diff --git a/src/nfa/arm/truffle.hpp b/src/nfa/arm/truffle.hpp
index 6f9a61fc9..7d2cfa6b9 100644
--- a/src/nfa/arm/truffle.hpp
+++ b/src/nfa/arm/truffle.hpp
@@ -230,6 +230,34 @@ svuint8_t blockSingleMaskWide(svuint8_t shuf_mask_lo_highclear, svuint8_t shuf_m
 #endif //HAVE_SVE2
 #endif //HAVE_SVE
 
+#if defined(CAN_USE_WIDE_TRUFFLE) && !defined(HAVE_SVE)
+/* require wide truffle compilation. The 256b mask is split between the two parameters */
+
+template <uint16_t S>
+static really_inline
+const SuperVector<S> blockSingleMaskWide(SuperVector<S> shuf_mask_lo_highclear, SuperVector<S> shuf_mask_lo_highset, SuperVector<S> chars) {
+    chars.print8("chars");
+    shuf_mask_lo_highclear.print8("shuf_mask_lo_highclear");
+
+    uint8x16x2_t tbl = {{shuf_mask_lo_highclear.u.u8x16[0], shuf_mask_lo_highset.u.u8x16[0]}};
+
+    SuperVector<S> idx = chars & SuperVector<S>::dup_u8(31);
+    idx.print8("idx");
+    SuperVector<S> byte_select = SuperVector<S>(vqtbl2q_u8(tbl, idx.u.u8x16[0]));
+    byte_select.print8("byte_select");
+
+    SuperVector<S> bits = chars.template vshr_8_imm<5>();
+    bits.print8("bits");
+    SuperVector<S> bit_select = SuperVector<S>(
+        vshlq_u8(vdupq_n_u8(1), vreinterpretq_s8_u8(bits.u.u8x16[0])));
+    bit_select.print8("bit_select");
+
+    SuperVector<S> res = bit_select & byte_select;
+    res.print8("bit_select & byte_select");
+    return !res.eq(SuperVector<S>::Zeroes());
+}
+
+#endif
 /* require normal truffle compilation. The 256b mask is split between the two parameters */
 template <uint16_t S>
 static really_inline
diff --git a/src/nfa/truffle.cpp b/src/nfa/truffle.cpp
index df3a4a3f5..054cb7f7e 100644
--- a/src/nfa/truffle.cpp
+++ b/src/nfa/truffle.cpp
@@ -58,7 +58,15 @@ const u8 *rtruffleExecWide(m256 mask, const u8 *buf,
     }
 }
 #else // HAVE_SVE
-#error "Wide truffle enabled for the target architecture but no implementation found"
+const u8 *truffleExecWide(m256 mask, const u8 *buf,
+                      const u8 *buf_end) {
+    return truffleExecReal<VECTORSIZE, true>(mask.lo, mask.hi, buf, buf_end);
+}
+
+const u8 *rtruffleExecWide(m256 mask, const u8 *buf,
+                       const u8 *buf_end) {
+    return rtruffleExecReal<VECTORSIZE, true>(mask.lo, mask.hi, buf, buf_end);
+}
 #endif // HAVE_SVE
 #endif // CAN_USE_WIDE_TRUFFLE
 
@@ -84,11 +92,11 @@ const u8 *rtruffleExec(m128 mask_lo, m128 mask_hi, const u8 *buf,
 #else
 const u8 *truffleExec(m128 mask_lo, m128 mask_hi, const u8 *buf,
                       const u8 *buf_end) {
-    return truffleExecReal<VECTORSIZE>(mask_lo, mask_hi, buf, buf_end);
+    return truffleExecReal<VECTORSIZE, false>(mask_lo, mask_hi, buf, buf_end);
 }
 
 const u8 *rtruffleExec(m128 mask_lo, m128 mask_hi, const u8 *buf,
                        const u8 *buf_end) {
-    return rtruffleExecReal<VECTORSIZE>(mask_lo, mask_hi, buf, buf_end);
+    return rtruffleExecReal<VECTORSIZE, false>(mask_lo, mask_hi, buf, buf_end);
 }
 #endif //HAVE_SVE
diff --git a/src/nfa/truffle_simd.hpp b/src/nfa/truffle_simd.hpp
index 8a2a06280..4b508ecf4 100644
--- a/src/nfa/truffle_simd.hpp
+++ b/src/nfa/truffle_simd.hpp
@@ -54,6 +54,11 @@ static really_inline
 svuint8_t blockSingleMaskWide(svuint8_t shuf_mask_lo_highclear, svuint8_t shuf_mask_lo_highset, svuint8_t chars);
 #endif //HAVE_SVE2
 #else
+#ifdef CAN_USE_WIDE_TRUFFLE
+template <uint16_t S>
+static really_inline
+const SuperVector<S> blockSingleMaskWide(SuperVector<S> shuf_mask_lo_highclear, SuperVector<S> shuf_mask_lo_highset, SuperVector<S> chars);
+#endif
 template <uint16_t S>
 static really_inline
 const SuperVector<S> blockSingleMask(SuperVector<S> shuf_mask_lo_highclear, SuperVector<S> shuf_mask_lo_highset, SuperVector<S> chars);
@@ -95,6 +100,7 @@ const u8 *scanBlock(svuint8_t shuf_mask_lo_highclear, svuint8_t shuf_mask_lo_hig
 #else
             DEBUG_PRINTF("Wide Truffle is not supported with 128b vectors unless SVE2 is enabled");
             assert(false);
+            return nullptr;
 #endif
         } else {
             result_mask = blockSingleMaskWide32(shuf_mask_lo_highclear, chars);
@@ -260,14 +266,24 @@ const u8 *rtruffleExecSVE(m256 shuf_mask_32, const u8 *buf, const u8 *buf_end){
     return buf - 1;
 }
 #else
-template <uint16_t S>
+template <uint16_t S, bool is_wide>
 static really_inline
 const u8 *fwdBlock(SuperVector<S> shuf_mask_lo_highclear, SuperVector<S> shuf_mask_lo_highset, SuperVector<S> chars, const u8 *buf) {
-    SuperVector<S> res = blockSingleMask(shuf_mask_lo_highclear, shuf_mask_lo_highset, chars);
-    return first_zero_match_inverted<S>(buf, res);
+    if constexpr (is_wide) {
+#ifdef CAN_USE_WIDE_TRUFFLE
+        SuperVector<S> res = blockSingleMaskWide(shuf_mask_lo_highclear, shuf_mask_lo_highset, chars);
+        return first_zero_match_inverted<S>(buf, res);
+#else
+        assert(false);
+        return nullptr;
+#endif
+    } else {
+        SuperVector<S> res = blockSingleMask(shuf_mask_lo_highclear, shuf_mask_lo_highset, chars);
+        return first_zero_match_inverted<S>(buf, res);
+    }
 }
 
-template <uint16_t S>
+template <uint16_t S, bool is_wide>
 const u8 *truffleExecReal(const m128 &shuf_mask_lo_highclear, m128 shuf_mask_lo_highset, const u8 *buf, const u8 *buf_end) {
     assert(buf && buf_end);
     assert(buf < buf_end);
@@ -289,7 +305,7 @@ const u8 *truffleExecReal(const m128 &shuf_mask_lo_highclear, m128 shuf_mask_lo_
         if (!ISALIGNED_N(d, S)) {
             SuperVector<S> chars = SuperVector<S>::loadu(d);
             const u8 *dup = ROUNDUP_PTR(d, S);
-            rv = fwdBlock(wide_shuf_mask_lo_highclear, wide_shuf_mask_lo_highset, chars, d);
+            rv = fwdBlock<S, is_wide>(wide_shuf_mask_lo_highclear, wide_shuf_mask_lo_highset, chars, d);
             if (rv && rv < dup) return rv;
             d = dup;
         }
@@ -298,7 +314,7 @@ const u8 *truffleExecReal(const m128 &shuf_mask_lo_highclear, m128 shuf_mask_lo_
             __builtin_prefetch(d + 16*64);
             DEBUG_PRINTF("d %p \n", d);
             SuperVector<S> chars = SuperVector<S>::load(d);
-            rv = fwdBlock(wide_shuf_mask_lo_highclear, wide_shuf_mask_lo_highset, chars, d);
+            rv = fwdBlock<S, is_wide>(wide_shuf_mask_lo_highclear, wide_shuf_mask_lo_highset, chars, d);
             if (rv) return rv;
             d += S;
         }
@@ -317,7 +333,7 @@ const u8 *truffleExecReal(const m128 &shuf_mask_lo_highclear, m128 shuf_mask_lo_
           chars = SuperVector<S>::loadu(buf_end - S);
           end_buf = buf_end - S;
         }
-        rv = fwdBlock(wide_shuf_mask_lo_highclear, wide_shuf_mask_lo_highset, chars, end_buf);
+        rv = fwdBlock<S, is_wide>(wide_shuf_mask_lo_highclear, wide_shuf_mask_lo_highset, chars, end_buf);
         DEBUG_PRINTF("rv %p \n", rv);
         if (rv && rv < buf_end) return rv;
     }
@@ -325,15 +341,25 @@ const u8 *truffleExecReal(const m128 &shuf_mask_lo_highclear, m128 shuf_mask_lo_
     return buf_end;
 }
 
-template <uint16_t S>
+template <uint16_t S, bool is_wide>
 static really_inline
-const u8 *revBlock(SuperVector<S> shuf_mask_lo_highclear, SuperVector<S> shuf_mask_lo_highset, SuperVector<S> v, 
+const u8 *revBlock(SuperVector<S> shuf_mask_lo_highclear, SuperVector<S> shuf_mask_lo_highset, SuperVector<S> v,
                     const u8 *buf) {
-    SuperVector<S> res = blockSingleMask(shuf_mask_lo_highclear, shuf_mask_lo_highset, v);
-    return last_zero_match_inverted<S>(buf, res);
+    if constexpr (is_wide) {
+#ifdef CAN_USE_WIDE_TRUFFLE
+        SuperVector<S> res = blockSingleMaskWide(shuf_mask_lo_highclear, shuf_mask_lo_highset, v);
+        return last_zero_match_inverted<S>(buf, res);
+#else
+        assert(false);
+        return nullptr;
+#endif
+    } else {
+        SuperVector<S> res = blockSingleMask(shuf_mask_lo_highclear, shuf_mask_lo_highset, v);
+        return last_zero_match_inverted<S>(buf, res);
+    }
 }
 
-template <uint16_t S>
+template <uint16_t S, bool is_wide>
 const u8 *rtruffleExecReal(m128 shuf_mask_lo_highclear, m128 shuf_mask_lo_highset, const u8 *buf, const u8 *buf_end){
     assert(buf && buf_end);
     assert(buf < buf_end);
@@ -355,7 +381,7 @@ const u8 *rtruffleExecReal(m128 shuf_mask_lo_highclear, m128 shuf_mask_lo_highse
         if (!ISALIGNED_N(d, S)) {
             SuperVector<S> chars = SuperVector<S>::loadu(d - S);
             const u8 *dbot = ROUNDDOWN_PTR(d, S);
-            rv = revBlock(wide_shuf_mask_lo_highclear, wide_shuf_mask_lo_highset, chars, d - S);
+            rv = revBlock<S, is_wide>(wide_shuf_mask_lo_highclear, wide_shuf_mask_lo_highset, chars, d - S);
             DEBUG_PRINTF("rv %p \n", rv);
             if (rv >= dbot) return rv;
             d = dbot;
@@ -368,7 +394,7 @@ const u8 *rtruffleExecReal(m128 shuf_mask_lo_highclear, m128 shuf_mask_lo_highse
 
             d -= S;
             SuperVector<S> chars = SuperVector<S>::load(d);
-            rv = revBlock(wide_shuf_mask_lo_highclear, wide_shuf_mask_lo_highset, chars, d);
+            rv = revBlock<S, is_wide>(wide_shuf_mask_lo_highclear, wide_shuf_mask_lo_highset, chars, d);
             if (rv) return rv;
         }
     }
@@ -383,7 +409,7 @@ const u8 *rtruffleExecReal(m128 shuf_mask_lo_highclear, m128 shuf_mask_lo_highse
         } else {
           chars = SuperVector<S>::loadu(buf);
         }
-        rv = revBlock(wide_shuf_mask_lo_highclear, wide_shuf_mask_lo_highset, chars, buf);
+        rv = revBlock<S, is_wide>(wide_shuf_mask_lo_highclear, wide_shuf_mask_lo_highset, chars, buf);
         DEBUG_PRINTF("rv %p \n", rv);
         if (rv && rv < buf_end) return rv;
     }
diff --git a/src/util/arch/arm/arm.h b/src/util/arch/arm/arm.h
index 3735d066e..9b743c1c0 100644
--- a/src/util/arch/arm/arm.h
+++ b/src/util/arch/arm/arm.h
@@ -57,6 +57,8 @@
 #define CAN_USE_WIDE_TRUFFLE 1
 #elif defined(HAVE_SVE)
 #define CAN_USE_WIDE_TRUFFLE (svcntb() >= 32)
+#elif defined(HAVE_NEON) && defined(ARCH_AARCH64) && !defined(VS_SIMDE_BACKEND)
+#define CAN_USE_WIDE_TRUFFLE 1
 #endif
 
 #endif // UTIL_ARCH_ARM_H_
diff --git a/src/util/supervector/arch/arm/impl.cpp b/src/util/supervector/arch/arm/impl.cpp
index b4cca4409..c19b281a4 100644
--- a/src/util/supervector/arch/arm/impl.cpp
+++ b/src/util/supervector/arch/arm/impl.cpp
@@ -363,6 +363,7 @@ template SuperVector<16> SuperVector<16>::vshl_128_imm<1>() const;
 template SuperVector<16> SuperVector<16>::vshl_128_imm<4>() const;
 template SuperVector<16> SuperVector<16>::vshr_8_imm<1>() const;
 template SuperVector<16> SuperVector<16>::vshr_8_imm<4>() const;
+template SuperVector<16> SuperVector<16>::vshr_8_imm<5>() const;
 template SuperVector<16> SuperVector<16>::vshr_16_imm<1>() const;
 template SuperVector<16> SuperVector<16>::vshr_64_imm<1>() const;
 template SuperVector<16> SuperVector<16>::vshr_64_imm<4>() const;
diff --git a/unit/internal/truffleWide.cpp b/unit/internal/truffleWide.cpp
index 733e6ffb8..c5afb1726 100644
--- a/unit/internal/truffleWide.cpp
+++ b/unit/internal/truffleWide.cpp
@@ -36,7 +36,7 @@
 #include "util/simd_utils.h"
 
 #include "util/arch.h"
-#ifdef HAVE_SVE
+#ifdef CAN_USE_WIDE_TRUFFLE
 using namespace ue2;
 
 #define SKIP_IF_NO_WIDE_AVAILABLE() \