ADR-0001 Rev 2: 51-bit PhysAddr layout with concrete sub-unit tables

Remove rack_id (4 bits), rename sip_seg→die_id, shift fields to enable
42-bit local_offset (4 TB per die). Define PE_LOCAL/MCPU_LOCAL/CUBE_SRAM
sub-unit tables for AHBM dies and IOCPU sub-unit table for IOCHIPLET
dies (1 TB window). Supersedes ADR-0031.

Also fixes latent VA/PA confusion in pe_dma pipeline DMA path where
virtual addresses were decoded as physical addresses without MMU
translation — previously masked by coincidental bit-position alignment.

529 passed (+6 recovered), 10 pre-existing failures unchanged.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

src/kernbench/sim_engine/event_log.py

@@ -212,7 +212,7 @@ def _generate_probe_h2d(graph, edge_map) -> list[dict]:
     t_offset = 0.0
     for rid, (name, cube, hops) in enumerate(cases):
         pa = PhysAddr.pe_hbm_addr(
-            rack_id=0, sip_id=0, cube_id=cube, pe_id=0,
+            sip_id=0, die_id=cube, pe_id=0,
             pe_local_hbm_offset=0x1000, slice_size_bytes=slice_bytes,
         )
         dst_node = resolver.resolve(pa)
@@ -256,7 +256,7 @@ def _generate_probe_d2h(graph, edge_map) -> list[dict]:
     t_offset = 0.0
     for rid, (name, cube, hops) in enumerate(cases):
         pa = PhysAddr.pe_hbm_addr(
-            rack_id=0, sip_id=0, cube_id=cube, pe_id=0,
+            sip_id=0, die_id=cube, pe_id=0,
             pe_local_hbm_offset=0x1000, slice_size_bytes=slice_bytes,
         )
         dst_node = resolver.resolve(pa)
@@ -310,7 +310,7 @@ def _generate_probe_pe_dma(graph, edge_map) -> list[dict]:
     t_offset = 0.0
     for rid, (name, sip, src_cube, src_pe, dst_cube, dst_pe) in enumerate(cases):
         pa = PhysAddr.pe_hbm_addr(
-            rack_id=0, sip_id=sip, cube_id=dst_cube, pe_id=dst_pe,
+            sip_id=sip, die_id=dst_cube, pe_id=dst_pe,
             pe_local_hbm_offset=0x1000, slice_size_bytes=slice_bytes,
         )
         dst_node = resolver.resolve(pa)