KinGeorge/Dr.Sparse-RL-level4
Dr.Sparse RL rollout pool, level4_huge members The SpGEMM online-RL rollout pool is 130 SuiteSparse matrices: the 118 level1-3 matrices in the Dr.Sparse git repo (level_sparse/level{1_small,2_medium,3_large}) plus the 12 level4 matrices in this dataset. It is disjoint from the 100-matrix OTF test set. The 12 level4 (>= 1.1M rows) matrices, in Dr.Sparse, in the repo's raw .bin layout ([int32 rows, cols, nnz][int32 row_ptr][int32 col_ind][float32 values][float32 x], read by… See the full description on the dataset page: https://huggingface.co/datasets/KinGeorge/Dr.Sparse-RL-level4.
Dr.Sparse RL rollout pool, level4_huge members
The SpGEMM online-RL rollout pool is 130 SuiteSparse matrices: the 118 level1-3 matrices in the Dr.Sparse git repo (level_sparse/level{1_small,2_medium,3_large}) plus the 12 level4 matrices in this dataset. It is disjoint from the 100-matrix OTF test set.
The 12 level4 (>= 1.1M rows) matrices, in Dr.Sparse, in the repo's raw .bin layout ([int32 rows, cols, nnz][int32 row_ptr][int32 col_ind][float32 values][float32 x], read by level_sparse/data_loader.h). The level1-3 members (118 matrices) ship in the git repo; the 100-matrix OTF test set is KinGeorge/Dr.Sparse-OTF-test-set.
Stage into a checkout with:
python level_sparse/stage_rl_pool.py --download # -> level_sparse/level4_huge/level4_list.txt is the member list; spgemm_excluded.txt records the ten level4 matrices that were dropped because cuSPARSE cannot produce an SpGEMM baseline for them on one GPU. Values are the real matrix values (float32); x is a fixed standard-normal vector baked in at build time.