nx2: profile engine method calls and semaphore acquires; notes 140-141

EDEN_NX2_GPU_PROFILE now times engine method calls (the work a split GPU thread's back half
would run) and counts semaphore acquires and those that wait. Title 1: method calls ~343 ms
a second (6.9 million), front-half work (command parsing, macros) ~333 ms, acquires ~2,700 a
second of which ~220 wait. Notes 140 (macro compiler) and 141 (draw setup).

1 files changed, 3 insertions, 0 deletions

diff --git a/src/video_core/engines/puller.cpp b/src/video_core/engines/puller.cpp
index b607f20161..c9073f1eb0 100644
--- a/src/video_core/engines/puller.cpp
+++ b/src/video_core/engines/puller.cpp
@@ -11,6 +11,7 @@
 #include "common/logging.h"
 #include "common/settings.h"
 #include "core/core.h"
+#include "video_core/nx2_profile.h"
 #include "video_core/control/channel_state.h"
 #include "video_core/dma_pusher.h"
 #include "video_core/engines/fermi_2d.h"
@@ -217,6 +218,7 @@ void Puller::ProcessSemaphoreTriggerMethod(DmaPusher& dma_pusher) {
             LOG_ERROR(HW_GPU, "Invalid semaphore operation {}", static_cast<u32>(op));
             return;
         }
+        VideoCommon::Nx2Profile::Add(VideoCommon::Nx2Profile::Counter::AcquireChecks, 1);
         if (acquire.Matches(ReadSemaphoreWord(dma_pusher, acquire.address))) {
             return;
         }
@@ -229,6 +231,7 @@ void Puller::ProcessSemaphoreTriggerMethod(DmaPusher& dma_pusher) {
         dma_pusher.rasterizer->ReleaseFences();
         if (!waiters.Latched(this) && !acquire.Matches(ReadSemaphoreWord(dma_pusher, acquire.address))) {
             pending_acquire = acquire;
+            VideoCommon::Nx2Profile::Add(VideoCommon::Nx2Profile::Counter::AcquireWaits, 1);
         } else {
             waiters.Remove(this);
         }