Add PE-level IPCQ collective infra + unified ccl_allreduce bench (ADR-0023)

Major changes:

PE-level IPCQ infrastructure:
- New PE_IPCQ component: ring-buffer control plane with 4-direction
  neighbor mapping, head/tail pointers, backpressure (poll/sleep).
- PE_DMA extended with vc_comm channel for IPCQ outbound/inbound DMA,
  including in-flight data snapshot (D9) and op_log recording at
  outbound time for Phase 2 replay correctness.
- IpcqDmaToken piggyback model: data + metadata travel together,
  atomic visibility at receiver (invariant I6).
- Credit return fast path: bottleneck-BW latency, no fabric vc_comm.

Phase 2 data execution (ADR-0020 integration):
- op_log extended: DmaWriteCmd now captures src_space/src_addr for
  Phase 2 dma_write copy; ipcq_copy ops recorded at outbound time.
- DataExecutor replays dma_write + ipcq_copy in t_start order.
- Engine._flush_data_phase: incremental cursor-based replay after
  each engine.wait() so host reads see post-Phase-2 data.
- KernelRunner Phase 1 writes disabled when op_log is active to
  prevent stale data from corrupting the MemoryStore snapshot.

TLContext / kernel API:
- tl.send(dir, src=TensorHandle), tl.recv(dir, shape, dtype),
  tl.recv_async, tl.wait(RecvFuture), copy_to_dst mode.
- TensorHandle operator overloading (add/sub/mul/div) via thread-local
  active TLContext → MathCmd dispatch through PE_MATH.
- PE-local scratch allocator for math output handles.
- tl.load returns space="hbm" handles for correct Phase 2 addressing.
- Additional math functions: maximum, minimum, fma, clamp, softmax, cdiv.

Unified ccl_allreduce bench (PyTorch-compat host code):
- Single benches/ccl_allreduce.py with run() + worker(rank, ws, torch)
  split matching real PyTorch DDP worker pattern.
- torch.distributed facade: init_process_group, get_world_size,
  get_rank, get_backend, all_reduce, barrier — only real PyTorch names.
- AhbmCCLBackend: eager install_ipcq at init, all_reduce dispatches
  kernel via tensor shard metadata (n_elem from shards[0].nbytes).
- world_size derived from topology spec (sips × cubes × pes_per_cube)
  with optional algorithm-level override in ccl.yaml.

Tensor API (PyTorch-compat surface):
- Tensor.numpy(): gather-aware (all shards via VA-based addressing).
- Tensor.copy_(source): scatter from host tensor into sharded target.
- RuntimeContext.from_numpy(arr): host-side staging tensor.
- Tensor.data property fixed to use numpy() (was shards[0]-only).

Algorithm modules moved to src/kernbench/ccl/algorithms/:
- ring_allreduce, mesh_allreduce, tree_allreduce, hello_send.
- Each module exports kernel_args(world_size, n_elem) helper.
- ccl.yaml module paths updated to kernbench.ccl.algorithms.*.

Dead code removed:
- 7 per-variant bench files (ccl_allreduce_{tcm,hbm,sram}, etc.).
- _run_ccl_bench greenlet-per-SIP scheduler.
- benches.loader.is_ccl_bench + run_rank detection.
- benches/ccl/ directory.

Tests:
- New test_ccl_allreduce_matrix.py: 7 parametrized cases
  (ring×3 buffers, ring 8/16, mesh 4, tree 7).
- New test_runtime_api_tensor.py: copy_/numpy/from_numpy unit tests.
- Existing tests updated for new import paths + world_size_override.

Docs:
- Korean ccl-author-guide.md and ADR-0023 paths updated.
- New English versions: ccl-author-guide.en.md, ADR-0023.en.md.

502 tests pass.

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

tests/test_ccl_hello_world_guide.py

@@ -0,0 +1,81 @@
+"""Validate the hello-world example from docs/ccl-author-guide.md.
+
+This is the simplest possible CCL kernel — each PE sends its tile E
+and receives a tile from W. After running, each rank's slice should
+contain the data of the previous rank.
+"""
+from __future__ import annotations
+
+import numpy as np
+
+from kernbench.ccl.algorithms import hello_send
+from kernbench.ccl.testing import run_kernel_in_mock
+
+
+def test_hello_send_4_ranks_mock():
+    n_elem = 8
+    inputs = [np.full((n_elem,), float(r + 1), dtype=np.float16) for r in range(4)]
+
+    outputs = run_kernel_in_mock(
+        kernel_fn=hello_send.kernel,
+        world_size=4,
+        topology="ring_1d",
+        inputs=inputs,
+        kernel_args=(n_elem,),
+    )
+
+    # rank r should have rank (r-1) % 4's data
+    for r in range(4):
+        prev = inputs[(r - 1) % 4]
+        assert np.array_equal(outputs[r], prev), f"rank {r}: got {outputs[r]}"
+
+
+def test_hello_send_via_simpy_runner():
+    """Same but through real SimPy + IPCQ."""
+    from kernbench.policy.placement.dp import DPPolicy
+    from kernbench.runtime_api.bench_runner import run_bench
+    from kernbench.runtime_api.types import resolve_device
+    from kernbench.sim_engine.engine import GraphEngine
+    from kernbench.topology.builder import resolve_topology
+
+    topo = resolve_topology("topology.yaml")
+    n_elem = 8
+    world_size = 8
+
+    def run(torch):
+        # World size for this hello test is 8 (one cube). ccl.yaml no
+        # longer carries a default world_size — pass it explicitly.
+        plan = torch.install_ipcq(
+            algorithm="ring_allreduce_tcm", world_size_override=world_size,
+        )
+        a = torch.zeros(
+            (1, world_size * n_elem), dtype="f16",
+            dp=DPPolicy(
+                sip="replicate", cube="replicate", pe="column_wise",
+                num_sips=1, num_cubes=1,
+            ),
+            name="hello_in",
+        )
+        store = torch.engine.memory_store
+        base = a._handle.va_base or a._handle.shards[0].pa
+        nbytes = n_elem * 2
+        for r in range(world_size):
+            store.write("hbm", base + r * nbytes,
+                        np.full((n_elem,), float(r + 1), dtype=np.float16))
+
+        torch.launch("hello_send", hello_send.kernel, a, n_elem)
+
+        # Each rank should hold the previous rank's data after the round
+        for r in range(world_size):
+            arr = store.read("hbm", base + r * nbytes, shape=(n_elem,), dtype="f16")
+            prev_value = float(((r - 1) % world_size) + 1)
+            assert np.allclose(arr, prev_value), f"rank {r}: got {arr}, expected {prev_value}"
+
+    result = run_bench(
+        topology=topo, bench_fn=run,
+        device=resolve_device("all"),
+        engine_factory=lambda t, d: GraphEngine(
+            getattr(t, "topology_obj", t), enable_data=True
+        ),
+    )
+    assert result.completion.ok