Add --warmup-request-count to CLI help message and docs (#180)

docs/cli.md

@@ -1,5 +1,5 @@
 <!--
-Copyright (c) 2023, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+Copyright (c) 2024, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
 
 Redistribution and use in source and binary forms, with or without
 modification, are permitted provided that the following conditions
@@ -157,6 +157,12 @@ will also be reported in the results.
 Default is `-1` indicating that the average latency is used to determine
 stability.
 
+#### `--warmup-request-count=<n>`
+
+Specifies the number of warmup requests to send before benchmarking.
+
+Default is `0`, which means that no warmup requests will be sent.
+
 #### `--request-count=<n>`
 
 Specifies a total number of requests to use for measurement.

src/command_line_parser.cc

@@ -138,6 +138,8 @@ CLParser::Usage(const std::string& msg)
             << std::endl;
   std::cerr << "\t--percentile <percentile>" << std::endl;
   std::cerr << "\t--request-count <number of requests>" << std::endl;
+  std::cerr << "\t--warmup-request-count <number of warmup requests>"
+            << std::endl;
   std::cerr << "\tDEPRECATED OPTIONS" << std::endl;
   std::cerr << "\t-t <number of concurrent requests>" << std::endl;
   std::cerr << "\t-c <maximum concurrency>" << std::endl;
@@ -472,6 +474,12 @@ CLParser::Usage(const std::string& msg)
              "until stabilization is detected.",
              18)
       << std::endl;
+  std::cerr << FormatMessage(
+                   " --warmup-request-count: Specifies the number of warmup "
+                   "requests to send before benchmarking. The default is 0, "
+                   "which means that no warmup requests will be sent.",
+                   18)
+            << std::endl;
   std::cerr << FormatMessage(
                    " --serial-sequences: Enables serial sequence mode "
                    "where a maximum of one request is outstanding at a time "