diff --git a/src/layer/x86/lstm_int8.h b/src/layer/x86/lstm_int8.h
index 752d3b16802..0bc9cda343a 100644
--- a/src/layer/x86/lstm_int8.h
+++ b/src/layer/x86/lstm_int8.h
@@ -1676,11 +1676,14 @@ static void lstm_dynamic_quantize_scale2int8(const float* ptr, int size, float s
         __m128 _p = _mm_loadu_ps(ptr);
         _p = _mm_mul_ps(_p, _scale);
         *(int32_t*)outptr = float2int8_sse(_p);
+#ifndef _MSC_VER
+        // vs2019 crash on 128bit vnni :L   --- nihui
 #if __AVXVNNI__ || __AVX512VNNI__
         outptr[0] += 127;
         outptr[1] += 127;
         outptr[2] += 127;
         outptr[3] += 127;
+#endif
 #endif
         ptr += 4;
         outptr += 4;
@@ -1781,10 +1784,17 @@ static void lstm_int8(const Mat& bottom_blob_int8, const Mat& bottom_blob_int8_d
                 }
                 for (; i + 3 < num_output; i += 4)
                 {
+#ifdef _MSC_VER
+                    hs[0] = 0;
+                    hs[1] = 0;
+                    hs[2] = 0;
+                    hs[3] = 0;
+#else
                     hs[0] = 127;
                     hs[1] = 127;
                     hs[2] = 127;
                     hs[3] = 127;
+#endif
                     hs += 4;
                 }
                 for (; i < num_output; i++)
@@ -1899,6 +1909,9 @@ static void lstm_int8(const Mat& bottom_blob_int8, const Mat& bottom_blob_int8_d
                 __m512i _xi = _mm512_set1_epi32(((const int*)(x + i))[0]);
                 __m512i _w = _mm512_loadu_si512((const __m512i*)kptr);
 
+#ifdef _MSC_VER
+                _xi = _mm512_add_epi32(_xi, _mm512_set1_epi8(127));
+#endif
                 _lstm_IFOGx0 = _mm512_dpbusd_epi32(_lstm_IFOGx0, _xi, _w);
 
                 kptr += 64;
@@ -2083,6 +2096,9 @@ static void lstm_int8(const Mat& bottom_blob_int8, const Mat& bottom_blob_int8_d
                 __m512i _h_cont = _mm512_set1_epi32(((const int*)(hs + i))[0]);
                 __m512i _w = _mm512_loadu_si512((const __m512i*)kptr);
 
+#ifdef _MSC_VER
+                _h_cont = _mm512_add_epi32(_h_cont, _mm512_set1_epi8(127));
+#endif
                 _lstm_IFOGh0 = _mm512_dpbusd_epi32(_lstm_IFOGh0, _h_cont, _w);
 
                 kptr += 64;
@@ -2295,6 +2311,9 @@ static void lstm_int8(const Mat& bottom_blob_int8, const Mat& bottom_blob_int8_d
                 __m256i _xi = _mm256_castps_si256(_mm256_broadcast_ss((const float*)(x + i)));
                 __m256i _w = _mm256_loadu_si256((const __m256i*)kptr);
 
+#ifdef _MSC_VER
+                _xi = _mm256_add_epi32(_xi, _mm256_set1_epi8(127));
+#endif
                 _lstm_IFOGx0 = _mm256_dpbusd_epi32(_lstm_IFOGx0, _xi, _w);
 
                 kptr += 32;
@@ -2453,6 +2472,9 @@ static void lstm_int8(const Mat& bottom_blob_int8, const Mat& bottom_blob_int8_d
                 __m256i _h_cont = _mm256_castps_si256(_mm256_broadcast_ss((const float*)(hs + i)));
                 __m256i _w = _mm256_loadu_si256((const __m256i*)kptr);
 
+#ifdef _MSC_VER
+                _h_cont = _mm256_add_epi32(_h_cont, _mm256_set1_epi8(127));
+#endif
                 _lstm_IFOGh0 = _mm256_dpbusd_epi32(_lstm_IFOGh0, _h_cont, _w);
 
                 kptr += 32;
@@ -2652,6 +2674,9 @@ static void lstm_int8(const Mat& bottom_blob_int8, const Mat& bottom_blob_int8_d
                 __m128i _xi = _mm_castps_si128(_mm_load1_ps((const float*)(x + i)));
                 __m128i _w = _mm_loadu_si128((const __m128i*)kptr);
 
+#ifdef _MSC_VER
+                _xi = _mm_add_epi32(_xi, _mm_set1_epi8(127));
+#endif
                 _lstm_IFOGx0 = _mm_dpbusd_epi32(_lstm_IFOGx0, _xi, _w);
 
                 kptr += 16;
@@ -2858,6 +2883,9 @@ static void lstm_int8(const Mat& bottom_blob_int8, const Mat& bottom_blob_int8_d
                 __m128i _h_cont = _mm_castps_si128(_mm_load1_ps((const float*)(hs + i)));
                 __m128i _w = _mm_loadu_si128((const __m128i*)kptr);
 
+#ifdef _MSC_VER
+                _h_cont = _mm_add_epi32(_h_cont, _mm_set1_epi8(127));
+#endif
                 _lstm_IFOGh0 = _mm_dpbusd_epi32(_lstm_IFOGh0, _h_cont, _w);
 
                 kptr += 16;