Skip to content

kali20gakki/msAgent / skills/operator/ops-performance-tuning

ops-performance-tuning

Verified

Compile, profile, diagnose, optimize, and compare Ascend NPU operators across CANN versions and A2/A3/A5 for Ascend C, CATLASS, Triton-Ascend, TileLang-Ascend, PyPTO, and SHMEM/MC2. Use when an operator already has a runnable implementation and the user asks for msOpProf collection, bottleneck analysis, source-level tuning, or a reproducible before/after performance report. Do not use as the primary workflow for operator creation, migration, or unresolved correctness failures.

FileKindSize
SKILL.mdskill md15.4 KB
evals/evals.jsonconfig3.6 KB
references/case-routing.mdmarkdown30.0 KB
references/cases/ascendc/pr_ascendc_allgathermm_scalar_align.mdmarkdown1.7 KB
references/cases/ascendc/pr_ascendc_cann_samples_perf_story.mdmarkdown3.9 KB
references/cases/ascendc/pr_ascendc_cgdr_vf_inverse.mdmarkdown1.6 KB
references/cases/ascendc/pr_ascendc_conv3ddx_tque_tbuf.mdmarkdown1.6 KB
references/cases/ascendc/pr_ascendc_layernorm_areg.mdmarkdown1.6 KB
references/cases/ascendc/pr_ascendc_logit_a5_microapi.mdmarkdown1.9 KB
references/cases/ascendc/pr_ascendc_matmul_case_18x.mdmarkdown1.5 KB
references/cases/ascendc/pr_ascendc_matmul_full_load_wiki.mdmarkdown1.6 KB
references/cases/ascendc/pr_ascendc_matmul_vcv_wiki.mdmarkdown1.8 KB
references/cases/ascendc/pr_ascendc_mc2_fusion.mdmarkdown1.4 KB
references/cases/ascendc/pr_ascendc_mhc_24x.mdmarkdown1.2 KB
references/cases/ascendc/pr_ascendc_msopprof_add_block_imbalance.mdmarkdown1.7 KB
references/cases/ascendc/pr_ascendc_quantize_a2_micro_opt.mdmarkdown2.0 KB
references/cases/ascendc/pr_ascendc_rotate_quant_blaze.mdmarkdown2.8 KB
references/cases/ascendc/pr_ascendc_tiling_l2_official.mdmarkdown1.6 KB
references/cases/catlass/pr_catlass_baseline.mdmarkdown1.3 KB
references/cases/catlass/pr_catlass_fixpipe_dualdst.mdmarkdown1.6 KB
references/cases/catlass/pr_catlass_grouped_matmul_tiling.mdmarkdown1.5 KB
references/cases/catlass/pr_catlass_mstuner_3p8x.mdmarkdown1.9 KB
references/cases/catlass/pr_catlass_swizzle_balance.mdmarkdown1.6 KB
references/cases/pypto/pr_pypto_gqa_decode_vector_merge.mdmarkdown2.4 KB
references/cases/pypto/pr_pypto_l2_hit_layout_s13.mdmarkdown1.9 KB
references/cases/pypto/pr_pypto_per_matmul_tile_shapes.mdmarkdown2.4 KB
references/cases/pypto/pr_pypto_small_shape_matmul.mdmarkdown2.3 KB
references/cases/pypto/pr_pypto_weight_none_cacheable.mdmarkdown2.6 KB
references/cases/shmem/pr_shmem_alltoallv_mr152.mdmarkdown2.4 KB
references/cases/shmem/pr_shmem_reduce_scatter_serial.mdmarkdown2.6 KB
references/cases/tilelang/pr_tilelang_benchmark_vs_ascendc.mdmarkdown1.7 KB
references/cases/tilelang/pr_tilelang_compiler_prs.mdmarkdown2.3 KB
references/cases/tilelang/pr_tilelang_fa_pr698.mdmarkdown1.5 KB
references/cases/tilelang/pr_tilelang_gqa_bwd_pr1494.mdmarkdown1.3 KB
references/cases/tilelang/pr_tilelang_persistent_mqa_decode.mdmarkdown2.2 KB
references/cases/tilelang/pr_tilelang_sfa_pr665.mdmarkdown914 B
references/cases/triton/pr_triton_atomic_simd_24x.mdmarkdown1.8 KB
references/cases/triton/pr_triton_autotune_background.mdmarkdown1.0 KB
references/cases/triton/pr_triton_constexpr_recompile.mdmarkdown1.1 KB
references/cases/triton/pr_triton_lgamma_simt.mdmarkdown1.6 KB
references/compile/compile-ascendc.mdmarkdown9.0 KB
references/compile/compile-catlass.mdmarkdown2.4 KB
references/compile/compile-pypto.mdmarkdown2.0 KB
references/compile/compile-shmem.mdmarkdown2.6 KB
references/compile/compile-tilelang.mdmarkdown2.7 KB
references/compile/compile-triton.mdmarkdown4.4 KB
references/diagnose/diagnose-l2-cache.mdmarkdown2.2 KB
references/diagnose/diagnose-occupancy.mdmarkdown4.0 KB
references/diagnose/diagnose-pipeline.mdmarkdown4.1 KB
references/diagnose/diagnose-roofline.mdmarkdown3.1 KB
references/optimize/optimize-api-usage.mdmarkdown1.7 KB
references/optimize/optimize-ascendc-regbase-vf.mdmarkdown5.9 KB
references/optimize/optimize-ascendc-tiling.mdmarkdown935 B
references/optimize/optimize-catlass.mdmarkdown3.0 KB
references/optimize/optimize-data-copy.mdmarkdown870 B
references/optimize/optimize-index.mdmarkdown3.6 KB
references/optimize/optimize-memory-hierarchy.mdmarkdown599 B
references/optimize/optimize-pipeline.mdmarkdown1.0 KB
references/optimize/optimize-pypto.mdmarkdown1.5 KB
references/optimize/optimize-shmem.mdmarkdown5.1 KB
references/optimize/optimize-tilelang.mdmarkdown1.7 KB
references/optimize/optimize-triton.mdmarkdown3.8 KB
references/profile/benchmark-harness.mdmarkdown2.2 KB
references/profile/profile-msopprof.mdmarkdown7.4 KB
references/troubleshooting.mdmarkdown5.8 KB
scripts/bound_analyzer.pyscript9.8 KB
scripts/delta_report.pyscript5.5 KB
scripts/env_check.shscript5.7 KB
scripts/result_saver.pyscript5.8 KB

references/cases/pypto/pr_pypto_gqa_decode_vector_merge.md

View on GitHub
# GQA Decode Attention:Vector 合轴 + 合图优化

## 基本信息
- 算子类别:attention
- DSL/框架:pypto
- 类型:非PR(内部案例库,cannbot-skills 官方调优案例库)
- 来源可信度:官方文档(cann/cannbot-skills 仓 `ops/pypto-op-perf-tune` 调优案例库,含 4 轮迭代失败分析与实测数字)

## 链接
- PR/出处链接:https://gitcode.com/cann/cannbot-skills/blob/master/ops/pypto-op-perf-tune/tune-frontecases/vector-axis-merge-softmax.md( 已验证可达,验证日期 2026-08)
- 优化代码查看:非代码 MR 类案例,无 diff/files 页;出处链接即为案例原文,内含完整优化前后代码与迭代记录

## 问题与瓶颈
GQA decode attention 算子中,softmax 及前后 vector 操作(mul/amax/sub/exp/sum/div/cast)在 3D shape `[8, 4, 2048]` 下执行,产生 6 个独立 vector 子图,调度开销大。属开箱调优(PHASE_FRONTEND)阶段案例。

## 优化方法(理论手段)
①matmul 输出 `reshape(inplace)` 为 2D `[32, 2048]`;②`set_vec_tile_shapes(8, 2048)` + `set_pass_options(sg_set_scope=1)` 使所有 vector 操作在 2D 下合图执行;③reshape 回 3D 传给下一个 matmul。机制:

1. **归约轴必须不切分**:vec_tile_shapes 第二维应等于实际归约轴长度(2048),切分为 512 会产生跨子图 reduce 开销(迭代 1 实测 +16.6% 回退);
2. **合图减少调度开销**:`sg_set_scope` 将连续 vector 操作合并为单个子图,任务数 168→137;
3. **UB 容量决定第一维上限**:`第一维 × 第二维 × dtype × tensor 总数` 不能超 UB(约 128KB 保守估计),`(32,2048)` 即因超 UB 编译失败;
4. 合轴后必须显式设 vec_tile_shapes,否则 reduce op 尾轴 32B 对齐推导失败。

## 性能对比
| 指标 | 优化前 | 优化后 | 变化 |
|---|---|---|---|
| 耗时 | 275.44 us | 258.98 us | -6.0% |
| 任务数 | 168 | 137 | -18.5% |
| 子图数 | 6 | 1 | 合并为单图 |
| 精度(Max difference) | 0.000031 | 0.000031 | 无变化 |

## 适用范围与警示
- 适用:GQA decode attention 等 softmax 前后连续 vector 操作链、多子图调度开销大的场景。
- 警示:归约轴切分会回退(迭代 1 实测 +16.6%);第一维受 UB 容量限制(约 128KB 保守估计),`(32,2048)` 超 UB 编译失败;合轴后必须显式设 vec_tile_shapes,否则 reduce op 尾轴 32B 对齐推导失败。
planning-and-task-breakdownaddyosmani/agent-skillsBreaks work into ordered tasks. Use when you have a spec or clear requirements and need to break work into implementable tasks. Use when a task feels too large to start, when you need to estimate scope, or when parallel work is possible.Stars: 119kUpdated Claude CodeCodexVerifiedInstall
screenpipethedotmack/claude-memSearch screen and audio history recorded by Screenpipe and optionally save selected context into claude-mem. Use when the user asks to recall desktop activity, a meeting, or something seen in another app through Screenpipe.Stars: 118kUpdated Claude CodeCodexVerifiedInstall
contact-widgetnexu-io/open-designSelf-contained floating chat widget with welcome screen, social links, meeting button, and message input. Single HTML file, zero dependencies.Stars: 108kUpdated Claude CodeCodexVerifiedInstall
meeting-notesnexu-io/open-designMeeting notes page — title bar with attendees, agenda checklist, decisions block, action items table with owners + dates, and a "next meeting" footer. Use when the brief mentions "meeting notes", "minutes", "1:1 notes", "all-hands recap", or "会议纪要".Stars: 108kUpdated Claude CodeCodexVerifiedInstall