add to benchmark

python/examples/pipeline/simple_web_service/benchmark.py

+import sys
+import os
+import yaml
+import requests
+import time
+import json
+try:
+    from paddle_serving_server_gpu.pipeline import PipelineClient
+except ImportError:
+    from paddle_serving_server.pipeline import PipelineClient
+import numpy as np
+from paddle_serving_client.utils import MultiThreadRunner
+from paddle_serving_client.utils import benchmark_args, show_latency
+
+def gen_yml():
+    fin = open("config.yml", "r")
+    config = yaml.load(fin)
+    fin.close()
+    config["dag"]["tracer"] = {"interval_s": 5}
+    with open("config2.yml", "w") as fout: 
+        yaml.dump(config, fout, default_flow_style=False)
+
+def run_http(idx, batch_size):
+    print("start thread ({})".format(idx))
+    url = "http://127.0.0.1:18082/uci/prediction"    
+    start = time.time()
+    data = {"key": ["x"], "value": ["0.0137, -0.1136, 0.2553, -0.0692, 0.0582, -0.0727, -0.1583, -0.0584, 0.6283, 0.4919, 0.1856, 0.0795, -0.0332"]}
+    for i in range(1000):
+        r = requests.post(url=url, data=json.dumps(data))
+    print(r.json())
+    end = time.time()
+    return [[end - start]]
+
+def multithread_http(thread, batch_size):
+    multi_thread_runner = MultiThreadRunner()
+    result = multi_thread_runner.run(
+        run_http , thread, batch_size)
+
+
+def multithread_rpc(thraed, batch_size):
+    multi_thread_runner = MultiThreadRunner()
+
+
+def run_rpc(thread, batch_size):
+    client = PipelineClient()
+    client.connect(['127.0.0.1:9998'])
+    data = {"key": "x", "value": "0.0137, -0.1136, 0.2553, -0.0692, 0.0582, -0.0727, -0.1583, -0.0584, 0.6283, 0.4919, 0.1856, 0.0795, -0.0332"}
+    ret = client.predict(feed_dict={data["key"]: data["value"]}, fetch=["res"])
+    print(ret)
+
+if __name__ == "__main__":
+    if sys.argv[1] == "yaml":
+        mode = sys.argv[2] # brpc/  local predictor
+        thread = int(sys.argv[3])
+        gen_yml()
+    elif sys.argv[1] == "run":
+        mode = sys.argv[2] # http/ rpc
+        thread = int(sys.argv[3])
+        batch_size = int(sys.argv[4])
+        if mode == "http":
+            multithread_http(thread, batch_size)
+        elif mode == "rpc":
+            run_rpc(thread, batch_size)
+
+    

python/examples/pipeline/simple_web_service/benchmark.sh

+ps -ef | grep web_service | awk '{print $2}' | xargs kill -9 
+rm -rf PipelineServingLogs
+rm -rf cpu_utilization.py
+sleep 3
+python3 benchmark.py yaml local_predictor 1 
+
+for thread_num in 1 8 32
+do
+python3 web_service.py >web.log 2>&1 &
+sleep 3
+echo "import psutil\ncpu_utilization=psutil.cpu_percent(1,False)\nprint('CPU_UTILIZATION:', cpu_utilization)\n" > cpu_utilization.py
+python3 benchmark.py run http $thread_num 1
+python3 cpu_utilization.py
+echo "------------Fit a line pipeline benchmark (Thread: $thread_num) (Benchmark: 1)"
+tail -n 25 PipelineServingLogs/pipeline.tracer 
+ps -ef | grep web_service | awk '{print $2}' | xargs kill -9
+done

python/examples/pipeline/simple_web_service/web_service.py

@@ -61,5 +61,5 @@ class UciService(WebService):
 
 
 uci_service = UciService(name="uci")
-uci_service.prepare_pipeline_config("config.yml")
+uci_service.prepare_pipeline_config("config2.yml")
 uci_service.run_service()