[CI/Build] Add CI end-to-end (E2E) tests (#139)

* [CI/Build] Add CI end-to-end (E2E) tests
Signed-off-by: Chenchao Hu <huchenchao@example.com>

.github/workflows/_e2e_singlecard.yml

@@ -0,0 +1,141 @@
+name: e2e-test
+
+on:
+  workflow_call:
+  pull_request:
+    branches: [main]
+    types: [opened, synchronize, reopened]
+  push:
+    branches: [main]
+
+concurrency:
+  group: e2e-singlecard
+  cancel-in-progress: false
+
+jobs:
+  e2e:
+    name: e2e-test-singlecard
+    runs-on:
+      - self-hosted
+      - Linux
+      - X64
+
+    steps:
+      - name: Checkout PR code
+        uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
+
+      - name: Verify PR workspace
+        run: |
+          echo "===== WORKSPACE ====="
+          pwd
+          ls -l
+          echo "===== GIT INFO ====="
+          git rev-parse HEAD
+          git log -1 --oneline
+          git status --porcelain
+
+      - name: Start docker
+        run: |
+          bash ci/scripts/docker/start_docker.sh
+
+      - name: Install enviroments
+        run: |
+          bash ci/scripts/env/install_env.sh
+
+      - name: Start vLLM server
+        run: |
+          bash ci/scripts/server/start_vllm.sh
+
+      - name: Wait for vLLM ready
+        run: |
+          bash ci/scripts/server/wait_vllm.sh
+
+      - name: Accuracy testing
+        run: |
+          bash ci/scripts/tests/run_accuracy.sh
+
+      - name: Performance testing
+        run: |
+          docker exec aiak-e2e-singlecard bash -lc '
+            source ci/scripts/common/env.sh
+            source ci/scripts/common/log.sh
+            #!/bin/bash
+            # ==========================================
+            # 1. Define test dimensions
+            #    (can be easily extended, e.g., add "2048x2048")
+            # ==========================================
+            DIMENSIONS=("1024x1024")
+
+            # ==========================================
+            # 2. Define concurrency generation logic (densification strategy)
+            # ============x==============================
+            # Use array concatenation to combine different density ranges
+            # Syntax: seq [start] [step] [end]
+            CONCURRENCIES=(1)
+
+            # ==========================================
+            # 3. Automatically assemble test cases
+            # ==========================================
+            TEST_COMBINATIONS=() # Initialize empty array
+
+            # 🔄 Modified: outer loop over batch size (concurrency), inner loop over dimensions
+            for bs in "${CONCURRENCIES[@]}"; do    # ← outer loop: concurrency
+                for dim in "${DIMENSIONS[@]}"; do  # ← inner loop: dimensions
+                    case_str="${bs}x${dim}"
+                    TEST_COMBINATIONS+=("$case_str")
+                done
+            done
+
+            # ==========================================
+            # 4. (Optional) Print generated cases for sanity check
+            # ==========================================
+            echo "Generated ${#TEST_COMBINATIONS[@]} test cases in total:"
+            echo "${TEST_COMBINATIONS[@]}" # Uncomment if you want to print all cases
+
+            # Progress counters
+            TOTAL_TESTS=${#TEST_COMBINATIONS[@]}
+            CURRENT_TEST=0
+
+            # Iterate over all test combinations
+            for COMBINATION in "${TEST_COMBINATIONS[@]}"; do
+                # Parse parameters from combination string
+                NUM_PROMPTS=$(echo $COMBINATION | cut -d'x' -f1)
+                INPUT_LEN=$(echo $COMBINATION | cut -d'x' -f2)
+                OUTPUT_LEN=$(echo $COMBINATION | cut -d'x' -f3)
+
+                # Update progress
+                CURRENT_TEST=$((CURRENT_TEST + 1))
+
+                echo "=========================================================="
+                echo "Test progress: $CURRENT_TEST / $TOTAL_TESTS"
+                echo "Current configuration: concurrency=$NUM_PROMPTS, input_len=$INPUT_LEN, output_len=$OUTPUT_LEN"
+                echo "=========================================================="
+
+                #OUTPUT_FILE="$RESULT_DIR/p800_${NUM_PROMPTS}_${INPUT_LEN}_${OUTPUT_LEN}.log"
+
+                # Run benchmark
+                python3 -m vllm.entrypoints.cli.main bench serve \
+                    --host 127.0.0.1 \
+                    --port ${VLLM_PORT:-8356}\
+                    --backend vllm \
+                    --model ${SERVED_MODEL_NAME:-Qwen3-8B} \
+                    --dataset-name random \
+                    --num-prompts $NUM_PROMPTS \
+                    --random-input-len $INPUT_LEN \
+                    --random-output-len $OUTPUT_LEN \
+                    --tokenizer ${MODEL_PATH:-/ssd3/models/Qwen3-8B} \
+                    --ignore-eos
+            done
+          '
+      
+      - name: Set permissions
+        if: always()
+        run: |
+          bash ci/scripts/docker/set_permissions.sh
+
+      - name: Cleanup docker
+        if: always()
+        run: |
+          bash ci/scripts/docker/stop_docker.sh