| 文件 | 最后提交记录 | 最后更新时间 |
|---|---|---|
Add the int8 x int8 -> int32 data type route to tla.mmad Co-authored-by: Alessio<alessio.russo@h-partners.com> # message auto-generated for no-merge-commit merge: !1120 merge mmad-int8 into master Add the int8 x int8 -> int32 data type route to tla.mmad Created-by: arusso Commit-by: Alessio Merged-by: cann-robot Description: ## 描述 tla.mmad previously routed only f16, bf16 and f32 operands, all accumulating into an fp32 L0C. This adds i8 x i8 -> i32, which the cube supports through the same non-MX mad intrinsic the float routes already use. The device wrapper is a single Catlass::Gemm::Mmad<int8_t, int8_t, int32_t> instantiation alongside the existing three, exposed as mmad_int8_int8_int32. TlaCubeRegionPass gains one more branch in the element-type routing to select it, and core_api.py adds ("i8", "i8", "i32") to the tla.mmad element-type contract. The bc layer registers the int8 operand copies the route needs: GM->L1 in zN and nZ, and L1->L0A / L1->L0B in both orientations. The one structural difference from every existing route is the accumulator: int8 accumulates into an **i32** L0C rather than fp32. Every copy out of L0C therefore had to stop assuming an fp32 accumulator. That contract is now expressed once and applied to all four L0C routes (GM, UB row-major, UB column-major, L1 zN): cpp // Whether fixpipe can carry this L0C element type out to the given destination // type. An fp32 accumulator (the float MMAD routes) may be narrowed to f16/bf16 // on the way out; an i32 accumulator (the int8 MMAD route) has no narrowing path // and must land as i32. static bool isLegalFixpipeElementType(StringRef srcElementType, StringRef dstElementType); Each of those exits is registered for int32_t -> int32_t, so the integer route has the same shape as the float ones: L0C -> GM, L0C -> UB (NO_SPLIT / SPLIT_M / SPLIT_N, both layouts) and L0C -> L1 (zN). The frontend enforces the matching rule for callers: an i32 accumulator may only be copied out as i32, while an fp32 one keeps its existing [f32, f16, bf16] choice. basic_matmul.py gains the route as a data-type choice: --dtype-a i8 --dtype-b i8 --dtype-c i32. The kernel takes an acc_is_int Constexpr and allocates the L0C tile as tla.Int32 when set; the host generates small integer operands, builds an exact int32 reference, and compares with element-wise equality instead of a tolerance. ## 关联的Issue N/A ## 原因 The DSL exposed only three of the operand formats the cube supports, which blocked int8-quantised workloads from being written in it. int8 is the cheapest of the missing formats to add: it needs no microscaling machinery, no sub-byte packing and no new element type, since i8 and i32 already exist in the DSL. CATLASS already selects an int32_t accumulator for this operand pair, so the device side is a single template instantiation. ## 测试 lit, in mmad-runtime-from-python.test, with both inputs emitted from legal @tla.kernel functions rather than hand-written IR: - an I8 case asserting the lowering reaches @mmad_int8_int8_int32 and that no tla.mmad survives; - an I8FIX case exercising every exit from the i32 accumulator in one kernel, asserting each copy resolves to an int32_t -> int32_t callee — copy_l0c_to_gm_row_major_int32_t, copy_l0c_to_ub_row_major_splitm_int32_t, copy_l0c_to_l1_zN_int32_t — and that no tla.mmad / tla.copy survives. This is what guards the shared fixpipe predicate: an i32 accumulator must never pick one of the fp32 callees. On-device, basic_matmul.py --dtype-a i8 --dtype-b i8 --dtype-c i32 runs the route end to end and checks the result with torch.equal against an exact int32 reference; integer MMAD is exact, so any tolerance would mask a defect. The example rejects a partial integer selection (--dtype-a i8 without i8/i32 on the other two), so the route cannot be entered half-configured. Full suites: pytest 1494 passed / 7 skipped, lit 124/124. ## 文档更新 docs/en/kernel_api_reference.md is regenerated, since core_api.py changed and the generated reference embeds source line numbers; the diff is limited to those mechanical Source: links. No new API surface is introduced, so there is no new entry: i8 operands with an i32 accumulator are a data-type choice on the existing tla.mmad, not a new op or parameter. The one thing a caller must know — that the integer route accumulates into an i32 L0C rather than fp32 — is visible in basic_matmul.py, which selects the L0C element type accordingly, and in the tla.mmad error message, which now names the supported element-type combinations including i8,i8 -> i32. ## 类型标签 - [ ] Bug修复 - [x] 新特性 - [ ] 性能优化 - [ ] 文档更新 - [ ] 其他，请描述： See merge request: cann/catlass!1120 | 2 个月前 | |
dsl: 补充迁移时收口的测试样例并纳入 hf32 相关测试 Co-authored-by: CheaterAbec<niyuhao2@huawei.com> Co-authored-by: cann-robot<cann@cann.team> # message auto-generated for no-merge-commit merge: !1143 merge restore/dsl-samples-tests-hf32 into master dsl: 补充迁移时收口的测试样例并纳入 hf32 相关测试 Created-by: CheaterAbec Commit-by: cann-robot;CheaterAbec Merged-by: cann-robot Description: ## 描述 在 CATLASS DSL 合入 master（!1122）时，公开树收口删除了 python/tla_dsl/tests/ 与部分未就绪 / 非中文 README 的 e2e 样例。本 PR 在当前 master（已含 hf32 运行时支持，见 !1133 / dsl@4f335a2e）上： 1. **补回** 此前收口去掉的 pytest / lit 测试树与相关 e2e 样例（debug_print、print_tensor、simt、tensor_index、dataclass_arg、lazy_conditions、scalar_arg_alignment、cross_flag_two_way 等）；**仍不带入** block_sparse_attention（BSA）； 2. **同步纳入** dsl 分支 4f335a2e（hf32 support）携带的测试：mmad-hf32-mode.mlir / mmad-hf32-mode-verify.test、以及 tests/test_frontend_lowering.py 中的 hf32 用例； 3. 恢复文档 / pyproject.toml / tests/dsl_battery / tests/run_dsl_test.sh 中对应入口，并与 master 上 basic_mmad_epilogue 命名保持一致。 **未改动** master 已落地的 hf32 运行时与 basic_matmul / common/golden.py / common/utils.py 等端到端实现（以 !1133 为准）。 ### 源分支 - cann/catlass：restore/dsl-samples-tests-hf32（基于 master @ 769cd40a） - tip：e8e24273 — dsl: restore pytest/lit and e2e samples, include hf32 tests ### MR https://gitcode.com/cann/catlass/merge_requests/1143 ## 关联的Issue 无。 ## 原因 迁移合入时为降低公开树噪音暂时下架了全量单测与部分样例；现 master 已具备 hf32 能力，需要把测试样例与 hf32 相关回归一并补回，便于主线持续验证。 ## 测试 - [ ] cd python/tla_dsl && ./build.sh 后：python -m pytest -q tests - [ ] python -m pytest -q tests/test_frontend_lowering.py -k hf32 - [ ] lit -sv csrc/mlir/build/tests/lit --filter=mmad-hf32-mode - [ ] bash tests/run_dsl_test.sh --device <id>（含补回样例） ## 文档更新 - 恢复 python/tla_dsl/README.md 与开发指南中对 pytest / lit / 仓库级回归的入口说明 - 在 01_build_and_test.md 补充 hf32 相关 pytest / lit 过滤命令 ## 类型标签 - [x] 其他：测试样例 / 单测树补充 See merge request: cann/catlass!1143 | 2 个月前 |