diff --git a/src/crypto/asm/cn_main_loop.S b/src/crypto/asm/cn_main_loop.S
index 417fd414..e9ac64f5 100644
--- a/src/crypto/asm/cn_main_loop.S
+++ b/src/crypto/asm/cn_main_loop.S
@@ -19,7 +19,10 @@ FN_PREFIX(cnv2_mainloop_ivybridge_asm):
 	#include "cn2/cnv2_main_loop_ivybridge.inc"
 	add rsp, 48
 	ret 0
-	nop;nop;nop;nop;
+	nop
+	nop
+	nop
+	nop
 
 ALIGN 16
 FN_PREFIX(cnv2_mainloop_ryzen_asm):
@@ -28,7 +31,10 @@ FN_PREFIX(cnv2_mainloop_ryzen_asm):
 	#include "cn2/cnv2_main_loop_ryzen.inc"
 	add rsp, 48
 	ret 0
-	nop;nop;nop;nop;
+	nop
+	nop
+	nop
+	nop
 
 ALIGN 16
 FN_PREFIX(cnv2_mainloop_bulldozer_asm):
@@ -37,7 +43,10 @@ FN_PREFIX(cnv2_mainloop_bulldozer_asm):
 	#include "cn2/cnv2_main_loop_bulldozer.inc"
 	add rsp, 48
 	ret 0
-	nop;nop;nop;nop;
+	nop
+	nop
+	nop
+	nop
 
 ALIGN 16
 FN_PREFIX(cnv2_double_mainloop_sandybridge_asm):
@@ -47,4 +56,7 @@ FN_PREFIX(cnv2_double_mainloop_sandybridge_asm):
 	#include "cn2/cnv2_double_main_loop_sandybridge.inc"
 	add rsp, 48
 	ret 0
-	nop;nop;nop;nop;
+	nop
+	nop
+	nop
+	nop
diff --git a/src/crypto/asm/cn_main_loop.asm b/src/crypto/asm/cn_main_loop.asm
index 9d4cede0..9c8a6ea9 100644
--- a/src/crypto/asm/cn_main_loop.asm
+++ b/src/crypto/asm/cn_main_loop.asm
@@ -8,28 +8,40 @@ ALIGN 64
 cnv2_mainloop_ivybridge_asm PROC
 	INCLUDE cn2/cnv2_main_loop_ivybridge.inc
 	ret 0
-	nop;nop;nop;nop;
+	nop
+	nop
+	nop
+	nop
 cnv2_mainloop_ivybridge_asm ENDP
 
 ALIGN 64
 cnv2_mainloop_ryzen_asm PROC
 	INCLUDE cn2/cnv2_main_loop_ryzen.inc
 	ret 0
-	nop;nop;nop;nop;
+	nop
+	nop
+	nop
+	nop
 cnv2_mainloop_ryzen_asm ENDP
 
 ALIGN 64
 cnv2_mainloop_bulldozer_asm PROC
 	INCLUDE cn2/cnv2_main_loop_bulldozer.inc
 	ret 0
-	nop;nop;nop;nop;
+	nop
+	nop
+	nop
+	nop
 cnv2_mainloop_bulldozer_asm ENDP
 
 ALIGN 64
 cnv2_double_mainloop_sandybridge_asm PROC
 	INCLUDE cn2/cnv2_double_main_loop_sandybridge.inc
 	ret 0
-	nop;nop;nop;nop;
+	nop
+	nop
+	nop
+	nop
 cnv2_double_mainloop_sandybridge_asm ENDP
 
 _TEXT_CNV2_MAINLOOP ENDS
diff --git a/src/crypto/asm/win64/cn_main_loop.S b/src/crypto/asm/win64/cn_main_loop.S
index 4caaa1a1..ea5a63b8 100644
--- a/src/crypto/asm/win64/cn_main_loop.S
+++ b/src/crypto/asm/win64/cn_main_loop.S
@@ -10,22 +10,34 @@ ALIGN 16
 cnv2_mainloop_ivybridge_asm:
 	#include "../cn2/cnv2_main_loop_ivybridge.inc"
 	ret 0
-	nop;nop;nop;nop;
+	nop
+	nop
+	nop
+	nop
 
 ALIGN 16
 cnv2_mainloop_ryzen_asm:
 	#include "../cn2/cnv2_main_loop_ryzen.inc"
 	ret 0
-	nop;nop;nop;nop;
+	nop
+	nop
+	nop
+	nop
 
 ALIGN 16
 cnv2_mainloop_bulldozer_asm:
 	#include "../cn2/cnv2_main_loop_bulldozer.inc"
 	ret 0
-	nop;nop;nop;nop;
+	nop
+	nop
+	nop
+	nop
 
 ALIGN 16
 cnv2_double_mainloop_sandybridge_asm:
 	#include "../cn2/cnv2_double_main_loop_sandybridge.inc"
 	ret 0
-	nop;nop;nop;nop;
+	nop
+	nop
+	nop
+	nop
diff --git a/src/crypto/asm/win64/cn_main_loop.asm b/src/crypto/asm/win64/cn_main_loop.asm
index 9d4cede0..9c8a6ea9 100644
--- a/src/crypto/asm/win64/cn_main_loop.asm
+++ b/src/crypto/asm/win64/cn_main_loop.asm
@@ -8,28 +8,40 @@ ALIGN 64
 cnv2_mainloop_ivybridge_asm PROC
 	INCLUDE cn2/cnv2_main_loop_ivybridge.inc
 	ret 0
-	nop;nop;nop;nop;
+	nop
+	nop
+	nop
+	nop
 cnv2_mainloop_ivybridge_asm ENDP
 
 ALIGN 64
 cnv2_mainloop_ryzen_asm PROC
 	INCLUDE cn2/cnv2_main_loop_ryzen.inc
 	ret 0
-	nop;nop;nop;nop;
+	nop
+	nop
+	nop
+	nop
 cnv2_mainloop_ryzen_asm ENDP
 
 ALIGN 64
 cnv2_mainloop_bulldozer_asm PROC
 	INCLUDE cn2/cnv2_main_loop_bulldozer.inc
 	ret 0
-	nop;nop;nop;nop;
+	nop
+	nop
+	nop
+	nop
 cnv2_mainloop_bulldozer_asm ENDP
 
 ALIGN 64
 cnv2_double_mainloop_sandybridge_asm PROC
 	INCLUDE cn2/cnv2_double_main_loop_sandybridge.inc
 	ret 0
-	nop;nop;nop;nop;
+	nop
+	nop
+	nop
+	nop
 cnv2_double_mainloop_sandybridge_asm ENDP
 
 _TEXT_CNV2_MAINLOOP ENDS
diff --git a/src/workers/CpuThread.cpp b/src/workers/CpuThread.cpp
index 8f3457dc..cf366860 100644
--- a/src/workers/CpuThread.cpp
+++ b/src/workers/CpuThread.cpp
@@ -57,10 +57,15 @@ xmrig::CpuThread::CpuThread(size_t index, Algo algorithm, AlgoVariant av, Multiw
 
 #ifndef XMRIG_NO_ASM
 template<typename T, typename U>
-static void patchCode(T& dst, U src, const uint32_t iterations, const uint32_t mask)
+static void patchCode(T dst, U src, const uint32_t iterations, const uint32_t mask)
 {
     const uint8_t* p = reinterpret_cast<const uint8_t*>(src);
 
+    // Workaround for Visual Studio placing trampoline in debug builds
+    if (p[0] == 0xE9) {
+        p += *(int32_t*)(p + 1) + 5;
+    }
+
     size_t size = 0;
     while (*(uint32_t*)(p + size) != 0x90909090) {
         ++size;