Instrumentation on the cycle/update/render performance. So far looks like update and then render are the worst.

src/chip8.cpp

@@ -94,8 +94,10 @@ void Chip8::Cycle() {
   // Increment the PC before we execute anything
   pc += 2;
 
+  auto func_ptr = table[(opcode & 0xF000u) >> 12u];
+
   // decode and Execute
-  ((*this).*(table[(opcode & 0xF000u) >> 12u]))();
+  (this->*(func_ptr))();
 
   // Decrement the delay timer if it's been set
   if(delayTimer > 0) {

src/main.cpp

@@ -94,6 +94,7 @@ int main(int argc, char* argv[]) {
   Display display{window, {0, 0}, {1000, 1000}};
   Stats cycle_stats;
   Stats render_stats;
+  Stats update_stats;
 
   while (display.active()) {
     auto currentTime = std::chrono::high_resolution_clock::now();
@@ -107,12 +108,16 @@ int main(int argc, char* argv[]) {
         cycle_stats.sample_time(t);
     }
 
-    auto t = render_stats.time_start();
+    auto ut = update_stats.time_start();
     display.update(chip8);
+    update_stats.sample_time(ut);
+
+    auto rt = render_stats.time_start();
     display.render();
-    render_stats.sample_time(t);
+    render_stats.sample_time(rt);
   }
 
-  cycle_stats.dump("cycle times");
-  render_stats.dump("render times");
+  cycle_stats.dump("CYCLE TIMES");
+  update_stats.dump("UPDATE TIMES");
+  render_stats.dump("RENDER TIMES");
 }