diff --git uppsrc/Core/Cpu.cpp uppsrc/Core/Cpu.cpp
--- uppsrc/Core/Cpu.cpp
+++ uppsrc/Core/Cpu.cpp
@@ -149,7 +149,10 @@ void EndianSwap(int *v, size_t count) ENDIAN_SWAP
 void EndianSwap(int64 *v, size_t count) ENDIAN_SWAP
 void EndianSwap(uint64 *v, size_t count) ENDIAN_SWAP
 
-#ifdef CPU_X86
+#if defined(CPU_X86) && defined(__SSE2__)
+
+#include <emmintrin.h>
+
 void huge_memsetd(void *p, dword c, int len)
 { // bypasses the cache, good for >4MB
 	dword *t = (dword *)p;
@@ -173,6 +176,48 @@ void huge_memsetd(void *p, dword c, int len)
 	while(len--)
 		*t++ = c;
 }
+
+void memsetd(void *p, dword data, int len)
+{
+	dword *t = (dword *)p;
+	if(len < 4) {
+		if(len & 2) {
+			t[0] = t[1] = data;
+			t += 2;
+		}
+		if(len & 1)
+			t[0] = data;
+		return;
+	}
+
+	__m128i val4 = _mm_set1_epi32(data);
+	auto Set4 = [&](int at) { _mm_storeu_si128((__m128i *)(t + at), val4); };
+
+	Set4(len - 4); // fill tail
+	if(len >= 32) {
+		if(len >= 1024*1024) { // for really huge data, bypass the cache
+			huge_memsetd(t, data, len);
+			return;
+		}
+		const dword *e = t + len - 32;
+		do {
+			Set4(0); Set4(4); Set4(8); Set4(12);
+			Set4(16); Set4(20); Set4(24); Set4(28);
+			t += 32;
+		}
+		while(t <= e);
+	}
+	if(len & 16) {
+		Set4(0); Set4(4); Set4(8); Set4(12);
+		t += 16;
+	}
+	if(len & 8) {
+		Set4(0); Set4(4);
+		t += 8;
+	}
+	if(len & 4)
+		Set4(0);
+}
 #endif
 
 }
diff --git uppsrc/Core/Ops.h uppsrc/Core/Ops.h
--- uppsrc/Core/Ops.h
+++ uppsrc/Core/Ops.h
@@ -342,54 +342,9 @@ inline bool FitsInInt64(double x)
 	return x >= -9223372036854775808.0 && x < 9223372036854775808.0;
 }
 
-#ifdef CPU_X86
-
-#include <smmintrin.h>
-
+#if defined(CPU_X86) && defined(__SSE2__)
 void huge_memsetd(void *p, dword data, int len);
-
-inline
-void memsetd(void *p, dword data, int len)
-{
-	dword *t = (dword *)p;
-	if(len < 4) {
-		if(len & 2) {
-			t[0] = t[1] = data;
-			t += 2;
-		}
-		if(len & 1)
-			t[0] = data;
-		return;
-	}
-
-	__m128i val4 = _mm_set1_epi32(data);
-	auto Set4 = [&](int at) { _mm_storeu_si128((__m128i *)(t + at), val4); };
-
-	Set4(len - 4); // fill tail
-	if(len >= 32) {
-		if(len >= 1024*1024) { // for really huge data, bypass the cache
-			huge_memsetd(t, data, len);
-			return;
-		}
-		const dword *e = t + len - 32;
-		do {
-			Set4(0); Set4(4); Set4(8); Set4(12);
-			Set4(16); Set4(20); Set4(24); Set4(28);
-			t += 32;
-		}
-		while(t <= e);
-	}
-	if(len & 16) {
-		Set4(0); Set4(4); Set4(8); Set4(12);
-		t += 16;
-	}
-	if(len & 8) {
-		Set4(0); Set4(4);
-		t += 8;
-	}
-	if(len & 4)
-		Set4(0);
-}
+void memsetd(void *p, dword data, int len);
 #else
 inline
 void memsetd(void *p, RGBA c, int len)