|
| subroutine, public | prg_partition_mod::prg_metispartition (gp, ngroups, nnodes, xadj, adjncy, nparts, part, core_count, CH_count, Halo_count, sumCubes, maxCH, smooth_maxCH, pnorm) |
| | Create graph partitions minizing number of cut edges. More...
|
| |
| subroutine, public | prg_partition_mod::prg_costpartition (gp, xadj, adjncy, partNumber, core_count, CH_count, Halo_count, sumCubes, maxCH, smooth_maxCH, pnorm) |
| | Compute cost of a partition. More...
|
| |
| subroutine, public | prg_partition_mod::update_prg_costpartition (gp, xadj, adjncy, partNumber, core_count, CH_count, Halo_count, sumCubes, maxCH, smooth_maxCH, pnorm, node, new_part) |
| | Update cost of partition and the different parameters node is moves into new_part For each neighbor of node, the following cases hold: Case 1: neighbor is in old_part Case 2: neighbor is in new_part Case 3: neighbor is neither in old_ or new_part. More...
|
| |
| subroutine | prg_partition_mod::prg_accept_prob (it, prg_delta, r) |
| | Compute acceptance probability for simulated annealing. More...
|
| |
| subroutine | prg_partition_mod::prg_costindex (cost, sumCubes, maxCH, smooth_maxCH, obj_fun) |
| | Choose objective function to work with. More...
|
| |
| subroutine | prg_partition_mod::prg_rand_node (gp, node, seed) |
| | Pick a random node. More...
|
| |
| subroutine, public | prg_partition_mod::prg_simannealing (gp, xadj, adjncy, partNumber, core_count, CH_count, Halo_count, sumCubes, maxCH, smooth_maxCH, pnorm, niter, seed) |
| | Graph partitioning based on Simulated Annealing. More...
|
| |
| subroutine, public | prg_partition_mod::prg_kernlin (gp, xadj, adjncy, partNumber, core_count, CH_count, Halo_count, sumCubes, maxCH, smooth_maxCH, pnorm, nconverg, seed) |
| | Graph partitioning based on inspired by Kernighan-Lin Review METiS manual for description of k-way implementation of KL Pick a core together with its halos Place free vertices on a priority queue with (key, value) =(prg_delta, best_part),with prg_delta = change in obj_value Dequeue and allow hill climbing. More...
|
| |
| subroutine, public | prg_partition_mod::prg_update_gp (gp, partNumber, core_count) |
| |
| subroutine | prg_partition_mod::prg_rand_shuffle (array, seed) |
| | Randomly shuffle array. More...
|
| |
| subroutine, public | prg_partition_mod::prg_check_arrays (gp, core_count, CH_count, Halo_count) |
| | Error checking Checking that core_count, CH_count, Halo_count match. More...
|
| |
| subroutine, public | prg_partition_mod::prg_kernlin_queue (gp, xadj, adjncy, partNumber, core_count, CH_count, Halo_count, sumCubes, maxCH, smooth_maxCH, pnorm) |
| | Greedy algorithm. At each step it chooses the (vertex, new_part) pair with highest gain Currently implementation is very slow. More...
|
| |
| subroutine | prg_partition_mod::prg_find_best_move (gp, xadj, adjncy, partNumber, core_count, CH_count, Halo_count, sumCubes, maxCH, smooth_maxCH, pnorm, best_node, best_part) |
| | For kerlin_queue to find (vertex, new_part) pair with highest gain. More...
|
| |
| subroutine, public | prg_partition_mod::prg_kernlin2 (gp, xadj, adjncy, partNumber, core_count, CH_count, Halo_count, sumCubes, maxCH, smooth_maxCH, pnorm) |
| |
| subroutine | prg_partition_mod::prg_get_largest_hedge_in_part (gp, xadj, adjncy, partNumber, core_count, CH_count, Halo_count, sumCubes, maxCH, smooth_maxCH, pnorm, search_part, largest_Hedge) |
| |
| subroutine, public | prg_partition_mod::prg_simannealing_old (gp, xadj, adjncy, partNumber, core_count, CH_count, Halo_count, sumCubes, maxCH, smooth_maxCH, pnorm, niter, seed) |
| |
1.9.1