ADR-0003/0014: generalize "router mesh" to "NOC"

NOC topology is an implementation choice (mesh, ring, crossbar, etc.).
ADR-0017 covers the current 2D mesh choice; ADRs at the system-level
shouldn't bind to that specific implementation.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

docs/adr/ADR-0003-target-system-hierarchy.md

@@ -35,11 +35,13 @@ We model the system hierarchy explicitly:
 
 - A CUBE contains:
   - HBM + memory controller (HBM_CTRL)
-  - NOC router mesh: 2D grid of explicit routers (from cube_mesh.yaml) with XY routing;
+  - NOC (on-die fabric): carries all intra-cube traffic including HBM data,
-    carries all intra-cube traffic including HBM data, inter-cube (UCIe),
+    inter-cube (UCIe), command (M_CPU↔PE_CPU), and shared SRAM access.
-    command (M_CPU↔PE_CPU), and shared SRAM access.
+    Must provide: full-BW PE↔local HBM path, PE↔SRAM connectivity,
-    HBM_CTRL is attached to PE routers (local HBM = 0 hop).
+    PE↔UCIe connectivity, M_CPU↔PE command path.
-    See ADR-0017 and ADR-0019 for full architecture.
+    NOC topology is an implementation choice (e.g., 2D mesh, ring, crossbar);
+    current implementation uses a 2D mesh with XY routing (see ADR-0017).
+    HBM_CTRL is attached to each PE's local NOC port (local HBM = minimal hop).
   - Shared SRAM: cube-level shared memory accessible by all PEs via NOC
   - management/control CPU (M_CPU) coordinating PE command distribution and completion aggregation
   - multiple PEs

docs/adr/ADR-0014-pe-internal-execution-model.md

@@ -44,15 +44,15 @@ Each PE contains the following logical components.
 **PE_DMA**
 
 - Handles memory transfers between PE_TCM and external memory domains.
-- PE_DMA connects to the NOC router mesh at the CUBE level (ADR-0019):
+- PE_DMA connects to the cube-level NOC (on-die fabric):
-  - All destinations (HBM, shared SRAM, inter-cube UCIe) are reached via the router mesh
+  - All destinations (HBM, shared SRAM, inter-cube UCIe) are reached via the NOC
-  - Local HBM access: PE_DMA → local router → hbm_ctrl (switching overhead only)
+  - Local HBM access: PE_DMA → NOC → hbm_ctrl (minimal hop)
-  - Remote/shared: PE_DMA → local router → (mesh hops) → destination
+  - Remote/shared: PE_DMA → NOC → (fabric hops) → destination
 - Supported directions include:
-  - HBM → PE_TCM (via router mesh)
+  - HBM → PE_TCM (via NOC)
-  - PE_TCM → HBM (via router mesh)
+  - PE_TCM → HBM (via NOC)
-  - PE_TCM → shared SRAM (via router mesh)
+  - PE_TCM → shared SRAM (via NOC)
-  - PE_TCM → other memory domains (via router mesh, if supported by topology)
+  - PE_TCM → other memory domains (via NOC, if supported by topology)
 
 **PE_GEMM**
 
@@ -252,7 +252,7 @@ Compute operations use a TCM-centric dataflow model.
 **Input path (HBM)**
 
 ```text
-HBM → router mesh → PE_DMA (DMA_READ) → PE_TCM
+HBM → NOC → PE_DMA (DMA_READ) → PE_TCM
 ```
 
 **Input path (shared SRAM)**
@@ -269,14 +269,14 @@ Compute engines read input tensors from PE_TCM.
 PE_TCM → GEMM / MATH
 ```
 
-Weights for GEMM may optionally stream directly from HBM (via router mesh).
+Weights for GEMM may optionally stream directly from HBM (via NOC).
 
 **Output path (HBM)**
 
 Compute results are written to PE_TCM, then DMA writes to HBM.
 
 ```text
-PE_TCM → PE_DMA (DMA_WRITE) → router mesh → HBM
+PE_TCM → PE_DMA (DMA_WRITE) → NOC → HBM
 ```
 
 **Output path (shared SRAM)**
@@ -348,9 +348,9 @@ PE instances are derived from `cube.pe_layout`.
 
 External connectivity such as:
 
-- PE_DMA → router mesh → HBM (data path, ADR-0019)
+- PE_DMA → NOC → HBM (data path)
-- PE_DMA → router mesh → shared SRAM, inter-cube UCIe (non-HBM data path)
+- PE_DMA → NOC → shared SRAM, inter-cube UCIe (non-HBM data path)
-- router mesh → PE_CPU (command path from M_CPU)
+- NOC → PE_CPU (command path from M_CPU)
 
 is modeled at the CUBE level (see ADR-0003 D3).