Fdwic swimlane deps

.clang-format

@@ -0,0 +1,136 @@
+# Copyright (c) PyPTO Contributors.
+# This program is free software, you can redistribute it and/or modify it under the terms and conditions of
+# CANN Open Software License Agreement Version 2.0 (the "License").
+# Please refer to the License for details. You may not use this file except in compliance with the License.
+# THIS SOFTWARE IS PROVIDED ON AN "AS IS" BASIS, WITHOUT WARRANTIES OF ANY KIND, EITHER EXPRESS OR IMPLIED,
+# INCLUDING BUT NOT LIMITED TO NON-INFRINGEMENT, MERCHANTABILITY, OR FITNESS FOR A PARTICULAR PURPOSE.
+# See LICENSE in the root of the software repository for the full text of the License.
+# -----------------------------------------------------------------------------------------------------------
+
+BasedOnStyle: LLVM
+Language: Cpp
+
+# --- Indentation ---
+IndentWidth: 4
+TabWidth: 4
+UseTab: Never
+ContinuationIndentWidth: 4
+IndentCaseLabels: false
+IndentPPDirectives: None
+NamespaceIndentation: None
+
+# --- Column limit ---
+ColumnLimit: 120
+
+# --- Pointer & reference ---
+DerivePointerAlignment: false
+PointerAlignment: Right
+ReferenceAlignment: Right
+
+# --- Braces ---
+BreakBeforeBraces: Custom
+BraceWrapping:
+  AfterCaseLabel: false
+  AfterClass: false
+  AfterControlStatement: Never
+  AfterEnum: false
+  AfterFunction: false
+  AfterNamespace: false
+  AfterStruct: false
+  AfterUnion: false
+  AfterExternBlock: false
+  BeforeCatch: false
+  BeforeElse: false
+  BeforeLambdaBody: false
+  BeforeWhile: false
+  IndentBraces: false
+  SplitEmptyFunction: false
+  SplitEmptyRecord: false
+  SplitEmptyNamespace: false
+
+# --- Parameter packing & alignment ---
+AlignAfterOpenBracket: BlockIndent
+BinPackParameters: true
+BinPackArguments: true
+AllowAllParametersOfDeclarationOnNextLine: true
+AllowAllArgumentsOnNextLine: true
+
+# --- Short statements ---
+AllowShortBlocksOnASingleLine: Empty
+AllowShortFunctionsOnASingleLine: All
+AllowShortIfStatementsOnASingleLine: AllIfsAndElse
+AllowShortLambdasOnASingleLine: Empty
+AllowShortLoopsOnASingleLine: false
+AllowShortEnumsOnASingleLine: true
+AllowShortCaseLabelsOnASingleLine: false
+
+# --- Alignment ---
+AlignConsecutiveAssignments: false
+AlignConsecutiveDeclarations: false
+AlignConsecutiveMacros: false
+AlignTrailingComments: true
+AlignEscapedNewlines: Left
+AlignOperands: Align
+
+# --- Constructor initializers ---
+BreakConstructorInitializers: AfterColon
+PackConstructorInitializers: Never
+
+# --- Inheritance ---
+BreakInheritanceList: AfterComma
+
+# --- Template ---
+AlwaysBreakTemplateDeclarations: Yes
+
+# --- Binary & ternary operators ---
+BreakBeforeBinaryOperators: None
+BreakBeforeTernaryOperators: false
+
+# --- Includes ---
+SortIncludes: Never
+IncludeBlocks: Regroup
+IncludeCategories:
+  - Regex: '^<.*>'
+    Priority: 3
+  - Regex: '^".*"'
+    Priority: 1
+
+# --- Namespace ---
+FixNamespaceComments: true
+CompactNamespaces: false
+
+# --- Spaces ---
+SpaceBeforeParens: ControlStatements
+SpaceAfterCStyleCast: false
+SpaceAfterLogicalNot: false
+SpaceAfterTemplateKeyword: true
+SpaceBeforeAssignmentOperators: true
+SpaceBeforeCpp11BracedList: false
+SpaceBeforeCtorInitializerColon: true
+SpaceBeforeInheritanceColon: true
+SpaceBeforeRangeBasedForLoopColon: true
+SpaceInEmptyBlock: false
+SpacesBeforeTrailingComments: 2
+SpacesInAngles: Never
+SpacesInCStyleCastParentheses: false
+SpacesInConditionalStatement: false
+SpacesInContainerLiterals: false
+SpacesInParentheses: false
+SpacesInSquareBrackets: false
+
+# --- Empty lines ---
+MaxEmptyLinesToKeep: 1
+KeepEmptyLinesAtTheStartOfBlocks: false
+EmptyLineAfterAccessModifier: Never
+EmptyLineBeforeAccessModifier: LogicalBlock
+
+# --- Access modifier ---
+AccessModifierOffset: -4
+
+# --- Other ---
+Cpp11BracedListStyle: true
+BreakStringLiterals: true
+ReflowComments: true
+SortUsingDeclarations: true
+InsertBraces: false
+AlwaysBreakAfterReturnType: None

.clang-tidy

@@ -0,0 +1,46 @@
+# Copyright (c) PyPTO Contributors.
+# This program is free software, you can redistribute it and/or modify it under the terms and conditions of
+# CANN Open Software License Agreement Version 2.0 (the "License").
+# Please refer to the License for details. You may not use this file except in compliance with the License.
+# THIS SOFTWARE IS PROVIDED ON AN "AS IS" BASIS, WITHOUT WARRANTIES OF ANY KIND, EITHER EXPRESS OR IMPLIED,
+# INCLUDING BUT NOT LIMITED TO NON-INFRINGEMENT, MERCHANTABILITY, OR FITNESS FOR A PARTICULAR PURPOSE.
+# See LICENSE in the root of the software repository for the full text of the License.
+# -----------------------------------------------------------------------------------------------------------
+
+Checks: >
+  -*,
+  bugprone-*,
+  performance-*,
+  modernize-*,
+  -modernize-use-trailing-return-type,
+  -modernize-avoid-c-arrays,
+  -modernize-use-nodiscard,
+  -modernize-macro-to-enum,
+  -modernize-deprecated-headers,
+  -modernize-use-auto,
+  -modernize-use-nullptr,
+  -modernize-use-using,
+  -modernize-loop-convert,
+  -modernize-use-default-member-init,
+  -modernize-use-equals-delete,
+  -modernize-use-bool-literals,
+  -modernize-raw-string-literal,
+  -bugprone-easily-swappable-parameters,
+  -bugprone-narrowing-conversions,
+  -bugprone-reserved-identifier,
+  -bugprone-implicit-widening-of-multiplication-result,
+  -bugprone-multi-level-implicit-pointer-conversion,
+  -bugprone-branch-clone,
+  -bugprone-empty-catch,
+  -bugprone-lambda-function-name,
+  -bugprone-macro-parentheses,
+  -bugprone-infinite-loop,
+  -bugprone-argument-comment,
+  -performance-no-int-to-ptr,
+  -performance-enum-size
+
+WarningsAsErrors: '*'
+
+HeaderFilterRegex: '(src|include)/.*'
+
+FormatStyle: file

.claude/commands/perf-example-device.md

@@ -0,0 +1,10 @@
+# Benchmark the hardware performance of a single example at $ARGUMENTS
+
+Reference `tools/benchmark_rounds.sh` for the full implementation pattern (device log resolution, timing parsing, reporting format). This skill runs the same logic but for a single example only.
+
+1. Locate the test file under `$ARGUMENTS/`: pick the single `test_*.py` that lives directly in that directory. If none exists, tell the user the directory is not a scene test and stop.
+2. Check `command -v npu-smi` — if not found, tell the user this requires hardware and stop.
+3. **Detect platform**: Run `npu-smi info` and parse the chip name. Map `910B`/`910C` → `a2a3`, `950` → `a5`. If unrecognized, warn and default to `a2a3`.
+4. Find the lowest-ID idle device (HBM-Usage = 0) from the `npu-smi info` output. If none, stop.
+5. Run the example following the same pattern as `run_bench()` in `tools/benchmark_rounds.sh`:
+   - Snapshot logs, run `python $ARGUMENTS/test_<name>.py -p <platform> -d <device_id> --rounds 10 --skip-golden`, find the new log, parse timing, report results.

.claude/commands/perf-runtime-device.md

@@ -0,0 +1,14 @@
+# Benchmark the hardware performance of all scene tests under `tests/st/<arch>/<runtime>/`
+
+If `$ARGUMENTS` is provided, use it as the runtime name. Otherwise, default to `tensormap_and_ringbuffer`.
+
+Reference `tools/benchmark_rounds.sh` for the full implementation pattern (device log resolution, timing parsing, reporting format).
+
+1. Validate the runtime is one of: `host_build_graph`, `tensormap_and_ringbuffer`. If not, list valid runtimes and stop.
+2. Check `command -v npu-smi` — if not found, tell the user this requires hardware and stop.
+3. **Detect platform**: Run `npu-smi info` and parse the chip name. Map `910B`/`910C` → `a2a3`, `950` → `a5`. If unrecognized, warn and default to `a2a3`.
+4. Find the lowest-ID idle device (HBM-Usage = 0) from the `npu-smi info` output. If none, stop.
+5. Enumerate all subdirectories under `tests/st/<arch>/$ARGUMENTS/` that contain a `test_*.py` directly beneath them.
+6. For each example, run the same `run_bench()` pattern from `tools/benchmark_rounds.sh`:
+   - Snapshot logs, run `python <example>/test_<name>.py -p <platform> -d <device_id> --rounds 10 --skip-golden`, find the new log, parse timing, report results.
+7. Print a final summary table with example name, average latency, trimmed average, and pass/fail.

.claude/commands/profile.md

@@ -0,0 +1,7 @@
+# Run the example at $ARGUMENTS with profiling enabled on hardware
+
+1. Locate the test file under `$ARGUMENTS/`: pick the single `test_*.py` that lives directly in that directory. If none exists, tell the user the directory is not a scene test and stop.
+2. Pick an idle device (HBM-Usage = 0) from `npu-smi info`. If none, stop.
+3. Run: `python $ARGUMENTS/test_<name>.py -p a2a3 -d <device_id> --enable-profiling`
+4. If the test passes, report the swimlane output file location in `outputs/`.
+5. Summarize the task statistics from the console output (per-function timing breakdown).

.claude/commands/test-all-device.md

@@ -0,0 +1,19 @@
+# Run the full hardware CI pipeline with automatic device detection
+
+1. Check `command -v npu-smi` — if not found, tell the user to use `/test-all-sim` instead and stop
+2. **Detect platform**: Run `npu-smi info` and parse the chip name. Map `910B`/`910C` → `a2a3`, `950` → `a5`. If unrecognized, warn and default to `a2a3`
+3. Read `.github/workflows/ci.yml` to extract the current `--pto-isa-commit` and `--pto-session-timeout` values from the `st-onboard-<platform>` job's `pytest` invocation
+4. From the `npu-smi info` output, find devices whose **HBM-Usage is 0** (idle)
+5. From the idle devices, take **at most 4**. If no idle device is found, report the situation and stop
+6. Build the device range flag: from the idle devices, find the **longest consecutive sub-range** (at most 4). Pass as `--device <start>-<end>`. If no consecutive pair exists, use the lowest-ID idle device as `--device <id>`
+7. Run:
+
+   ```bash
+   pytest examples tests/st --platform <platform> --device <range-or-id> \
+     --pto-session-timeout <timeout> --clone-protocol https \
+     --pto-isa-commit <commit> -v
+   ```
+
+   Parallelism is auto-driven by `--device`: on hardware, one in-flight subprocess per device (`--max-parallel auto` = `len(--device)`); see `docs/testing.md` for the full reuse hierarchy.
+8. Report the results summary (pass/fail counts per task)
+9. If any tests fail, show the relevant error output and which device failed

.claude/commands/test-all-sim.md

@@ -0,0 +1,15 @@
+# Run the full simulation CI pipeline
+
+1. Read `.github/workflows/ci.yml` to extract the current `--pto-isa-commit` and `--pto-session-timeout` values from the `st-sim-*` jobs' `pytest` invocations
+2. **Detect platform**: If `npu-smi` is available, parse the chip name from `npu-smi info`. Map `910B`/`910C` → `a2a3sim`, `950` → `a5sim`. If `npu-smi` is not found, default to `a2a3sim`
+3. Build the command:
+
+   ```bash
+   pytest examples tests/st --platform <platform> \
+     --pto-session-timeout <timeout> --clone-protocol https \
+     --pto-isa-commit <commit> -v
+   ```
+
+4. Run the command (xdist parallelism is auto-enabled via `--max-parallel`, see `docs/testing.md`)
+5. Report the results summary (pass/fail counts)
+6. If any tests fail, show the relevant error output

.claude/commands/test-example-device.md

@@ -0,0 +1,15 @@
+# Run the hardware device test for the example at $ARGUMENTS
+
+1. Locate the test file under `$ARGUMENTS/`: pick the single `test_*.py` that lives directly in that directory (not in a subdirectory). If none exists, tell the user the directory is not a scene test and stop.
+2. Check `command -v npu-smi` — if not found, tell the user to use `/test-example-sim` instead and stop.
+3. **Detect platform**: Run `npu-smi info` and parse the chip name. Map `910B`/`910C` → `a2a3`, `950` → `a5`. If unrecognized, warn and default to `a2a3`.
+4. Read `.github/workflows/ci.yml` to extract the current `--pto-isa-commit` value from the `st-onboard-<platform>` job's `pytest` invocation.
+5. Pick an idle device: from `npu-smi info`, find one whose HBM-Usage is 0. If none is free, report and stop.
+6. Run standalone:
+
+   ```bash
+   python $ARGUMENTS/test_<name>.py -p <platform> -d <device_id> \
+     --clone-protocol https --pto-isa-commit <commit>
+   ```
+
+7. Report pass/fail status with any error output.

.claude/commands/test-example-sim.md

@@ -0,0 +1,13 @@
+# Run the simulation test for the example at $ARGUMENTS
+
+1. Locate the test file under `$ARGUMENTS/`: pick the single `test_*.py` that lives directly in that directory (not in a subdirectory). If none exists, tell the user the directory is not a scene test and stop.
+2. Read `.github/workflows/ci.yml` to extract the current `--pto-isa-commit` value from the `st-sim-*` jobs' `pytest` invocations.
+3. **Detect platform**: Infer the architecture from the path (e.g., `examples/a2a3/...` or `tests/st/a2a3/...` → `a2a3sim`; `examples/a5/...` or `tests/st/a5/...` → `a5sim`). If the path doesn't contain an arch prefix, default to `a2a3sim`.
+4. Run standalone:
+
+   ```bash
+   python $ARGUMENTS/test_<name>.py -p <platform> \
+     --clone-protocol https --pto-isa-commit <commit>
+   ```
+
+5. Report pass/fail status with any error output.

.claude/commands/test-runtime-device.md

@@ -0,0 +1,21 @@
+# Run hardware device tests for a single runtime specified by $ARGUMENTS
+
+1. Validate that `$ARGUMENTS` is one of: `host_build_graph`, `tensormap_and_ringbuffer`. If not, list the valid runtimes and stop.
+2. Check `command -v npu-smi` — if not found, tell the user to use `/test-runtime-sim` instead and stop.
+3. **Detect platform**: Run `npu-smi info` and parse the chip name. Map `910B`/`910C` → `a2a3`, `950` → `a5`. If unrecognized, warn and default to `a2a3`.
+4. Read `.github/workflows/ci.yml` to extract the current `--pto-isa-commit` and `--pto-session-timeout` values from the `st-onboard-<platform>` job's `pytest` invocation.
+5. From the `npu-smi info` output, find devices whose **HBM-Usage is 0** (idle).
+6. From the idle devices, take **at most 4**. If no idle device is found, report the situation and stop.
+7. Build the device range flag: from the idle devices, find the **longest consecutive sub-range** (at most 4). Pass as `--device <start>-<end>`. If no consecutive pair exists, use the lowest-ID idle device as `--device <id>`.
+8. Run:
+
+   ```bash
+   pytest examples tests/st --platform <platform> --runtime $ARGUMENTS \
+     --device <range-or-id> \
+     --pto-session-timeout <timeout> --clone-protocol https \
+     --pto-isa-commit <commit> -v
+   ```
+
+   Hardware parallelism is auto-driven by `--device` (one subprocess per device); no extra flag needed.
+9. Report the results summary (pass/fail counts per task).
+10. If any tests fail, show the relevant error output and which device failed.

.claude/commands/test-runtime-sim.md

@@ -0,0 +1,16 @@
+# Run simulation tests for a single runtime specified by $ARGUMENTS
+
+1. Validate that `$ARGUMENTS` is one of: `host_build_graph`, `tensormap_and_ringbuffer`. If not, list the valid runtimes and stop.
+2. Read `.github/workflows/ci.yml` to extract the current `--pto-isa-commit` and `--pto-session-timeout` values from the `st-sim-*` jobs' `pytest` invocations.
+3. **Detect platform**: If `npu-smi` is available, parse the chip name from `npu-smi info`. Map `910B`/`910C` → `a2a3sim`, `950` → `a5sim`. If `npu-smi` is not found, default to `a2a3sim`.
+4. Run:
+
+   ```bash
+   pytest examples tests/st --platform <platform> --runtime $ARGUMENTS \
+     --pto-session-timeout <timeout> --clone-protocol https \
+     --pto-isa-commit <commit> -v
+   ```
+
+   xdist parallelism is auto-selected via `--max-parallel auto` (min of nproc and device pool size on sim); see `docs/testing.md`.
+5. Report the results summary (pass/fail counts).
+6. If any tests fail, show the relevant error output.

.claude/lib/github/README.md

@@ -0,0 +1,35 @@
+# GitHub Shared Procedures
+
+Reusable procedures for GitHub PR workflows.
+
+## Available Procedures
+
+| Procedure | Description | Used By |
+| --------- | ----------- | ------- |
+| [setup](setup.md) | Authenticate and detect repository context | All |
+| [lookup-pr](lookup-pr.md) | Find PR by number, branch, or list all | All |
+| [detect-permission](detect-permission.md) | Check push access to PR | fix-pr |
+| [commit-and-push](commit-and-push.md) | Squash commits, rebase, and push | github-pr, fix-pr |
+| [fetch-comments](fetch-comments.md) | Get unresolved PR review comments | fix-pr |
+| [reply-and-resolve](reply-and-resolve.md) | Reply to and resolve review threads | fix-pr |
+| [branch-naming](branch-naming.md) | Generate branch name from commit | github-pr |
+| [common-issues](common-issues.md) | Troubleshooting reference | All |
+
+## Standard Variables
+
+After running `detect-context`, these variables are available:
+
+| Variable | Description | Example (owner) | Example (fork) |
+| -------- | ----------- | --------------- | -------------- |
+| `REPO_OWNER` | Origin repo owner | `hw-native-sys` | `contributor` |
+| `REPO_NAME` | Origin repo name | `simpler` | `simpler` |
+| `PR_REPO_OWNER` | PR target owner | `hw-native-sys` | `hw-native-sys` |
+| `PR_REPO_NAME` | PR target name | `simpler` | `simpler` |
+| `DEFAULT_BRANCH` | Base branch name | `main` | `main` |
+| `BASE_REF` | Full base ref | `origin/main` | `upstream/main` |
+| `PUSH_REMOTE` | Where to push | `origin` | `origin` |
+| `PR_HEAD_PREFIX` | PR head prefix | *(empty)* | `contributor:` |
+| `ROLE` | Repository role | `owner` | `fork` |
+| `BRANCH_NAME` | Current branch | `feat/new` | `feat/new` |
+| `COMMITS_AHEAD` | Commits ahead of base | `1` | `1` |
+| `UNCOMMITTED` | Uncommitted changes | *(empty or files)* | *(empty or files)* |

.claude/lib/github/branch-naming.md

@@ -0,0 +1,29 @@
+# Branch Naming
+
+Generate branch name from commit message.
+
+## Rules
+
+1. Determine the commit type prefix:
+
+| Commit type | Branch prefix |
+| ----------- | ------------- |
+| Add | `feat/` |
+| Fix | `fix/` |
+| Update | `feat/` |
+| Refactor | `refactor/` |
+| Support | `support/` |
+| Sim | `sim/` |
+| CI | `support/` |
+
+2. Take the commit subject description (after `Type: `), lowercase it, replace spaces and special characters with hyphens, strip trailing hyphens.
+
+3. Truncate to 50 characters.
+
+## Examples
+
+**Short commit (no truncation):**
+`Refactor: inline ring buffer hot paths` → `refactor/inline-ring-buffer-hot-paths` (37 chars)
+
+**Long commit (with truncation):**
+`Support: add complete-phase poll hit-rate logging with detailed statistics` → `support/add-complete-phase-poll-hit-rate-logging-w` (50 chars)