[CICD] Add P2P engine perf benchmark (one-sided read/write) (#498) #387

	name: "Unit Tests: runner"

	on:
	push:
	branches:
	- main
	pull_request:
	branches:
	- main

	jobs:
	unittest-runner:
	runs-on: [self-hosted, cx-build]
	container:
	image: localhost:5000/flagscale:cuda12.8.1-cudnn9.7.1-python3.12-torch2.7.0-time2507111538
	options: --gpus all --privileged --ipc=host --ulimit memlock=-1 --ulimit stack=67108864
	steps:
	- name: Checkout Code
	uses: actions/checkout@v6
	with:
	repository: ${{ github.event.pull_request.head.repo.full_name }}
	ref: ${{ github.event.pull_request.head.ref }}
	ssh-strict: true
	ssh-user: git
	persist-credentials: true
	clean: true
	sparse-checkout-cone-mode: true
	fetch-tags: false
	show-progress: true
	lfs: false
	submodules: true
	set-safe-directory: true

	- name: Build Google Test
	run: \|
	cd /__w/FlagCX/FlagCX/third-party/googletest
	mkdir -p build
	cd build
	cmake ..
	make -j$(nproc)

	- name: Build FlagCX
	run: \|
	cd /__w/FlagCX/FlagCX
	export MPI_HOME=/usr/local/mpi
	make -j$(nproc) USE_NVIDIA=1 COMPILE_KERNEL=1

	- name: Build runner unit tests
	run: \|
	cd /__w/FlagCX/FlagCX/test/unittest/runner
	export MPI_HOME=/usr/local/mpi
	make -j$(nproc)

	- name: Run runner unit tests (no MPI)
	run: \|
	cd /__w/FlagCX/FlagCX/test/unittest/runner
	export MPI_HOME=/usr/local/mpi
	export LD_LIBRARY_PATH=/__w/FlagCX/FlagCX/build/lib:$LD_LIBRARY_PATH
	make run-unit

	- name: Run runner MPI collective tests
	run: \|
	cd /__w/FlagCX/FlagCX/test/unittest/runner
	export MPI_HOME=/usr/local/mpi
	export PATH=$MPI_HOME/bin:$PATH
	export LD_LIBRARY_PATH=/__w/FlagCX/FlagCX/build/lib:$LD_LIBRARY_PATH
	make run-mpi

Provide feedback