diff --git a/XenonRecomp/recompiler.cpp b/XenonRecomp/recompiler.cpp
index 27f1a3e..6268df9 100644
--- a/XenonRecomp/recompiler.cpp
+++ b/XenonRecomp/recompiler.cpp
@@ -2457,6 +2457,11 @@ bool Recompiler::Recompile(
         println("\t_mm_store_si128((__m128i*){}.u8, _mm_or_si128(_mm_andnot_si128(_mm_load_si128((__m128i*){}.u8), _mm_load_si128((__m128i*){}.u8)), _mm_and_si128(_mm_load_si128((__m128i*){}.u8), _mm_load_si128((__m128i*){}.u8))));", v(insn.operands[0]), v(insn.operands[3]), v(insn.operands[1]), v(insn.operands[3]), v(insn.operands[2]));
         break;
 
+    case PPC_INST_VSL:
+        println("\t_mm_store_si128((__m128i*){}.u8, _mm_vsl(_mm_load_si128((__m128i*){}.u8), _mm_load_si128((__m128i*){}.u8)));", 
+            v(insn.operands[0]), v(insn.operands[1]), v(insn.operands[2]));
+        break;
+
     case PPC_INST_VSLB:
         // TODO: vectorize
         for (size_t i = 0; i < 16; i++)
diff --git a/XenonUtils/ppc_context.h b/XenonUtils/ppc_context.h
index 2225748..7b30689 100644
--- a/XenonUtils/ppc_context.h
+++ b/XenonUtils/ppc_context.h
@@ -708,4 +708,20 @@ inline __m128i _mm_vctuxs(__m128 src1)
     return _mm_blendv_epi8(result, _mm_set1_epi32(-1), _mm_castps_si128(saturate_mask));
 }
 
+inline __m128i _mm_vsl(__m128i a, __m128i b)
+{
+    // Extract shift count from last byte of b (accounting for endianness)
+    uint32_t shift = _mm_extract_epi8(b, 15) & 0x7;
+    
+    if (shift == 0) return a;
+    
+    // Shift left by bits
+    __m128i shifted = _mm_or_si128(
+        _mm_slli_epi64(a, shift),
+        _mm_srli_epi64(_mm_slli_si128(a, 8), 64 - shift)
+    );
+    
+    return shifted;
+}
+
 #endif