echo "Select Benchmark Suite:"

echo "1) Test 1: Automated Context & Memory Profile (TPS & RAM matrix)"

echo "2) Test 2: Prompt Cache & Sliding Window Regression Test"

echo "3) Quit"

read -p "Option (1-3): " suite_opt

if [ "$suite_opt" == "3" ] || [ -z "$suite_opt" ]; then

echo "Exiting."

exit 0

if [ -f ".build/arm64-apple-macosx/release/SwiftLM" ]; then

BIN=".build/arm64-apple-macosx/release/SwiftLM"

elif [ -f ".build/release/SwiftLM" ]; then

BIN=".build/release/SwiftLM"

echo "⚠️ SwiftLM release binary not found! Please compile the project by running ./build.sh first."

exit 1

if [ "$suite_opt" == "2" ]; then

echo ""

echo "=> Starting Prompt Cache Regression Test"

echo "Generating /tmp/big_prompt.json (approx 5K tokens)..."

python3 -c 'import json; open("/tmp/big_prompt.json", "w").write(json.dumps({"messages": [{"role": "user", "content": "apple "*4500}], "max_tokens": 30}))'

echo "Starting Server in background..."

killall SwiftLM 2>/dev/null

mkdir -p tmp

$BIN --model mlx-community/gemma-4-26b-a4b-it-4bit --port 5431 --turbo-kv --stream-experts --ctx-size 16384 > ./tmp/regression_server.log 2>&1 &

SERVER_PID=$!

echo "Waiting for server to be ready on port 5431 (this may take a minute if downloading)..."

for i in {1..300}; do

if curl -s http://127.0.0.1:5431/health > /dev/null; then break; fi

sleep 1

done

echo ""

echo "Server is up! Running 4-request sliding window validation..."

echo "=== Req 1 (Big 5537t) ===" && curl -sS --max-time 120 http://127.0.0.1:5431/v1/chat/completions -H "Content-Type: application/json" -d @/tmp/big_prompt.json 2>&1 | python3 -c "import sys,json;d=json.load(sys.stdin);print('OK:',d['choices'][0]['message']['content'])" && \

echo "=== Req 2 (Short 18t) ===" && curl -sS --max-time 60 http://127.0.0.1:5431/v1/chat/completions -H "Content-Type: application/json" -d '{"messages":[{"role":"user","content":"What is today?"}],"max_tokens":30}' 2>&1 | python3 -c "import sys,json;d=json.load(sys.stdin);print('OK:',d['choices'][0]['message']['content'])" && \

echo "=== Req 3 (Big 5537t) ===" && curl -sS --max-time 120 http://127.0.0.1:5431/v1/chat/completions -H "Content-Type: application/json" -d @/tmp/big_prompt.json 2>&1 | python3 -c "import sys,json;d=json.load(sys.stdin);print('OK:',d['choices'][0]['message']['content'])" && \

echo "=== Req 4 (Big Full Cache Hit) ===" && curl -sS --max-time 120 http://127.0.0.1:5431/v1/chat/completions -H "Content-Type: application/json" -d @/tmp/big_prompt.json 2>&1 | python3 -c "import sys,json;d=json.load(sys.stdin);print('OK:',d['choices'][0]['message']['content'])" && \

echo "=== ALL 4 PASSED ==="

echo ""

echo "Cleaning up..."

killall SwiftLM

wait $SERVER_PID 2>/dev/null

exit 0

echo ""

PS3="Select a model to profile (1-7): "