📄 sparse.c
字号:
/* * Revision Control Information * * $Source: /projects/mvsis/Repository/mvsis-1.3/src/sis/espresso/sparse.c,v $ * $Author: wjiang $ * $Revision: 1.1.1.1 $ * $Date: 2003/02/24 22:24:08 $ * *//* module: sparse.c make_sparse is a last-step cleanup to reduce the total number of literals in the cover. This is done by reducing the "sparse" variables (using a modified version of irredundant rather than reduce), followed by expanding the "dense" variables (using modified version of expand).*/#include "espresso.h"pcover make_sparse(F, D, R)pcover F, D, R;{ cost_t cost, best_cost; cover_cost(F, &best_cost); do { EXECUTE(F = mv_reduce(F, D), MV_REDUCE_TIME, F, cost); if (cost.total == best_cost.total) break; copy_cost(&cost, &best_cost); EXECUTE(F = expand(F, R, TRUE), RAISE_IN_TIME, F, cost); if (cost.total == best_cost.total) break; copy_cost(&cost, &best_cost); } while (force_irredundant); return F;}/* mv_reduce -- perform an "optimal" reduction of the variables which we desire to be sparse This could be done using "reduce" and then saving just the desired part of the reduction. Instead, this version uses IRRED to find which cubes of an output are redundant. Note that this gets around the cube-ordering problem. In normal use, it is expected that the cover is irredundant and hence no cubes will be reduced to the empty cube (however, this is checked for and such cubes will be deleted)*/pcovermv_reduce(F, D)pcover F, D;{ register int i, var; register pcube p, p1, last; int index; pcover F1, D1; pcube *F_cube_table; /* loop for each multiple-valued variable */ for(var = 0; var < cube.num_vars; var++) { if (cube.sparse[var]) { /* loop for each part of the variable */ for(i = cube.first_part[var]; i <= cube.last_part[var]; i++) { /* remember mapping of F1 cubes back to F cubes */ F_cube_table = ALLOC(pcube, F->count); /* 'cofactor' against part #i of variable #var */ F1 = new_cover(F->count); foreach_set(F, last, p) { if (is_in_set(p, i)) { F_cube_table[F1->count] = p; p1 = GETSET(F1, F1->count++); (void) set_diff(p1, p, cube.var_mask[var]); set_insert(p1, i); } } /* 'cofactor' against part #i of variable #var */ /* not really necessary -- just more efficient ? */ D1 = new_cover(D->count); foreach_set(D, last, p) { if (is_in_set(p, i)) { p1 = GETSET(D1, D1->count++); (void) set_diff(p1, p, cube.var_mask[var]); set_insert(p1, i); } } mark_irredundant(F1, D1); /* now remove part i from cubes which are redundant */ index = 0; foreach_set(F1, last, p1) { if (! TESTP(p1, ACTIVE)) { p = F_cube_table[index]; /* don't reduce a variable which is full * (unless it is the output variable) */ /* if (var == cube.num_vars-1 || ! setp_implies(cube.var_mask[var], p)) { */ /* reduce anyways (changed by WJIANG 1/9/01) */ if (! setp_implies(cube.var_mask[var], p)) { set_remove(p, i); } RESET(p, PRIME); } index++; } free_cover(F1); free_cover(D1); FREE(F_cube_table); } } } /* Check if any cubes disappeared */ (void) sf_active(F); for(var = 0; var < cube.num_vars; var++) { if (cube.sparse[var]) { foreach_active_set(F, last, p) { if (setp_disjoint(p, cube.var_mask[var])) { RESET(p, ACTIVE); F->active_count--; } } } } if (F->count != F->active_count) { F = sf_inactive(F); } return F;}
⌨️ 快捷键说明
复制代码
Ctrl + C
搜索代码
Ctrl + F
全屏模式
F11
切换主题
Ctrl + Shift + D
显示快捷键
?
增大字号
Ctrl + =
减小字号
Ctrl + -