菜单
Convert LLK lib/API tile-size args to ckernel::TensorShape and maintain TRISC TensorShape coverage. Use when adding TensorShape parameters, replacing face_r_dim/num_faces, editing LLK_VALIDATE_TENSOR_SHAPE_*, regenerating tensor_shape_coverage_*.h, or reviewing TensorShape PRs.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Create, extend, debug, and validate Quasar LLK performance tests and their PerfRunType kernel paths. Use when adding perf_[op]_quasar.py, wiring PerfConfig, implementing UNPACK_ISOLATE / MATH_ISOLATE / PACK_ISOLATE / L1_CONGESTION, or investigating implausible Quasar perf metrics and dvalid handshakes.
Create, extend, debug, and validate Quasar LLK performance tests and their PerfRunType kernel paths. Use when adding perf_[op]_quasar.py, wiring PerfConfig, implementing UNPACK_ISOLATE / MATH_ISOLATE / PACK_ISOLATE / L1_CONGESTION, or investigating implausible Quasar perf metrics and dvalid handshakes.
Route gh-aw workflow design/create/debug/upgrade requests to the right prompts.
Audit LLK code for races on the backend CONFIG register file where differently-named fields share the SAME 32-bit config word — both cross-thread (unpack/math/pack write the same word) and intra-thread (a full-word write clobbers a sibling field the same thread set elsewhere). Use after adding/changing any ALU_FORMAT_SPEC / ALU_ACC_CTRL / ALU_ROUNDING_MODE / STACC_RELU / THCON_SEC* write, any WRCFG_32b/cfg[]= full-word write to a multi-field word, or any cfg_reg_rmw_tensix on a word another thread also touches.
Audit hand-written Tensix/SFPU instruction sequences for missing pipeline-latency padding — where a dependent instruction consumes a multi-cycle-latency result before it is ready and a NOP (or independent-instruction spacing) is required. Use after touching any raw TTI_SFP*/TTI_* sequence, ckernel_sfpu_* kernels, or hand-assembled instruction streams. NOT a cross-thread race — an intra-thread micro-architectural hazard.
Audit cross-core NoC synchronization in dataflow kernels — noc_semaphore_wait/set/inc balance and direction, multicast fan-out counts, and data-before-signal NoC ordering (noc_async_write_barrier / noc_async_writes_flushed before a remote credit). The half of dataflow that dataflow-cb-sync-audit (CB credits) does not reach. Use after touching reader/writer kernels, noc_semaphore_*, noc_async_*_barrier, or any cross-core handshake not expressed as a CB.
| name | tensor-shape |
| description | Convert LLK lib/API tile-size args to ckernel::TensorShape and maintain TRISC TensorShape coverage. Use when adding TensorShape parameters, replacing face_r_dim/num_faces, editing LLK_VALIDATE_TENSOR_SHAPE_*, regenerating tensor_shape_coverage_*.h, or reviewing TensorShape PRs. |
| argument-hint | [convert|coverage|review] [op-or-path] |
| user_invocable | true |
/tensor-shape convert unpack_A
/tensor-shape coverage harvest unpack_A
/tensor-shape review
Read the canonical reference and follow it:
Read(".claude/references/tensor-shape.md")
For metal-layer propagation after signature changes, also read .claude/references/metal-integration.md.
Plumb TensorShape through an LLK op (lib → metal API → tests/fuser):
face_r_dim / num_faces (and similar) with ckernel::TensorShape.tensor_shape.h + the matching tensor_shape_coverage_{unpack|math|pack}.h.LLK_VALIDATE_TENSOR_SHAPE_{UNPACK|MATH|PACK}("_llk_foo_...", tensor_shape) — string literal name; no API registry edit in coverage.h.num_faces_r_dim / num_faces_c_dim for new logic; keep flat total_num_faces() only when intentionally preserving old mop behavior.get_operand_tensor_shape(operand_id).make_tensor_shape / cpp_tensor_shape; never legacy helper for 32×16.Discover or regenerate TRISC allowlists:
TT_LLK_DISABLE_ASSERTS=1 + DPRINT, then:
helpers/tensor_shape_coverage_parser.py harvest|emit|summaryTENSOR_SHAPE_FR*_NF* in tensor_shape_coverage.h if needed + TRISC predicate update via emit.fn_name → add to MATH_FUNCTIONS / UNPACK_FUNCTIONS in the parser only.Apply the Review Checklist in .claude/references/tensor-shape.md. Especially correct the misconception that every new API must be listed in coverage.h — coverage is TRISC-scoped; call sites only need the validate macro + string literal.
| Role | Path |
|---|---|
| Struct / helpers | common/tensor_shape.h |
| Macros / named shapes | common/tensor_shape_coverage.h |
| TRISC tables | common/tensor_shape_coverage_{math,unpack,pack}.h |
| Parser | tests/python_tests/helpers/tensor_shape_coverage_parser.py |
| Python TileShape | tests/python_tests/helpers/tile_shape.py |
| Metal operand helper | tt_metal/hw/ckernels/{arch}/metal/llk_io/llk_operands.h |