-
baseline: Time = 1.32774 ms. 带宽 = 96.404126 GB/s.
-
no_divergence_branch: Time = 1.11318 ms. 带宽 = 114.985486 GB/s.
-
no_bank_conflict: Time = 0.8944 ms. 带宽 = 143.112701 GB/s.
-
add_during_load: Time = 0.487872 ms. 带宽 = 262.363896 GB/s.
-
unroll_last_warp: Time = 0.345984 ms. 带宽 = 369.959292 GB/s.
-
completely_unroll: Time = 0.335776 ms. 带宽 = 381.206517 GB/s.
-
multi_add: Time = 0.275168 ms. 带宽 = 465.170366 GB/s.
daydream0929 / how_to_optimize_reduce Goto Github PK
View Code? Open in Web Editor NEWLicense: MIT License