fix: critical bugs, architecture improvements, and test infrastructure #55

	# Continuous Integration
	# Validates code formatting, CUDA compilation, and CPU-only test execution
	#
	# Workflow separation:
	# - This workflow: formatting + CUDA compile-time checks + CPU tests (no GPU runtime)
	# - pages.yml: docs tests/build and GitHub Pages buildability
	name: CI

	on:
	push:
	branches: [master]
	paths-ignore:
	- '*/.md'
	- 'docs/**'
	pull_request:
	branches: [master]
	paths-ignore:
	- '*/.md'
	- 'docs/**'
	workflow_dispatch:

	permissions:
	contents: read

	concurrency:
	group: ci-${{ github.ref }}
	cancel-in-progress: true

	jobs:
	format-check:
	name: Format Check
	runs-on: ubuntu-latest
	steps:
	- name: Checkout
	uses: actions/checkout@v4

	- name: Check formatting
	uses: jidicula/clang-format-action@v4.14.0
	with:
	clang-format-version: '17'
	fallback-style: 'LLVM'

	cuda-build:
	name: CUDA Build
	runs-on: ubuntu-latest
	container:
	image: nvidia/cuda:12.4.1-devel-ubuntu22.04
	steps:
	- name: Install dependencies
	run: apt-get update && apt-get install -y cmake git build-essential

	- name: Checkout
	uses: actions/checkout@v4

	- name: Configure
	run: cmake -S . -B build -DCMAKE_BUILD_TYPE=Release -DBUILD_TESTS=ON

	- name: Build
	run: cmake --build build -j2

	- name: Run CPU tests
	run: ctest --test-dir build -L cpu --output-on-failure

	- name: Info
	run: \|
	echo "✅ CUDA compilation successful for all targets (benchmark + tests)"
	echo "✅ CPU-only tests passed"
	echo "ℹ️ GPU runtime tests require a CUDA-capable machine"

Provide feedback