From ca9e7c2c2dffaf0f5b1832955d1704c6c1ba16aa Mon Sep 17 00:00:00 2001
From: Tatsuhiro Tsujikawa <tatsuhiro.t@gmail.com>
Date: Fri, 9 Oct 2015 21:36:28 +0900
Subject: [PATCH] h2load: Start thread execution using conditional variable

When thread is created, we pause them.  After all threads are created,
master thread sends signal to all worker threads and let them start to
benchmark.  This will make thread start almost at the same time since
we can avoid thread creation overhead.  It also exclude thread
creating time from benchmark time.  We also simplified thread creation
routine, and now we always use dedicted worker thread to issue
requests even if -t1.
---
 src/h2load.cc | 98 +++++++++++++++++++++++++++------------------------
 1 file changed, 52 insertions(+), 46 deletions(-)

diff --git a/src/h2load.cc b/src/h2load.cc
index a1113105..924e8c20 100644
--- a/src/h2load.cc
+++ b/src/h2load.cc
@@ -1254,6 +1254,24 @@ void read_script_from_file(std::istream &infile,
 }
 } // namespace
 
+namespace {
+std::unique_ptr<Worker> create_worker(uint32_t id, SSL_CTX *ssl_ctx,
+                                      size_t nreqs, size_t nclients,
+                                      size_t rate) {
+  std::stringstream rate_report;
+  if (config.is_rate_mode() && nclients > rate) {
+    rate_report << "Up to " << rate << " client(s) will be created every "
+                << std::setprecision(3) << config.rate_period << " seconds. ";
+  }
+
+  std::cout << "spawning thread #" << id << ": " << nclients
+            << " total client(s). " << rate_report.str() << nreqs
+            << " total requests" << std::endl;
+
+  return make_unique<Worker>(id, ssl_ctx, nreqs, nclients, rate, &config);
+}
+} // namespace
+
 namespace {
 void print_version(std::ostream &out) {
   out << "h2load nghttp2/" NGHTTP2_VERSION << std::endl;
@@ -1905,6 +1923,12 @@ int main(int argc, char **argv) {
 
   resolve_host();
 
+  std::cout << "starting benchmark..." << std::endl;
+
+  std::vector<std::unique_ptr<Worker>> workers;
+  workers.reserve(config.nthreads);
+
+#ifndef NOTHREADS
   size_t nreqs_per_thread = 0;
   ssize_t nreqs_rem = 0;
 
@@ -1919,16 +1943,12 @@ int main(int argc, char **argv) {
   size_t rate_per_thread = config.rate / config.nthreads;
   ssize_t rate_per_thread_rem = config.rate % config.nthreads;
 
-  std::cout << "starting benchmark..." << std::endl;
+  std::mutex mu;
+  std::condition_variable cv;
+  auto ready = false;
 
-  auto start = std::chrono::steady_clock::now();
-
-  std::vector<std::unique_ptr<Worker>> workers;
-
-  workers.reserve(config.nthreads);
-#ifndef NOTHREADS
   std::vector<std::future<void>> futures;
-  for (size_t i = 0; i < config.nthreads - 1; ++i) {
+  for (size_t i = 0; i < config.nthreads; ++i) {
     auto rate = rate_per_thread;
     if (rate_per_thread_rem > 0) {
       --rate_per_thread_rem;
@@ -1955,55 +1975,41 @@ int main(int argc, char **argv) {
       }
     }
 
-    std::stringstream rate_report;
-    if (config.is_rate_mode() && nclients > rate) {
-      rate_report << "Up to " << rate << " client(s) will be created every "
-                  << std::setprecision(3) << config.rate_period << " seconds. ";
-    }
-
-    std::cout << "spawning thread #" << i << ": " << nclients
-              << " total client(s). " << rate_report.str() << nreqs
-              << " total requests" << std::endl;
-
-    workers.push_back(
-        make_unique<Worker>(i, ssl_ctx, nreqs, nclients, rate, &config));
+    workers.push_back(create_worker(i, ssl_ctx, nreqs, nclients, rate));
     auto &worker = workers.back();
     futures.push_back(
-        std::async(std::launch::async, [&worker]() { worker->run(); }));
+        std::async(std::launch::async, [&worker, &mu, &cv, &ready]() {
+          {
+            std::unique_lock<std::mutex> ulk(mu);
+            cv.wait(ulk, [&ready] { return ready; });
+          }
+          worker->run();
+        }));
   }
-#endif // NOTHREADS
-
-  assert(rate_per_thread_rem == 0);
-  assert(nclients_rem == 0);
-  assert(nreqs_rem == 0);
 
   {
-    auto rate_last = rate_per_thread;
-    auto nclients_last = nclients_per_thread;
-    auto nreqs_last =
-        config.timing_script ? config.nreqs * nclients_last : nreqs_per_thread;
-    std::stringstream rate_report;
-    if (config.is_rate_mode() && nclients_last > rate_last) {
-      rate_report << "Up to " << rate_last
-                  << " client(s) will be created every " << std::setprecision(3)
-                  << config.rate_period << " seconds. ";
-    }
-
-    std::cout << "spawning thread #" << (config.nthreads - 1) << ": "
-              << nclients_last << " total client(s). " << rate_report.str()
-              << nreqs_last << " total requests" << std::endl;
-
-    workers.push_back(make_unique<Worker>(config.nthreads - 1, ssl_ctx,
-                                          nreqs_last, nclients_last, rate_last,
-                                          &config));
+    std::lock_guard<std::mutex> lg(mu);
+    ready = true;
+    cv.notify_all();
   }
 
-  workers.back()->run();
+  auto start = std::chrono::steady_clock::now();
 
-#ifndef NOTHREADS
   for (auto &fut : futures) {
     fut.get();
   }
+
+#else  // NOTHREADS
+  auto rate = config.rate;
+  auto nclients = config.nclients;
+  auto nreqs =
+      config.timing_script ? config.nreqs * config.nclients : config.nreqs;
+
+  workers.push_back(create_worker(0, ssl_ctx, nreqs, nclients, rate));
+
+  auto start = std::chrono::steady_clock::now();
+
+  workers.back()->run();
 #endif // NOTHREADS
 
   auto end = std::chrono::steady_clock::now();