Merge pull request #5877 from unknownbrackets/gpu-minor

Improve God of War performance

Core/HLE/sceDisplay.cpp

@@ -672,7 +672,8 @@ u32 sceDisplaySetFramebuf(u32 topaddr, int linesize, int pixelformat, int sync)
 		const int FLIP_DELAY_MIN_FLIPS = 30;
 
 		u64 now = CoreTiming::GetTicks();
-		u64 expected = msToCycles(1000) / g_Config.iForceMaxEmulatedFPS;
+		// 1001 to account for NTSC timing (59.94 fps.)
+		u64 expected = msToCycles(1001) / g_Config.iForceMaxEmulatedFPS;
 		u64 actual = now - lastFlipCycles;
 		if (actual < expected - FLIP_DELAY_CYCLES_MIN) {
 			if (lastFlipsTooFrequent >= FLIP_DELAY_MIN_FLIPS) {

GPU/GLES/Framebuffer.cpp

@@ -1674,13 +1674,13 @@ void FramebufferManager::UpdateFromMemory(u32 addr, int size, bool safe) {
 		if (!Memory::IsValidAddress(displayFramebufPtr_))
 			return;
 
-		fbo_unbind();
-		currentRenderVfb_ = 0;
-
 		bool needUnbind = false;
 		for (size_t i = 0; i < vfbs_.size(); ++i) {
 			VirtualFramebuffer *vfb = vfbs_[i];
 			if (MaskedEqual(vfb->fb_address, addr)) {
+				fbo_unbind();
+				currentRenderVfb_ = 0;
+
 				vfb->dirtyAfterDisplay = true;
 				vfb->reallyDirtyAfterDisplay = true;
 				// TODO: This without the fbo_unbind() above would be better than destroying the FBO.

GPU/GLES/GLES_GPU.cpp

@@ -1698,14 +1698,22 @@ void GLES_GPU::DoBlockTransfer() {
 	// Do the copy! (Hm, if we detect a drawn video frame (see below) then we could maybe skip this?)
 	// Can use GetPointerUnchecked because we checked the addresses above. We could also avoid them
 	// entirely by walking a couple of pointers...
-	// GetPointerUnchecked crash in windows 64 bit of issue 2301
-	for (int y = 0; y < height; y++) {
-		u32 srcLineStartAddr = srcBasePtr + ((y + srcY) * srcStride + srcX) * bpp;
-		u32 dstLineStartAddr = dstBasePtr + ((y + dstY) * dstStride + dstX) * bpp;
-
+	if (srcStride == dstStride && width == srcStride) {
+		// Common case in God of War, let's do it all in one chunk.
+		u32 srcLineStartAddr = srcBasePtr + (srcY * srcStride + srcX) * bpp;
+		u32 dstLineStartAddr = dstBasePtr + (dstY * dstStride + dstX) * bpp;
 		const u8 *src = Memory::GetPointerUnchecked(srcLineStartAddr);
 		u8 *dst = Memory::GetPointerUnchecked(dstLineStartAddr);
-		memcpy(dst, src, width * bpp);
+		memcpy(dst, src, width * height * bpp);
+	} else {
+		for (int y = 0; y < height; y++) {
+			u32 srcLineStartAddr = srcBasePtr + ((y + srcY) * srcStride + srcX) * bpp;
+			u32 dstLineStartAddr = dstBasePtr + ((y + dstY) * dstStride + dstX) * bpp;
+
+			const u8 *src = Memory::GetPointerUnchecked(srcLineStartAddr);
+			u8 *dst = Memory::GetPointerUnchecked(dstLineStartAddr);
+			memcpy(dst, src, width * bpp);
+		}
 	}
 
 	// TODO: Notify all overlapping FBOs that they need to reload.