diff --git a/rpcs3/Emu/Cell/PPUInterpreter.h b/rpcs3/Emu/Cell/PPUInterpreter.h
index 3f27b769ca..94cc1cde92 100644
--- a/rpcs3/Emu/Cell/PPUInterpreter.h
+++ b/rpcs3/Emu/Cell/PPUInterpreter.h
@@ -2104,6 +2104,7 @@ private:
 					Emu.GetSFuncManager()[CPU.GPR[11]]->name, CPU.GPR[3], CPU.PC);
 			}
 			break;
+		case 0x4: CPU.FastStop(); break;
 		case 0x22: UNK("HyperCall LV1"); break;
 		default: UNK(fmt::Format("Unknown sc: %x", sc_code));
 		}
@@ -2368,13 +2369,9 @@ private:
 	}
 	void LWARX(u32 rd, u32 ra, u32 rb)
 	{
-		const u64 addr = ra ? CPU.GPR[ra] + CPU.GPR[rb] : CPU.GPR[rb];
-
-		SMutexLockerR lock(reservation.mutex);
-		reservation.owner = lock.tid;
-		reservation.addr = addr;
-		reservation.size = 4;
-		reservation.data32 = CPU.GPR[rd] = Memory.Read32(addr);
+		CPU.R_ADDR = ra ? CPU.GPR[ra] + CPU.GPR[rb] : CPU.GPR[rb];
+		CPU.R_VALUE = (u32&)Memory[CPU.R_ADDR];
+		CPU.GPR[rd] = re32((u32)CPU.R_VALUE);
 	}
 	void LDX(u32 rd, u32 ra, u32 rb)
 	{
@@ -2523,13 +2520,9 @@ private:
 	}
 	void LDARX(u32 rd, u32 ra, u32 rb)
 	{
-		const u64 addr = ra ? CPU.GPR[ra] + CPU.GPR[rb] : CPU.GPR[rb];
-
-		SMutexLockerR lock(reservation.mutex);
-		reservation.owner = lock.tid;
-		reservation.addr = addr;
-		reservation.size = 8;
-		reservation.data64 = CPU.GPR[rd] = Memory.Read64(addr);
+		CPU.R_ADDR = ra ? CPU.GPR[ra] + CPU.GPR[rb] : CPU.GPR[rb];
+		CPU.R_VALUE = (u64&)Memory[CPU.R_ADDR];
+		CPU.GPR[rd] = re64(CPU.R_VALUE);
 	}
 	void DCBF(u32 ra, u32 rb)
 	{
@@ -2644,17 +2637,13 @@ private:
 	{
 		const u64 addr = ra ? CPU.GPR[ra] + CPU.GPR[rb] : CPU.GPR[rb];
 
-		SMutexLockerR lock(reservation.mutex);
-		if (lock.tid == reservation.owner && reservation.addr == addr && reservation.size == 4)
+		if (CPU.R_ADDR == addr)
 		{
-			// Memory.Write32(addr, CPU.GPR[rs]);
-			CPU.SetCR_EQ(0, InterlockedCompareExchange((volatile long*) (Memory + addr), re((u32) CPU.GPR[rs]), re(reservation.data32)) == re(reservation.data32));
-			reservation.clear();
+			CPU.SetCR_EQ(0, InterlockedCompareExchange((volatile u32*)(Memory + addr), re32((u32)CPU.GPR[rs]), (u32)CPU.R_VALUE) == (u32)CPU.R_VALUE);
 		}
 		else
 		{
 			CPU.SetCR_EQ(0, false);
-			if (lock.tid == reservation.owner) reservation.clear();
 		}
 	}
 	void STWX(u32 rs, u32 ra, u32 rb)
@@ -2705,17 +2694,13 @@ private:
 	{
 		const u64 addr = ra ? CPU.GPR[ra] + CPU.GPR[rb] : CPU.GPR[rb];
 
-		SMutexLockerR lock(reservation.mutex);
-		if (lock.tid == reservation.owner && reservation.addr == addr && reservation.size == 8)
+		if (CPU.R_ADDR == addr)
 		{
-			// Memory.Write64(addr, CPU.GPR[rs]);
-			CPU.SetCR_EQ(0, InterlockedCompareExchange64((volatile long long*)(Memory + addr), re(CPU.GPR[rs]), re(reservation.data64)) == re(reservation.data64));
-			reservation.clear();
+			CPU.SetCR_EQ(0, InterlockedCompareExchange((volatile u64*)(Memory + addr), re64(CPU.GPR[rs]), CPU.R_VALUE) == CPU.R_VALUE);
 		}
 		else
 		{
 			CPU.SetCR_EQ(0, false);
-			if (lock.tid == reservation.owner) reservation.clear();
 		}
 	}
 	void STBX(u32 rs, u32 ra, u32 rb)
diff --git a/rpcs3/Emu/Cell/PPUThread.cpp b/rpcs3/Emu/Cell/PPUThread.cpp
index 0b039e4886..c4907dc6e3 100644
--- a/rpcs3/Emu/Cell/PPUThread.cpp
+++ b/rpcs3/Emu/Cell/PPUThread.cpp
@@ -219,3 +219,34 @@ int FPRdouble::Cmp(PPCdouble a, PPCdouble b)
 
 	return CR_SO;
 }
+
+u64 PPUThread::FastCall(u64 addr, u64 arg1, u64 arg2, u64 arg3, u64 arg4, u64 arg5, u64 arg6, u64 arg7, u64 arg8)
+{
+	auto old_status = m_status;
+	auto old_PC = PC;
+	auto old_LR = LR;
+
+	PC = addr;
+	GPR[3] = arg1;
+	GPR[4] = arg2;
+	GPR[5] = arg3;
+	GPR[6] = arg4;
+	GPR[7] = arg5;
+	GPR[8] = arg6;
+	GPR[9] = arg7;
+	GPR[10] = arg8;
+	LR = Emu.m_ppu_thr_stop;
+
+	Task();
+
+	LR = old_LR;
+	PC = old_PC;
+	m_status = old_status;
+
+	return GPR[3];
+}
+
+void PPUThread::FastStop()
+{
+	m_status = Stopped;
+}
\ No newline at end of file
diff --git a/rpcs3/Emu/Cell/PPUThread.h b/rpcs3/Emu/Cell/PPUThread.h
index 1fe6fffd01..7d8cd62197 100644
--- a/rpcs3/Emu/Cell/PPUThread.h
+++ b/rpcs3/Emu/Cell/PPUThread.h
@@ -607,6 +607,9 @@ public:
 
 	u64 cycle;
 
+	u64 R_ADDR; // reservation address
+	u64 R_VALUE; // reservation value (BE)
+
 public:
 	PPUThread();
 	virtual ~PPUThread();
@@ -839,6 +842,8 @@ public:
 public:
 	virtual void InitRegs(); 
 	virtual u64 GetFreeStackSize() const;
+	u64 FastCall(u64 addr, u64 arg1 = 0, u64 arg2 = 0, u64 arg3 = 0, u64 arg4 = 0, u64 arg5 = 0, u64 arg6 = 0, u64 arg7 = 0, u64 arg8 = 0);
+	void FastStop();
 
 protected:
 	virtual void DoReset() override;
diff --git a/rpcs3/Emu/SysCalls/Modules/cellSysutil.cpp b/rpcs3/Emu/SysCalls/Modules/cellSysutil.cpp
index 5bacd9f287..e3e538286d 100644
--- a/rpcs3/Emu/SysCalls/Modules/cellSysutil.cpp
+++ b/rpcs3/Emu/SysCalls/Modules/cellSysutil.cpp
@@ -550,7 +550,7 @@ int cellAudioOutGetNumberOfDevice(u32 audioOut)
 
 int cellAudioOutGetDeviceInfo(u32 audioOut, u32 deviceIndex, mem_ptr_t<CellAudioOutDeviceInfo> info)
 {
-	cellSysutil->Todo("Unimplemented function: cellAudioOutGetDeviceInfo(audioOut=%u, deviceIndex=%u, info_addr=0x%x)",
+	cellSysutil->Todo("cellAudioOutGetDeviceInfo(audioOut=%u, deviceIndex=%u, info_addr=0x%x)",
 		audioOut, deviceIndex, info.GetAddr());
 
 	if(deviceIndex) return CELL_AUDIO_OUT_ERROR_DEVICE_NOT_FOUND;
diff --git a/rpcs3/Emu/SysCalls/SysCalls.cpp b/rpcs3/Emu/SysCalls/SysCalls.cpp
index da13269249..ddea3cdd5b 100644
--- a/rpcs3/Emu/SysCalls/SysCalls.cpp
+++ b/rpcs3/Emu/SysCalls/SysCalls.cpp
@@ -45,7 +45,7 @@ static func_caller* sc_table[kSyscallTableLength] =
 	bind_func(sys_process_kill),                            //19  (0x013)
 	null_func,                                              //20  (0x014)  UNS
 	null_func,//bind_func(_sys_process_spawn),              //21  (0x015)  DBG
-	bind_func(sys_process_exit),                            //22  (0x016)
+	null_func,//bind_func(sys_process_exit),                //22  (0x016)
 	bind_func(sys_process_wait_for_child2),                 //23  (0x017)  DBG
 	null_func,//bind_func(),                                //24  (0x018)  DBG
 	bind_func(sys_process_get_sdk_version),                 //25  (0x019)
diff --git a/rpcs3/Emu/System.cpp b/rpcs3/Emu/System.cpp
index 0745e1c49c..ec21d85dbd 100644
--- a/rpcs3/Emu/System.cpp
+++ b/rpcs3/Emu/System.cpp
@@ -324,11 +324,17 @@ void Emulator::Load()
 		m_ppu_thr_exit = Memory.MainMem.AllocAlign(4 * 4);
 
 		mem32_ptr_t ppu_thr_exit_data(m_ppu_thr_exit);
-		ppu_thr_exit_data += ADDI(3, 0, 0);
+		//ppu_thr_exit_data += ADDI(3, 0, 0); // why it kills return value (GPR[3]) ?
 		ppu_thr_exit_data += ADDI(11, 0, 41);
 		ppu_thr_exit_data += SC(2);
 		ppu_thr_exit_data += BCLR(0x10 | 0x04, 0, 0, 0);
 
+		m_ppu_thr_stop = Memory.MainMem.AllocAlign(2 * 4);
+
+		mem32_ptr_t ppu_thr_stop_data(m_ppu_thr_stop);
+		ppu_thr_stop_data += SC(4);
+		ppu_thr_exit_data += BCLR(0x10 | 0x04, 0, 0, 0);
+
 		Memory.Write64(Memory.PRXMem.AllocAlign(0x10000), 0xDEADBEEFABADCAFE);
 	}
 	break;
@@ -422,11 +428,6 @@ void Emulator::Stop()
 			break;
 		}
 		std::this_thread::sleep_for(std::chrono::milliseconds(1));
-		if (counter++ > 3000)
-		{
-			LOG_ERROR(HLE, "%d threads not stopped (timeout)", (u32)(g_thread_count - uncounted));
-			break;
-		}
 	}
 
 	m_rsx_callback = 0;
diff --git a/rpcs3/Emu/System.h b/rpcs3/Emu/System.h
index f2500878ac..66a087589e 100644
--- a/rpcs3/Emu/System.h
+++ b/rpcs3/Emu/System.h
@@ -106,6 +106,7 @@ public:
 	std::string m_path;
 	std::string m_elf_path;
 	std::string m_title_id;
+	u32 m_ppu_thr_stop;
 	s32 m_sdk_version;
 
 	Emulator();