llama.cpp/docs/debugging-tests.md

Debugging Tests Tips

How to run & execute a specific test without anything else to keep the feedback loop short?

There is a script called run-single-test.sh in the scripts folder whose parameter takes a REGEX and an optional test number.

For example, running the following command will output an interactive list from which you can select a test. It takes this form:

run-single-test.sh [OPTION]... <test_regex> <test_number>

It will then build & run in the debugger for you.

./scripts/run-single-test.sh test-tokenizer

An example of a single test output is shown below. You will get either a green TEST PASS or a red TEST FAIL if a particular test is working or not. This shorter feedback loop will hopefully make it easier for you to figure out the problem you are trying to solve.

$ ./scripts/run-single-test.sh test 24
~/gitextern/llama.cpp ~/gitextern/llama.cpp

... prepping cmake environment ...
... building test binaries ...
... running test ...

Ran Test #24: test-eval-callback
Command: /home/mofosyne/gitextern/llama.cpp/build-ci-debug/bin/eval-callback "--hf-repo" "ggml-org/models" "--hf-file" "tinyllamas/stories260K.gguf" "--model" "stories260K.gguf" "--prompt" "hello" "--seed" "42" "-ngl" "0"
TEST PASS

For further reference use run-single-test.sh -h to print help.

How does the script work?

This is similar to debug-test.sh so you can follow the similar guide in this page for similar process. Just run the command directly rather than though gdb.

How to run & debug a specific test without anything else to keep the feedback loop short?

There is a script called debug-test.sh in the scripts folder whose parameter takes a REGEX and an optional test number.

For example, running the following command will output an interactive list from which you can select a test. It takes this form:

debug-test.sh [OPTION]... <test_regex> <test_number>

It will then build & run in the debugger for you.

./scripts/debug-test.sh test-tokenizer

# Once in the debugger, i.e. at the chevrons prompt, setting a breakpoint could be as follows:
>>> b main

For further reference use debug-test.sh -h to print help.

 

How does the script work?

If you want to be able to use the concepts contained in the script separately, the important ones are briefly outlined below.

Step 1: Reset and Setup folder context

From base of this repository, let's create build-ci-debug as our build context.

rm -rf build-ci-debug && mkdir build-ci-debug && cd build-ci-debug

Step 2: Setup Build Environment and Compile Test Binaries

Setup and trigger a build under debug mode. You may adapt the arguments as needed, but in this case these are sane defaults.

cmake -DCMAKE_BUILD_TYPE=Debug -DLLAMA_CUDA=1 -DLLAMA_FATAL_WARNINGS=ON ..
make -j

Step 3.1: Identify Test Command for Debugging

The output of this command will give you the command & arguments needed to run GDB.

• -R test-tokenizer : looks for all the test files named test-tokenizer* (R=Regex)
• -N : "show-only" disables test execution & shows test commands that you can feed to GDB.
• -V : Verbose Mode

ctest -R "test-tokenizer" -V -N

This may return output similar to below (focusing on key lines to pay attention to):

...
1: Test command: ~/llama.cpp/build-ci-debug/bin/test-tokenizer-0 "~/llama.cpp/tests/../models/ggml-vocab-llama-spm.gguf"
1: Working Directory: .
Labels: main
  Test  #1: test-tokenizer-0-llama-spm
...
4: Test command: ~/llama.cpp/build-ci-debug/bin/test-tokenizer-0 "~/llama.cpp/tests/../models/ggml-vocab-falcon.gguf"
4: Working Directory: .
Labels: main
  Test  #4: test-tokenizer-0-falcon
...

So for test #1 we can tell these two pieces of relevant information:

• Test Binary: ~/llama.cpp/build-ci-debug/bin/test-tokenizer-0
• Test GGUF Model: ~/llama.cpp/tests/../models/ggml-vocab-llama-spm.gguf

Step 3.2: Run GDB on test command

Based on the ctest 'test command' report above we can then run a gdb session via this command below:

gdb --args ${Test Binary} ${Test GGUF Model}

Example:

gdb --args ~/llama.cpp/build-ci-debug/bin/test-tokenizer-0 "~/llama.cpp/tests/../models/ggml-vocab-llama-spm.gguf"