Issues

• viz: gate/transistor-level visualization (function-colored device view)

  npip99/autogpu#55 · Filter by author npip99 opened on Jun 7, 2026

• Add elementwise pipeline (add/sub/mul/relu, reductions, transcendentals)

  npip99/autogpu#53 · Filter by author npip99 opened on Jun 4, 2026

• Hardened macros' .lib clock-tree characterization doesn't match parent CTS — root cause of #28 mask

  npip99/autogpu#52 · Filter by author npip99 opened on Jun 3, 2026

• Chip-level clock distribution: avoid fanning chip clock to all macros via parent CTS

  npip99/autogpu#50 · Filter by author npip99 opened on Jun 3, 2026
  1

• Adopt a chip-wide PDN alignment standard across leaf macros

  npip99/autogpu#49 · Filter by author npip99 opened on Jun 3, 2026
  1

• Decouple chip_top from leaf macros: autogen stub LEF/LIB as the team-API

  npip99/autogpu#47 · Filter by author npip99 opened on Jun 3, 2026
  1

• compute_array: absorb parent BCAST_PIPE flops into cmd_unit OR delete BCAST_PIPE entirely

  npip99/autogpu#45 · Filter by author npip99 opened on Jun 2, 2026

• skew_lane: dead 31-stage shift register at tap_index=0

  npip99/autogpu#44 · Filter by author npip99 opened on Jun 2, 2026

• Manifest-driven build system for hardened-macro abutment designs

  enhancement
  npip99/autogpu#43 · Filter by author npip99 opened on Jun 2, 2026
  2

• Better chip-wide clock (low-skew mesh / multi-source CTS) — BLOCKED on OpenROAD upstream

  npip99/autogpu#42 · Filter by author npip99 opened on Jun 1, 2026

• Full 32×32 compute_array abutment: re-harden cmd_unit + skew_lanes as abutment-ready tiles

  npip99/autogpu#40 · Filter by author npip99 opened on Jun 1, 2026
  8

• Tile-and-array compute_array via abutment (kill the 1089-macro parent P&R monster)

  npip99/autogpu#32 · Filter by author npip99 opened on May 31, 2026
  1

• Full 32×32 compute_array does not converge in detailed routing (~17K DRC violations) — routability/congestion

  npip99/autogpu#30 · Filter by author npip99 opened on May 29, 2026
  1

• chip_top MMA=32 hardening: M5 PDN density + placement-guard visibility (from #26 review)

  npip99/autogpu#29 · Filter by author npip99 opened on May 29, 2026

• Antenna sign-off: re-route pass with overlay loaded so repair_antennas can insert diodes

  npip99/autogpu#23 · Filter by author npip99 opened on May 28, 2026

• chip_top cocotb test_directed_dispatch failing: master CDC sync added 2-cycle latency on instr_push, tests not updated

  npip99/autogpu#21 · Filter by author npip99 opened on May 28, 2026

• IO pad ring / chip-pad placement (PDK-swap-time blocker)

  npip99/autogpu#19 · Filter by author npip99 opened on May 27, 2026

• Density fill insertion methodology (PDK-swap-time blocker)

  npip99/autogpu#18 · Filter by author npip99 opened on May 27, 2026

• LEC re-enable on chip_top (kepler-formal scaling probe + fix)

  npip99/autogpu#17 · Filter by author npip99 opened on May 27, 2026

• Multi-corner timing (SS / FF) — current flow only runs best-case, setup margin untested

  npip99/autogpu#16 · Filter by author npip99 opened on May 27, 2026

• lvs.py: NetlistComparer doesn't terminate on bigger blocks (cmdproc, mac_array_small, store, tile_buf_8row)

  npip99/autogpu#15 · Filter by author npip99 opened on May 27, 2026

• chip_top gate-level cocotb + SDF annotation

  npip99/autogpu#14 · Filter by author npip99 opened on May 27, 2026

• DFT / scan-chain insertion methodology

  npip99/autogpu#13 · Filter by author npip99 opened on May 27, 2026

• chip_top chip-scale hold/setup timing measurement

  npip99/autogpu#12 · Filter by author npip99 opened on May 27, 2026

• chip_top sign-off sweep — run verify_macro_power + lvs + ir_drop + antenna once route completes

  npip99/autogpu#11 · Filter by author npip99 opened on May 27, 2026