Commit Graph

27 Commits

Author SHA1 Message Date
tangxifan 148da80869 [Tool] Add new syntax about physical_pb_port_rotate_offset to support fracturable heterogeneous block mapping between operating modes and physical modes 2021-04-24 14:53:29 -06:00
tangxifan 0709e5bb81 [Tool] Fixed a bug in the routing trace finder for direct connections inside repacker 2021-04-24 13:27:44 -06:00
tangxifan 96ce6b545f [Tool] Patch repack to consider design constraints for pins that are not equivalent 2021-04-21 13:53:08 -06:00
tangxifan df7b436ac7 [Tool] Patch repacker to support duplicated nets due to adder nets 2021-02-23 19:01:18 -07:00
tangxifan a5b8b2a64a [Tool] Use dedicated function to identify wire LUT created by repacker 2021-02-18 19:37:44 -07:00
tangxifan aae03482f5 [Tool] Bug fix for wire LUT identification by repacker. Create a dedicated function to identify these LUTs and store the results in shared database 2021-02-18 19:37:17 -07:00
tangxifan 0c409b5bcc [Tool] Add bitstream annotation support 2021-02-01 20:49:36 -07:00
tangxifan 8c311b8282 [Tool] Bug fix in repacker for considering design constraints 2021-01-17 12:26:14 -07:00
tangxifan 2efe513122 [Tool] Now repack consider design constraints; test pending 2021-01-16 21:57:17 -07:00
tangxifan bb8e7e25c2 [Tool] Start deploying design constraints in repack engine 2021-01-16 21:27:12 -07:00
tangxifan 26f1a5d9ec [OpenFPGA Tool] Bug fix for repacking no local routing architecture 2020-09-21 22:22:03 -06:00
tangxifan 9cfb2f52ef [OpenFPGA code] bug fix for fully equivalent outputs of pb_type 2020-09-16 19:26:46 -06:00
tangxifan e10cafe0a5 Critical patch on repacking about wire LUT support.
Previously, the wire LUT identification is too naive and does not consider all the cases
2020-04-19 16:42:31 -06:00
tangxifan 2e3a811f4f critical bug fixed in repacking. This is due to depop50% local routing where the same net may be mapped to two different pins in the same pb_graph_pin. Now we restrict the pin searching. But in long term, we should sync the pb_route results to post routing results 2020-04-18 21:04:46 -06:00
tangxifan 808853db0b critical bug fixed for find proper pb_route traceback 2020-03-13 12:26:37 -06:00
tangxifan 8921905bec annotate multiple-source and multiple-sink nets from pb to lb router 2020-03-12 19:21:13 -06:00
tangxifan 410dcf6ab6 debugged LUT bitstream 2020-02-26 11:42:18 -07:00
tangxifan 9583731531 add results saver for lb router 2020-02-22 22:10:32 -07:00
tangxifan 926e429374 add save repacking results in physical pb 2020-02-21 20:39:49 -07:00
tangxifan 12f2888c7c add physical pb data structure and basic allocator 2020-02-21 17:47:27 -07:00
tangxifan b035b4c87f debugged with Lbrouter. Next step is to output routing traces to physical pb data structure 2020-02-21 12:16:50 -07:00
tangxifan 1b66e837ba bug fixing for lb router. Add physical mode to default node expanding settings 2020-02-21 11:29:00 -07:00
tangxifan 0b0e00b5f4 debugging the LbRouter 2020-02-20 21:56:15 -07:00
tangxifan 3e07d7d5e0 finish net addition to LbRouter. Found a bug in pb pin fix-up. Need to consider clustered I/O block z offset 2020-02-20 20:26:20 -07:00
tangxifan fdb27c5a6b move lb_rr_graph construction to repack command 2020-02-20 13:24:34 -07:00
tangxifan 409b3f6896 add lb_rr_graph builder for the refactored version 2020-02-17 21:11:56 -07:00
tangxifan 8e97443410 start working on repack 2020-02-17 17:57:43 -07:00