pzgssvx3d: DiagScale %d\n", ScalePermstruct.DiagScale);
-    //PrintInt10("after pdgssvx3d: ScalePermstruct.perm_c", (int_t) m_big, ScalePermstruct.perm_c);
+    //PrintInt32("after pdgssvx3d: ScalePermstruct.perm_c", m_big, ScalePermstruct.perm_c);
     //Printdouble5("big-B-solution", m_big, b);
 #endif
 

SRC/complex16/pzgssvx_ABglobal.c

@@ -491,8 +491,8 @@ pzgssvx_ABglobal(superlu_dist_options_t *options, SuperMatrix *A,
 	      structures are not used.                                  */
     fact_t   Fact;
     doublecomplex   *a;
-    int_t    *perm_r; /* row permutations from partial pivoting */
-    int_t    *perm_c; /* column permutation vector */
+    int    *perm_r; /* row permutations from partial pivoting */
+    int    *perm_c; /* column permutation vector */
     int_t    *etree;  /* elimination tree */
     int_t    *colptr, *rowind;
     int_t    Equil, factored, job, notran, colequ, rowequ;
@@ -724,7 +724,7 @@ pzgssvx_ABglobal(superlu_dist_options_t *options, SuperMatrix *A,
 
                 MPI_Bcast( &iinfo, 1, mpi_int_t, 0, grid->comm );
 		if ( iinfo == 0 ) {
-		    MPI_Bcast( perm_r, m, mpi_int_t, 0, grid->comm );
+		    MPI_Bcast( perm_r, m, MPI_INT, 0, grid->comm );
 		    if ( job == 5 && Equil ) {
 		       MPI_Bcast( R1, m, MPI_DOUBLE, 0, grid->comm );
 		       MPI_Bcast( C1, n, MPI_DOUBLE, 0, grid->comm );
@@ -733,7 +733,7 @@ pzgssvx_ABglobal(superlu_dist_options_t *options, SuperMatrix *A,
 	    } else {
 		MPI_Bcast( &iinfo, 1, mpi_int_t, 0, grid->comm );
 		if ( iinfo == 0 ) {
-		   MPI_Bcast( perm_r, m, mpi_int_t, 0, grid->comm );
+		   MPI_Bcast( perm_r, m, MPI_INT, 0, grid->comm );
 		   if ( job == 5 && Equil ) {
 		      MPI_Bcast( R1, m, MPI_DOUBLE, 0, grid->comm );
 		      MPI_Bcast( C1, n, MPI_DOUBLE, 0, grid->comm );

SRC/complex16/pzgstrs.c

@@ -170,7 +170,7 @@ pzReDistribute_B_to_X(doublecomplex *B, int_t m_loc, int nrhs, int_t ldb,
     int  *SendCnt, *SendCnt_nrhs, *RecvCnt, *RecvCnt_nrhs;
     int  *sdispls, *sdispls_nrhs, *rdispls, *rdispls_nrhs;
     int  *ptr_to_ibuf, *ptr_to_dbuf;
-    int_t  *perm_r, *perm_c; /* row and column permutation vectors */
+    int  *perm_r, *perm_c; /* row and column permutation vectors */
     int_t  *send_ibuf, *recv_ibuf;
     doublecomplex *send_dbuf, *recv_dbuf;
     int_t  *xsup, *supno;

SRC/complex16/pzgstrs3d.c

@@ -6288,7 +6288,7 @@ pzReDistribute3d_B_to_X (doublecomplex *B, int_t m_loc, int nrhs, int_t ldb,
     int *SendCnt, *SendCnt_nrhs, *RecvCnt, *RecvCnt_nrhs;
     int *sdispls, *sdispls_nrhs, *rdispls, *rdispls_nrhs;
     int *ptr_to_ibuf, *ptr_to_dbuf;
-    int_t *perm_r, *perm_c;     /* row and column permutation vectors */
+    int *perm_r, *perm_c;     /* row and column permutation vectors */
     int_t *send_ibuf, *recv_ibuf;
     doublecomplex *send_dbuf, *recv_dbuf;
     int_t *xsup, *supno;

SRC/complex16/pzsymbfact_distdata.c

@@ -783,8 +783,8 @@ zdist_A(SuperMatrix *A, zScalePermstruct_t *ScalePermstruct,
 {
   int    iam, p, procs;
   NRformat_loc *Astore;
-  int_t  *perm_r; /* row permutation vector */
-  int_t  *perm_c; /* column permutation vector */
+  int  *perm_r; /* row permutation vector */
+  int  *perm_c; /* column permutation vector */
   int_t  i, it, irow, fst_row, j, jcol, k, gbi, gbj, n, m_loc, jsize, isize;
   int_t  nsupers, nsupers_i, nsupers_j;
   int_t  nnz_loc, nnz_loc_ainf, nnz_loc_asup;    /* number of local nonzeros */

SRC/complex16/pzutil.c

@@ -297,7 +297,7 @@ int pzPermute_Dense_Matrix
  int_t fst_row,
  int_t m_loc,
  int_t row_to_proc[],
- int_t perm[],
+ int perm[],
  doublecomplex X[], int ldx,
  doublecomplex B[], int ldb,
  int nrhs,
@@ -649,7 +649,7 @@ if (get_acc_solve()){
  */
 int_t
 pzgstrs_init(int_t n, int_t m_loc, int_t nrhs, int_t fst_row,
-	     int_t perm_r[], int_t perm_c[], gridinfo_t *grid,
+	     int perm_r[], int perm_c[], gridinfo_t *grid,
 	     Glu_persist_t *Glu_persist, zSOLVEstruct_t *SOLVEstruct)
 {
 
@@ -1108,11 +1108,12 @@ pzgstrs_delete_device_lsum_x(zSOLVEstruct_t *SOLVEstruct)
 /*! \brief Initialize the data structure for the solution phase.
  */
 int zSolveInit(superlu_dist_options_t *options, SuperMatrix *A,
-	       int_t perm_r[], int_t perm_c[], int_t nrhs,
+	       int perm_r[], int perm_c[], int_t nrhs,
 	       zLUstruct_t *LUstruct, gridinfo_t *grid,
 	       zSOLVEstruct_t *SOLVEstruct)
 {
-    int_t *row_to_proc, *inv_perm_c, *itemp;
+    int_t *row_to_proc, *itemp;
+    int *inv_perm_c;
     NRformat_loc *Astore;
     int_t        i, fst_row, m_loc, p;
     int          procs;
@@ -1125,7 +1126,7 @@ int zSolveInit(superlu_dist_options_t *options, SuperMatrix *A,
     if ( !(row_to_proc = intMalloc_dist(A->nrow)) )
 	ABORT("Malloc fails for row_to_proc[]");
     SOLVEstruct->row_to_proc = row_to_proc;
-    if ( !(inv_perm_c = intMalloc_dist(A->ncol)) )
+    if ( !(inv_perm_c = int32Malloc_dist(A->ncol)) )
         ABORT("Malloc fails for inv_perm_c[].");
     for (i = 0; i < A->ncol; ++i) inv_perm_c[perm_c[i]] = i;
     SOLVEstruct->inv_perm_c = inv_perm_c;
@@ -1149,7 +1150,7 @@ int zSolveInit(superlu_dist_options_t *options, SuperMatrix *A,
     if ( !grid->iam ) {
       printf("fst_row = %d\n", fst_row);
       PrintInt10("row_to_proc", A->nrow, row_to_proc);
-      PrintInt10("inv_perm_c", A->ncol, inv_perm_c);
+      PrintInt32("inv_perm_c", A->ncol, inv_perm_c);
     }
 #endif
     SUPERLU_FREE(itemp);

SRC/complex16/z3DPartition.c

@@ -189,7 +189,7 @@ void zbcastPermutedSparseA(SuperMatrix *A,
                           zLUstruct_t *LUstruct, gridinfo3d_t *grid3d)
 {
     int_t m = A->nrow;
-    int_t n = A->ncol;
+	int_t n = A->ncol;
     Glu_persist_t *Glu_persist = LUstruct->Glu_persist;
     zLocalLU_t *Llu = LUstruct->Llu;
     NRformat_loc *Astore   = (NRformat_loc *) A->Store;
@@ -204,12 +204,12 @@ void zbcastPermutedSparseA(SuperMatrix *A,
 
     /* broadcast etree */
     int_t *etree = LUstruct->etree;
-    if(etree)
-        MPI_Bcast( etree, n, mpi_int_t, 0,  grid3d->zscp.comm);
+    if (etree)
+       MPI_Bcast( etree, n, mpi_int_t, 0,  grid3d->zscp.comm);
 
 
-    // list of all the arrays to be broadcasted:
-    //   A, ScalePermstruct, Glu_freeable, LUstruct
+    // list of all the arrays to be broadcast:
+    // A, ScalePermstruct, Glu_freeable, LUstruct
     int_t nsupers;
 
     if (!grid3d->zscp.Iam)
@@ -241,20 +241,20 @@ void zbcastPermutedSparseA(SuperMatrix *A,
 /***** YL: remove the allocation in the following as perm_r/perm_c has been allocated on all grids by dScalePermstructInit
 */
 #if 1
-    MPI_Bcast(ScalePermstruct->perm_r, m*sizeof(int_t), MPI_BYTE, 0, grid3d->zscp.comm);
-    MPI_Bcast(ScalePermstruct->perm_c, n*sizeof(int_t), MPI_BYTE, 0, grid3d->zscp.comm);
+    MPI_Bcast(ScalePermstruct->perm_r, m*sizeof(int), MPI_BYTE, 0, grid3d->zscp.comm);
+    MPI_Bcast(ScalePermstruct->perm_c, n*sizeof(int), MPI_BYTE, 0, grid3d->zscp.comm);
 #else
     allocBcastArray ( &(ScalePermstruct->perm_r), m*sizeof(int_t),
         0, grid3d->zscp.comm);
     allocBcastArray ( &(ScalePermstruct->perm_c), n*sizeof(int_t),
         0, grid3d->zscp.comm);
 #endif
     if(ScalePermstruct->DiagScale==ROW || ScalePermstruct->DiagScale==BOTH)
-    allocBcastArray ( (void **) &(ScalePermstruct->R), m*sizeof(double),
-        0, grid3d->zscp.comm);
+        allocBcastArray ( (void **) &(ScalePermstruct->R), m*sizeof(double),
+                          0, grid3d->zscp.comm);
     if(ScalePermstruct->DiagScale==COL || ScalePermstruct->DiagScale==BOTH)
-    allocBcastArray ( (void **) &(ScalePermstruct->C), n*sizeof(double),
-        0, grid3d->zscp.comm);
+        allocBcastArray ( (void **) &(ScalePermstruct->C), n*sizeof(double),
+                          0, grid3d->zscp.comm);
 
 
     /* ==== Broadcasting Glu_freeable ======= */

SRC/complex16/zldperm_dist.c

@@ -22,7 +22,7 @@ at the top-level directory.
 #include "superlu_zdefs.h"
 
 extern int mc64ad_dist(int *job, int *n, int_t *ne, int_t *ip,
-       int_t *irn, double *a, int *num, int_t *cperm,
+       int_t *irn, double *a, int *num, int *cperm,
        int_t *liw, int_t *iw, int_t *ldw, double *dw,
        int * icntl, int *info);
 
@@ -93,10 +93,10 @@ extern int mc64ad_dist(int *job, int *n, int_t *ne, int_t *ip,
 
 int
 zldperm_dist(int job, int n, int_t nnz, int_t colptr[], int_t adjncy[],
-	doublecomplex nzval[], int_t *perm, double u[], double v[])
+	doublecomplex nzval[], int *perm, double u[], double v[])
 {
     int i;
-    int num, icntl[10], info[10]; 
+    int num, icntl[10], info[10];
     int_t liw, ldw;
     int_t *iw;
     double *dw;
@@ -147,12 +147,12 @@ zldperm_dist(int job, int n, int_t nnz, int_t colptr[], int_t adjncy[],
 	        &liw, iw, &ldw, dw, icntl, info);
 
 #if ( DEBUGlevel>=2 )
-    PrintInt10("perm", n, perm);
+    PrintInt32("perm", n, perm);
     printf(".. After MC64AD info %d\tsize of matching %d\n", info[0], num);
 #endif
     if ( info[0] == 1 ) { /* Structurally singular */
         printf(".. The last %d permutations:\n", n-num);
-	PrintInt10("perm", n-num, &perm[num]);
+	PrintInt32("perm", n-num, &perm[num]);
     }
 
     /* Restore to 0-based indexing. */