Fix all remaining test failures: single-cube allreduce + matplotlib dep

- intercube_allreduce: add single-cube fast path that skips intra-SIP
  mesh reduce and goes directly to inter-SIP exchange. Fixes IPCQ
  deadlock when TP launches kernel on one cube per SIP.
- distributed.py: derive effective cube dims from tensor shard placement
  instead of hardcoding topology mesh size.
- pyproject.toml: add matplotlib>=3.7 to dependencies.
- pe_dma.py (prior commit): add MMU translation in pipeline DMA path.

577 passed, 0 failed (was 529 passed, 10 failed).

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

src/kernbench/runtime_api/distributed.py

@@ -113,7 +113,18 @@ class AhbmCCLBackend:
             )
         n_elem = shards[0].nbytes // tensor.itemsize
         kernel_fn = self._algo_module.kernel
-        kernel_args = self._algo_module.kernel_args(self._world_size, n_elem)
+        # Derive effective cube dims from tensor's actual shard placement
+        # (may differ from topology mesh when TP uses fewer cubes).
+        sip0_cubes = sorted({s.cube for s in shards if s.sip == shards[0].sip})
+        eff_n_cubes = len(sip0_cubes) if sip0_cubes else 1
+        if eff_n_cubes == 1:
+            eff_cube_w, eff_cube_h = 1, 1
+        else:
+            eff_cube_w, eff_cube_h = self._cube_w, self._cube_h
+        kernel_args = self._algo_module.kernel_args(
+            self._world_size, n_elem,
+            cube_w=eff_cube_w, cube_h=eff_cube_h,
+        )
 
         # Resolve sip_rank from the current greenlet's bound rank
         from greenlet import getcurrent as _gc