diff --git a/OpenCL/m07100.cl b/OpenCL/m07100.cl
index 1464b0ec1..963dd1651 100644
--- a/OpenCL/m07100.cl
+++ b/OpenCL/m07100.cl
@@ -5,6 +5,8 @@
 
 #define _PBKDF2_SHA512_
 
+#define NEW_SIMD_CODE
+
 #include "include/constants.h"
 #include "include/kernel_vendor.h"
 
@@ -16,6 +18,7 @@
 #include "include/kernel_functions.c"
 #include "OpenCL/types_ocl.c"
 #include "OpenCL/common.c"
+#include "OpenCL/simd.c"
 
 #define COMPARE_S "OpenCL/check_single_comp4.c"
 #define COMPARE_M "OpenCL/check_multi_comp4.c"
@@ -44,7 +47,7 @@ __constant u64 k_sha512[80] =
   SHA512C4c, SHA512C4d, SHA512C4e, SHA512C4f,
 };
 
-void sha512_transform (const u64 w[16], u64 dgst[8])
+void sha512_transform_S (const u64 w[16], u64 dgst[8])
 {
   u64 a = dgst[0];
   u64 b = dgst[1];
@@ -72,6 +75,197 @@ void sha512_transform (const u64 w[16], u64 dgst[8])
   u64 we_t = w[14];
   u64 wf_t = w[15];
 
+  #define ROUND_EXPAND()                              \
+  {                                                   \
+    w0_t = SHA512_EXPAND_S (we_t, w9_t, w1_t, w0_t);  \
+    w1_t = SHA512_EXPAND_S (wf_t, wa_t, w2_t, w1_t);  \
+    w2_t = SHA512_EXPAND_S (w0_t, wb_t, w3_t, w2_t);  \
+    w3_t = SHA512_EXPAND_S (w1_t, wc_t, w4_t, w3_t);  \
+    w4_t = SHA512_EXPAND_S (w2_t, wd_t, w5_t, w4_t);  \
+    w5_t = SHA512_EXPAND_S (w3_t, we_t, w6_t, w5_t);  \
+    w6_t = SHA512_EXPAND_S (w4_t, wf_t, w7_t, w6_t);  \
+    w7_t = SHA512_EXPAND_S (w5_t, w0_t, w8_t, w7_t);  \
+    w8_t = SHA512_EXPAND_S (w6_t, w1_t, w9_t, w8_t);  \
+    w9_t = SHA512_EXPAND_S (w7_t, w2_t, wa_t, w9_t);  \
+    wa_t = SHA512_EXPAND_S (w8_t, w3_t, wb_t, wa_t);  \
+    wb_t = SHA512_EXPAND_S (w9_t, w4_t, wc_t, wb_t);  \
+    wc_t = SHA512_EXPAND_S (wa_t, w5_t, wd_t, wc_t);  \
+    wd_t = SHA512_EXPAND_S (wb_t, w6_t, we_t, wd_t);  \
+    we_t = SHA512_EXPAND_S (wc_t, w7_t, wf_t, we_t);  \
+    wf_t = SHA512_EXPAND_S (wd_t, w8_t, w0_t, wf_t);  \
+  }
+
+  #define ROUND_STEP(i)                                                                     \
+  {                                                                                         \
+    SHA512_STEP_S (SHA512_F0o, SHA512_F1o, a, b, c, d, e, f, g, h, w0_t, k_sha512[i +  0]); \
+    SHA512_STEP_S (SHA512_F0o, SHA512_F1o, h, a, b, c, d, e, f, g, w1_t, k_sha512[i +  1]); \
+    SHA512_STEP_S (SHA512_F0o, SHA512_F1o, g, h, a, b, c, d, e, f, w2_t, k_sha512[i +  2]); \
+    SHA512_STEP_S (SHA512_F0o, SHA512_F1o, f, g, h, a, b, c, d, e, w3_t, k_sha512[i +  3]); \
+    SHA512_STEP_S (SHA512_F0o, SHA512_F1o, e, f, g, h, a, b, c, d, w4_t, k_sha512[i +  4]); \
+    SHA512_STEP_S (SHA512_F0o, SHA512_F1o, d, e, f, g, h, a, b, c, w5_t, k_sha512[i +  5]); \
+    SHA512_STEP_S (SHA512_F0o, SHA512_F1o, c, d, e, f, g, h, a, b, w6_t, k_sha512[i +  6]); \
+    SHA512_STEP_S (SHA512_F0o, SHA512_F1o, b, c, d, e, f, g, h, a, w7_t, k_sha512[i +  7]); \
+    SHA512_STEP_S (SHA512_F0o, SHA512_F1o, a, b, c, d, e, f, g, h, w8_t, k_sha512[i +  8]); \
+    SHA512_STEP_S (SHA512_F0o, SHA512_F1o, h, a, b, c, d, e, f, g, w9_t, k_sha512[i +  9]); \
+    SHA512_STEP_S (SHA512_F0o, SHA512_F1o, g, h, a, b, c, d, e, f, wa_t, k_sha512[i + 10]); \
+    SHA512_STEP_S (SHA512_F0o, SHA512_F1o, f, g, h, a, b, c, d, e, wb_t, k_sha512[i + 11]); \
+    SHA512_STEP_S (SHA512_F0o, SHA512_F1o, e, f, g, h, a, b, c, d, wc_t, k_sha512[i + 12]); \
+    SHA512_STEP_S (SHA512_F0o, SHA512_F1o, d, e, f, g, h, a, b, c, wd_t, k_sha512[i + 13]); \
+    SHA512_STEP_S (SHA512_F0o, SHA512_F1o, c, d, e, f, g, h, a, b, we_t, k_sha512[i + 14]); \
+    SHA512_STEP_S (SHA512_F0o, SHA512_F1o, b, c, d, e, f, g, h, a, wf_t, k_sha512[i + 15]); \
+  }
+
+  ROUND_STEP (0);
+
+  #ifdef _unroll
+  #pragma unroll
+  #endif
+  for (int i = 16; i < 80; i += 16)
+  {
+    ROUND_EXPAND (); ROUND_STEP (i);
+  }
+
+  dgst[0] += a;
+  dgst[1] += b;
+  dgst[2] += c;
+  dgst[3] += d;
+  dgst[4] += e;
+  dgst[5] += f;
+  dgst[6] += g;
+  dgst[7] += h;
+}
+
+void hmac_sha512_run_S (const u64 w1[16], const u64 ipad[8], const u64 opad[8], u64 dgst[8])
+{
+  dgst[0] = ipad[0];
+  dgst[1] = ipad[1];
+  dgst[2] = ipad[2];
+  dgst[3] = ipad[3];
+  dgst[4] = ipad[4];
+  dgst[5] = ipad[5];
+  dgst[6] = ipad[6];
+  dgst[7] = ipad[7];
+
+  sha512_transform_S (w1, dgst);
+
+  u64 w[16];
+
+  w[ 0] = dgst[0];
+  w[ 1] = dgst[1];
+  w[ 2] = dgst[2];
+  w[ 3] = dgst[3];
+  w[ 4] = dgst[4];
+  w[ 5] = dgst[5];
+  w[ 6] = dgst[6];
+  w[ 7] = dgst[7];
+  w[ 8] = 0x8000000000000000;
+  w[ 9] = 0;
+  w[10] = 0;
+  w[11] = 0;
+  w[12] = 0;
+  w[13] = 0;
+  w[14] = 0;
+  w[15] = (128 + 64) * 8;
+
+  dgst[0] = opad[0];
+  dgst[1] = opad[1];
+  dgst[2] = opad[2];
+  dgst[3] = opad[3];
+  dgst[4] = opad[4];
+  dgst[5] = opad[5];
+  dgst[6] = opad[6];
+  dgst[7] = opad[7];
+
+  sha512_transform_S (w, dgst);
+}
+
+void hmac_sha512_init_S (u64 w[16], u64 ipad[8], u64 opad[8])
+{
+  w[ 0] ^= 0x3636363636363636;
+  w[ 1] ^= 0x3636363636363636;
+  w[ 2] ^= 0x3636363636363636;
+  w[ 3] ^= 0x3636363636363636;
+  w[ 4] ^= 0x3636363636363636;
+  w[ 5] ^= 0x3636363636363636;
+  w[ 6] ^= 0x3636363636363636;
+  w[ 7] ^= 0x3636363636363636;
+  w[ 8] ^= 0x3636363636363636;
+  w[ 9] ^= 0x3636363636363636;
+  w[10] ^= 0x3636363636363636;
+  w[11] ^= 0x3636363636363636;
+  w[12] ^= 0x3636363636363636;
+  w[13] ^= 0x3636363636363636;
+  w[14] ^= 0x3636363636363636;
+  w[15] ^= 0x3636363636363636;
+
+  ipad[0] = SHA512M_A;
+  ipad[1] = SHA512M_B;
+  ipad[2] = SHA512M_C;
+  ipad[3] = SHA512M_D;
+  ipad[4] = SHA512M_E;
+  ipad[5] = SHA512M_F;
+  ipad[6] = SHA512M_G;
+  ipad[7] = SHA512M_H;
+
+  sha512_transform_S (w, ipad);
+
+  w[ 0] ^= 0x6a6a6a6a6a6a6a6a;
+  w[ 1] ^= 0x6a6a6a6a6a6a6a6a;
+  w[ 2] ^= 0x6a6a6a6a6a6a6a6a;
+  w[ 3] ^= 0x6a6a6a6a6a6a6a6a;
+  w[ 4] ^= 0x6a6a6a6a6a6a6a6a;
+  w[ 5] ^= 0x6a6a6a6a6a6a6a6a;
+  w[ 6] ^= 0x6a6a6a6a6a6a6a6a;
+  w[ 7] ^= 0x6a6a6a6a6a6a6a6a;
+  w[ 8] ^= 0x6a6a6a6a6a6a6a6a;
+  w[ 9] ^= 0x6a6a6a6a6a6a6a6a;
+  w[10] ^= 0x6a6a6a6a6a6a6a6a;
+  w[11] ^= 0x6a6a6a6a6a6a6a6a;
+  w[12] ^= 0x6a6a6a6a6a6a6a6a;
+  w[13] ^= 0x6a6a6a6a6a6a6a6a;
+  w[14] ^= 0x6a6a6a6a6a6a6a6a;
+  w[15] ^= 0x6a6a6a6a6a6a6a6a;
+
+  opad[0] = SHA512M_A;
+  opad[1] = SHA512M_B;
+  opad[2] = SHA512M_C;
+  opad[3] = SHA512M_D;
+  opad[4] = SHA512M_E;
+  opad[5] = SHA512M_F;
+  opad[6] = SHA512M_G;
+  opad[7] = SHA512M_H;
+
+  sha512_transform_S (w, opad);
+}
+
+void sha512_transform_V (const u64x w[16], u64x dgst[8])
+{
+  u64x a = dgst[0];
+  u64x b = dgst[1];
+  u64x c = dgst[2];
+  u64x d = dgst[3];
+  u64x e = dgst[4];
+  u64x f = dgst[5];
+  u64x g = dgst[6];
+  u64x h = dgst[7];
+
+  u64x w0_t = w[ 0];
+  u64x w1_t = w[ 1];
+  u64x w2_t = w[ 2];
+  u64x w3_t = w[ 3];
+  u64x w4_t = w[ 4];
+  u64x w5_t = w[ 5];
+  u64x w6_t = w[ 6];
+  u64x w7_t = w[ 7];
+  u64x w8_t = w[ 8];
+  u64x w9_t = w[ 9];
+  u64x wa_t = w[10];
+  u64x wb_t = w[11];
+  u64x wc_t = w[12];
+  u64x wd_t = w[13];
+  u64x we_t = w[14];
+  u64x wf_t = w[15];
+
   #define ROUND_EXPAND()                            \
   {                                                 \
     w0_t = SHA512_EXPAND (we_t, w9_t, w1_t, w0_t);  \
@@ -132,7 +326,7 @@ void sha512_transform (const u64 w[16], u64 dgst[8])
   dgst[7] += h;
 }
 
-void hmac_run (const u64 w1[16], const u64 ipad[8], const u64 opad[8], u64 dgst[8])
+void hmac_sha512_run_V (const u64x w1[16], const u64x ipad[8], const u64x opad[8], u64x dgst[8])
 {
   dgst[0] = ipad[0];
   dgst[1] = ipad[1];
@@ -143,9 +337,9 @@ void hmac_run (const u64 w1[16], const u64 ipad[8], const u64 opad[8], u64 dgst[
   dgst[6] = ipad[6];
   dgst[7] = ipad[7];
 
-  sha512_transform (w1, dgst);
+  sha512_transform_V (w1, dgst);
 
-  u64 w[16];
+  u64x w[16];
 
   w[ 0] = dgst[0];
   w[ 1] = dgst[1];
@@ -173,10 +367,10 @@ void hmac_run (const u64 w1[16], const u64 ipad[8], const u64 opad[8], u64 dgst[
   dgst[6] = opad[6];
   dgst[7] = opad[7];
 
-  sha512_transform (w, dgst);
+  sha512_transform_V (w, dgst);
 }
 
-void hmac_init (u64 w[16], u64 ipad[8], u64 opad[8])
+void hmac_sha512_init_V (u64x w[16], u64x ipad[8], u64x opad[8])
 {
   w[ 0] ^= 0x3636363636363636;
   w[ 1] ^= 0x3636363636363636;
@@ -204,7 +398,7 @@ void hmac_init (u64 w[16], u64 ipad[8], u64 opad[8])
   ipad[6] = SHA512M_G;
   ipad[7] = SHA512M_H;
 
-  sha512_transform (w, ipad);
+  sha512_transform_V (w, ipad);
 
   w[ 0] ^= 0x6a6a6a6a6a6a6a6a;
   w[ 1] ^= 0x6a6a6a6a6a6a6a6a;
@@ -232,7 +426,7 @@ void hmac_init (u64 w[16], u64 ipad[8], u64 opad[8])
   opad[6] = SHA512M_G;
   opad[7] = SHA512M_H;
 
-  sha512_transform (w, opad);
+  sha512_transform_V (w, opad);
 }
 
 __kernel void m07100_init (__global pw_t *pws, __global kernel_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global pbkdf2_sha512_tmp_t *tmps, __global void *hooks, __global u32 *bitmaps_buf_s1_a, __global u32 *bitmaps_buf_s1_b, __global u32 *bitmaps_buf_s1_c, __global u32 *bitmaps_buf_s1_d, __global u32 *bitmaps_buf_s2_a, __global u32 *bitmaps_buf_s2_b, __global u32 *bitmaps_buf_s2_c, __global u32 *bitmaps_buf_s2_d, __global plain_t *plains_buf, __global digest_t *digests_buf, __global u32 *hashes_shown, __global salt_t *salt_bufs, __global pbkdf2_sha512_t *esalt_bufs, __global u32 *d_return_buf, __global u32 *d_scryptV_buf, const u32 bitmap_mask, const u32 bitmap_shift1, const u32 bitmap_shift2, const u32 salt_pos, const u32 loop_pos, const u32 loop_cnt, const u32 il_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
@@ -247,31 +441,31 @@ __kernel void m07100_init (__global pw_t *pws, __global kernel_rule_t *rules_buf
 
   u32 w0[4];
 
-  w0[0] = swap32 (pws[gid].i[ 0]);
-  w0[1] = swap32 (pws[gid].i[ 1]);
-  w0[2] = swap32 (pws[gid].i[ 2]);
-  w0[3] = swap32 (pws[gid].i[ 3]);
+  w0[0] = swap32_S (pws[gid].i[ 0]);
+  w0[1] = swap32_S (pws[gid].i[ 1]);
+  w0[2] = swap32_S (pws[gid].i[ 2]);
+  w0[3] = swap32_S (pws[gid].i[ 3]);
 
   u32 w1[4];
 
-  w1[0] = swap32 (pws[gid].i[ 4]);
-  w1[1] = swap32 (pws[gid].i[ 5]);
-  w1[2] = swap32 (pws[gid].i[ 6]);
-  w1[3] = swap32 (pws[gid].i[ 7]);
+  w1[0] = swap32_S (pws[gid].i[ 4]);
+  w1[1] = swap32_S (pws[gid].i[ 5]);
+  w1[2] = swap32_S (pws[gid].i[ 6]);
+  w1[3] = swap32_S (pws[gid].i[ 7]);
 
   u32 w2[4];
 
-  w2[0] = swap32 (pws[gid].i[ 8]);
-  w2[1] = swap32 (pws[gid].i[ 9]);
-  w2[2] = swap32 (pws[gid].i[10]);
-  w2[3] = swap32 (pws[gid].i[11]);
+  w2[0] = swap32_S (pws[gid].i[ 8]);
+  w2[1] = swap32_S (pws[gid].i[ 9]);
+  w2[2] = swap32_S (pws[gid].i[10]);
+  w2[3] = swap32_S (pws[gid].i[11]);
 
   u32 w3[4];
 
-  w3[0] = swap32 (pws[gid].i[12]);
-  w3[1] = swap32 (pws[gid].i[13]);
-  w3[2] = swap32 (pws[gid].i[14]);
-  w3[3] = swap32 (pws[gid].i[15]);
+  w3[0] = swap32_S (pws[gid].i[12]);
+  w3[1] = swap32_S (pws[gid].i[13]);
+  w3[2] = swap32_S (pws[gid].i[14]);
+  w3[3] = swap32_S (pws[gid].i[15]);
 
   /**
    * salt
@@ -281,33 +475,33 @@ __kernel void m07100_init (__global pw_t *pws, __global kernel_rule_t *rules_buf
 
   u32 salt_len  = salt_bufs[salt_pos].salt_len;
 
-  esalt_buf[ 0] = hl32_to_64 (swap32 (esalt_bufs[salt_pos].salt_buf[ 0]), swap32 (esalt_bufs[salt_pos].salt_buf[ 1]));
-  esalt_buf[ 1] = hl32_to_64 (swap32 (esalt_bufs[salt_pos].salt_buf[ 2]), swap32 (esalt_bufs[salt_pos].salt_buf[ 3]));
-  esalt_buf[ 2] = hl32_to_64 (swap32 (esalt_bufs[salt_pos].salt_buf[ 4]), swap32 (esalt_bufs[salt_pos].salt_buf[ 5]));
-  esalt_buf[ 3] = hl32_to_64 (swap32 (esalt_bufs[salt_pos].salt_buf[ 6]), swap32 (esalt_bufs[salt_pos].salt_buf[ 7]));
-  esalt_buf[ 4] = hl32_to_64 (swap32 (esalt_bufs[salt_pos].salt_buf[ 8]), swap32 (esalt_bufs[salt_pos].salt_buf[ 9]));
-  esalt_buf[ 5] = hl32_to_64 (swap32 (esalt_bufs[salt_pos].salt_buf[10]), swap32 (esalt_bufs[salt_pos].salt_buf[11]));
-  esalt_buf[ 6] = hl32_to_64 (swap32 (esalt_bufs[salt_pos].salt_buf[12]), swap32 (esalt_bufs[salt_pos].salt_buf[13]));
-  esalt_buf[ 7] = hl32_to_64 (swap32 (esalt_bufs[salt_pos].salt_buf[14]), swap32 (esalt_bufs[salt_pos].salt_buf[15]));
-  esalt_buf[ 8] = hl32_to_64 (swap32 (esalt_bufs[salt_pos].salt_buf[16]), swap32 (esalt_bufs[salt_pos].salt_buf[17]));
-  esalt_buf[ 9] = hl32_to_64 (swap32 (esalt_bufs[salt_pos].salt_buf[18]), swap32 (esalt_bufs[salt_pos].salt_buf[19]));
-  esalt_buf[10] = hl32_to_64 (swap32 (esalt_bufs[salt_pos].salt_buf[20]), swap32 (esalt_bufs[salt_pos].salt_buf[21]));
-  esalt_buf[11] = hl32_to_64 (swap32 (esalt_bufs[salt_pos].salt_buf[22]), swap32 (esalt_bufs[salt_pos].salt_buf[23]));
-  esalt_buf[12] = hl32_to_64 (swap32 (esalt_bufs[salt_pos].salt_buf[24]), swap32 (esalt_bufs[salt_pos].salt_buf[25]));
-  esalt_buf[13] = hl32_to_64 (swap32 (esalt_bufs[salt_pos].salt_buf[26]), swap32 (esalt_bufs[salt_pos].salt_buf[27]));
+  esalt_buf[ 0] = hl32_to_64_S (swap32_S (esalt_bufs[salt_pos].salt_buf[ 0]), swap32_S (esalt_bufs[salt_pos].salt_buf[ 1]));
+  esalt_buf[ 1] = hl32_to_64_S (swap32_S (esalt_bufs[salt_pos].salt_buf[ 2]), swap32_S (esalt_bufs[salt_pos].salt_buf[ 3]));
+  esalt_buf[ 2] = hl32_to_64_S (swap32_S (esalt_bufs[salt_pos].salt_buf[ 4]), swap32_S (esalt_bufs[salt_pos].salt_buf[ 5]));
+  esalt_buf[ 3] = hl32_to_64_S (swap32_S (esalt_bufs[salt_pos].salt_buf[ 6]), swap32_S (esalt_bufs[salt_pos].salt_buf[ 7]));
+  esalt_buf[ 4] = hl32_to_64_S (swap32_S (esalt_bufs[salt_pos].salt_buf[ 8]), swap32_S (esalt_bufs[salt_pos].salt_buf[ 9]));
+  esalt_buf[ 5] = hl32_to_64_S (swap32_S (esalt_bufs[salt_pos].salt_buf[10]), swap32_S (esalt_bufs[salt_pos].salt_buf[11]));
+  esalt_buf[ 6] = hl32_to_64_S (swap32_S (esalt_bufs[salt_pos].salt_buf[12]), swap32_S (esalt_bufs[salt_pos].salt_buf[13]));
+  esalt_buf[ 7] = hl32_to_64_S (swap32_S (esalt_bufs[salt_pos].salt_buf[14]), swap32_S (esalt_bufs[salt_pos].salt_buf[15]));
+  esalt_buf[ 8] = hl32_to_64_S (swap32_S (esalt_bufs[salt_pos].salt_buf[16]), swap32_S (esalt_bufs[salt_pos].salt_buf[17]));
+  esalt_buf[ 9] = hl32_to_64_S (swap32_S (esalt_bufs[salt_pos].salt_buf[18]), swap32_S (esalt_bufs[salt_pos].salt_buf[19]));
+  esalt_buf[10] = hl32_to_64_S (swap32_S (esalt_bufs[salt_pos].salt_buf[20]), swap32_S (esalt_bufs[salt_pos].salt_buf[21]));
+  esalt_buf[11] = hl32_to_64_S (swap32_S (esalt_bufs[salt_pos].salt_buf[22]), swap32_S (esalt_bufs[salt_pos].salt_buf[23]));
+  esalt_buf[12] = hl32_to_64_S (swap32_S (esalt_bufs[salt_pos].salt_buf[24]), swap32_S (esalt_bufs[salt_pos].salt_buf[25]));
+  esalt_buf[13] = hl32_to_64_S (swap32_S (esalt_bufs[salt_pos].salt_buf[26]), swap32_S (esalt_bufs[salt_pos].salt_buf[27]));
   esalt_buf[14] = 0;
   esalt_buf[15] = (128 + salt_len + 4) * 8;
 
   u64 w[16];
 
-  w[ 0] = hl32_to_64 (w0[0], w0[1]);
-  w[ 1] = hl32_to_64 (w0[2], w0[3]);
-  w[ 2] = hl32_to_64 (w1[0], w1[1]);
-  w[ 3] = hl32_to_64 (w1[2], w1[3]);
-  w[ 4] = hl32_to_64 (w2[0], w2[1]);
-  w[ 5] = hl32_to_64 (w2[2], w2[3]);
-  w[ 6] = hl32_to_64 (w3[0], w3[1]);
-  w[ 7] = hl32_to_64 (w3[2], w3[3]);
+  w[ 0] = hl32_to_64_S (w0[0], w0[1]);
+  w[ 1] = hl32_to_64_S (w0[2], w0[3]);
+  w[ 2] = hl32_to_64_S (w1[0], w1[1]);
+  w[ 3] = hl32_to_64_S (w1[2], w1[3]);
+  w[ 4] = hl32_to_64_S (w2[0], w2[1]);
+  w[ 5] = hl32_to_64_S (w2[2], w2[3]);
+  w[ 6] = hl32_to_64_S (w3[0], w3[1]);
+  w[ 7] = hl32_to_64_S (w3[2], w3[3]);
   w[ 8] = 0;
   w[ 9] = 0;
   w[10] = 0;
@@ -320,7 +514,7 @@ __kernel void m07100_init (__global pw_t *pws, __global kernel_rule_t *rules_buf
   u64 ipad[8];
   u64 opad[8];
 
-  hmac_init (w, ipad, opad);
+  hmac_sha512_init_S (w, ipad, opad);
 
   tmps[gid].ipad[0] = ipad[0];
   tmps[gid].ipad[1] = ipad[1];
@@ -344,7 +538,7 @@ __kernel void m07100_init (__global pw_t *pws, __global kernel_rule_t *rules_buf
   {
     u64 dgst[8];
 
-    hmac_run (esalt_buf, ipad, opad, dgst);
+    hmac_sha512_run_S (esalt_buf, ipad, opad, dgst);
 
     tmps[gid].dgst[i + 0] = dgst[0];
     tmps[gid].dgst[i + 1] = dgst[1];
@@ -372,55 +566,53 @@ __kernel void m07100_loop (__global pw_t *pws, __global kernel_rule_t *rules_buf
 
   if (gid >= gid_max) return;
 
-  u64 ipad[8];
-
-  ipad[0] = tmps[gid].ipad[0];
-  ipad[1] = tmps[gid].ipad[1];
-  ipad[2] = tmps[gid].ipad[2];
-  ipad[3] = tmps[gid].ipad[3];
-  ipad[4] = tmps[gid].ipad[4];
-  ipad[5] = tmps[gid].ipad[5];
-  ipad[6] = tmps[gid].ipad[6];
-  ipad[7] = tmps[gid].ipad[7];
-
-  u64 opad[8];
-
-  opad[0] = tmps[gid].opad[0];
-  opad[1] = tmps[gid].opad[1];
-  opad[2] = tmps[gid].opad[2];
-  opad[3] = tmps[gid].opad[3];
-  opad[4] = tmps[gid].opad[4];
-  opad[5] = tmps[gid].opad[5];
-  opad[6] = tmps[gid].opad[6];
-  opad[7] = tmps[gid].opad[7];
+  u64x ipad[8];
+  u64x opad[8];
+
+  ipad[0] = pack64v (tmps, ipad, gid, 0);
+  ipad[1] = pack64v (tmps, ipad, gid, 1);
+  ipad[2] = pack64v (tmps, ipad, gid, 2);
+  ipad[3] = pack64v (tmps, ipad, gid, 3);
+  ipad[4] = pack64v (tmps, ipad, gid, 4);
+  ipad[5] = pack64v (tmps, ipad, gid, 5);
+  ipad[6] = pack64v (tmps, ipad, gid, 6);
+  ipad[7] = pack64v (tmps, ipad, gid, 7);
+
+  opad[0] = pack64v (tmps, opad, gid, 0);
+  opad[1] = pack64v (tmps, opad, gid, 1);
+  opad[2] = pack64v (tmps, opad, gid, 2);
+  opad[3] = pack64v (tmps, opad, gid, 3);
+  opad[4] = pack64v (tmps, opad, gid, 4);
+  opad[5] = pack64v (tmps, opad, gid, 5);
+  opad[6] = pack64v (tmps, opad, gid, 6);
+  opad[7] = pack64v (tmps, opad, gid, 7);
 
   for (u32 i = 0; i < 8; i += 8)
   {
-    u64 dgst[8];
-
-    dgst[0] = tmps[gid].dgst[i + 0];
-    dgst[1] = tmps[gid].dgst[i + 1];
-    dgst[2] = tmps[gid].dgst[i + 2];
-    dgst[3] = tmps[gid].dgst[i + 3];
-    dgst[4] = tmps[gid].dgst[i + 4];
-    dgst[5] = tmps[gid].dgst[i + 5];
-    dgst[6] = tmps[gid].dgst[i + 6];
-    dgst[7] = tmps[gid].dgst[i + 7];
-
-    u64 out[8];
-
-    out[0] = tmps[gid].out[i + 0];
-    out[1] = tmps[gid].out[i + 1];
-    out[2] = tmps[gid].out[i + 2];
-    out[3] = tmps[gid].out[i + 3];
-    out[4] = tmps[gid].out[i + 4];
-    out[5] = tmps[gid].out[i + 5];
-    out[6] = tmps[gid].out[i + 6];
-    out[7] = tmps[gid].out[i + 7];
+    u64x dgst[8];
+    u64x out[8];
+
+    dgst[0] = pack64v (tmps, dgst, gid, 0);
+    dgst[1] = pack64v (tmps, dgst, gid, 1);
+    dgst[2] = pack64v (tmps, dgst, gid, 2);
+    dgst[3] = pack64v (tmps, dgst, gid, 3);
+    dgst[4] = pack64v (tmps, dgst, gid, 4);
+    dgst[5] = pack64v (tmps, dgst, gid, 5);
+    dgst[6] = pack64v (tmps, dgst, gid, 6);
+    dgst[7] = pack64v (tmps, dgst, gid, 7);
+
+    out[0] = pack64v (tmps, out, gid, 0);
+    out[1] = pack64v (tmps, out, gid, 1);
+    out[2] = pack64v (tmps, out, gid, 2);
+    out[3] = pack64v (tmps, out, gid, 3);
+    out[4] = pack64v (tmps, out, gid, 4);
+    out[5] = pack64v (tmps, out, gid, 5);
+    out[6] = pack64v (tmps, out, gid, 6);
+    out[7] = pack64v (tmps, out, gid, 7);
 
     for (u32 j = 0; j < loop_cnt; j++)
     {
-      u64 w[16];
+      u64x w[16];
 
       w[ 0] = dgst[0];
       w[ 1] = dgst[1];
@@ -439,7 +631,7 @@ __kernel void m07100_loop (__global pw_t *pws, __global kernel_rule_t *rules_buf
       w[14] = 0;
       w[15] = (128 + 64) * 8;
 
-      hmac_run (w, ipad, opad, dgst);
+      hmac_sha512_run_V (w, ipad, opad, dgst);
 
       out[0] ^= dgst[0];
       out[1] ^= dgst[1];
@@ -451,23 +643,23 @@ __kernel void m07100_loop (__global pw_t *pws, __global kernel_rule_t *rules_buf
       out[7] ^= dgst[7];
     }
 
-    tmps[gid].dgst[i + 0] = dgst[0];
-    tmps[gid].dgst[i + 1] = dgst[1];
-    tmps[gid].dgst[i + 2] = dgst[2];
-    tmps[gid].dgst[i + 3] = dgst[3];
-    tmps[gid].dgst[i + 4] = dgst[4];
-    tmps[gid].dgst[i + 5] = dgst[5];
-    tmps[gid].dgst[i + 6] = dgst[6];
-    tmps[gid].dgst[i + 7] = dgst[7];
-
-    tmps[gid].out[i + 0] = out[0];
-    tmps[gid].out[i + 1] = out[1];
-    tmps[gid].out[i + 2] = out[2];
-    tmps[gid].out[i + 3] = out[3];
-    tmps[gid].out[i + 4] = out[4];
-    tmps[gid].out[i + 5] = out[5];
-    tmps[gid].out[i + 6] = out[6];
-    tmps[gid].out[i + 7] = out[7];
+    unpackv (tmps, dgst, gid, 0, dgst[0]);
+    unpackv (tmps, dgst, gid, 1, dgst[1]);
+    unpackv (tmps, dgst, gid, 2, dgst[2]);
+    unpackv (tmps, dgst, gid, 3, dgst[3]);
+    unpackv (tmps, dgst, gid, 4, dgst[4]);
+    unpackv (tmps, dgst, gid, 5, dgst[5]);
+    unpackv (tmps, dgst, gid, 6, dgst[6]);
+    unpackv (tmps, dgst, gid, 7, dgst[7]);
+
+    unpackv (tmps, out, gid, 0, out[0]);
+    unpackv (tmps, out, gid, 1, out[1]);
+    unpackv (tmps, out, gid, 2, out[2]);
+    unpackv (tmps, out, gid, 3, out[3]);
+    unpackv (tmps, out, gid, 4, out[4]);
+    unpackv (tmps, out, gid, 5, out[5]);
+    unpackv (tmps, out, gid, 6, out[6]);
+    unpackv (tmps, out, gid, 7, out[7]);
   }
 }
 
@@ -486,10 +678,10 @@ __kernel void m07100_comp (__global pw_t *pws, __global kernel_rule_t *rules_buf
   const u64 a = tmps[gid].out[0];
   const u64 b = tmps[gid].out[1];
 
-  const u32 r0 = l32_from_64 (a);
-  const u32 r1 = h32_from_64 (a);
-  const u32 r2 = l32_from_64 (b);
-  const u32 r3 = h32_from_64 (b);
+  const u32 r0 = l32_from_64_S (a);
+  const u32 r1 = h32_from_64_S (a);
+  const u32 r2 = l32_from_64_S (b);
+  const u32 r3 = h32_from_64_S (b);
 
   #define il_pos 0
 
diff --git a/OpenCL/m10900.cl b/OpenCL/m10900.cl
index 0107d84f0..c2f24576d 100644
--- a/OpenCL/m10900.cl
+++ b/OpenCL/m10900.cl
@@ -5,6 +5,8 @@
 
 #define _PBKDF2_SHA256_
 
+#define NEW_SIMD_CODE
+
 #include "include/constants.h"
 #include "include/kernel_vendor.h"
 
@@ -14,9 +16,9 @@
 #define DGST_R3 3
 
 #include "include/kernel_functions.c"
-
 #include "OpenCL/types_ocl.c"
 #include "OpenCL/common.c"
+#include "OpenCL/simd.c"
 
 #define COMPARE_S "OpenCL/check_single_comp4.c"
 #define COMPARE_M "OpenCL/check_multi_comp4.c"
@@ -41,7 +43,7 @@ __constant u32 k_sha256[64] =
   SHA256C3c, SHA256C3d, SHA256C3e, SHA256C3f,
 };
 
-void sha256_transform (const u32 w0[4], const u32 w1[4], const u32 w2[4], const u32 w3[4], u32 digest[8])
+void sha256_transform_S (const u32 w0[4], const u32 w1[4], const u32 w2[4], const u32 w3[4], u32 digest[8])
 {
   u32 a = digest[0];
   u32 b = digest[1];
@@ -69,6 +71,195 @@ void sha256_transform (const u32 w0[4], const u32 w1[4], const u32 w2[4], const
   u32 we_t = w3[2];
   u32 wf_t = w3[3];
 
+  #define ROUND_EXPAND()                              \
+  {                                                   \
+    w0_t = SHA256_EXPAND_S (we_t, w9_t, w1_t, w0_t);  \
+    w1_t = SHA256_EXPAND_S (wf_t, wa_t, w2_t, w1_t);  \
+    w2_t = SHA256_EXPAND_S (w0_t, wb_t, w3_t, w2_t);  \
+    w3_t = SHA256_EXPAND_S (w1_t, wc_t, w4_t, w3_t);  \
+    w4_t = SHA256_EXPAND_S (w2_t, wd_t, w5_t, w4_t);  \
+    w5_t = SHA256_EXPAND_S (w3_t, we_t, w6_t, w5_t);  \
+    w6_t = SHA256_EXPAND_S (w4_t, wf_t, w7_t, w6_t);  \
+    w7_t = SHA256_EXPAND_S (w5_t, w0_t, w8_t, w7_t);  \
+    w8_t = SHA256_EXPAND_S (w6_t, w1_t, w9_t, w8_t);  \
+    w9_t = SHA256_EXPAND_S (w7_t, w2_t, wa_t, w9_t);  \
+    wa_t = SHA256_EXPAND_S (w8_t, w3_t, wb_t, wa_t);  \
+    wb_t = SHA256_EXPAND_S (w9_t, w4_t, wc_t, wb_t);  \
+    wc_t = SHA256_EXPAND_S (wa_t, w5_t, wd_t, wc_t);  \
+    wd_t = SHA256_EXPAND_S (wb_t, w6_t, we_t, wd_t);  \
+    we_t = SHA256_EXPAND_S (wc_t, w7_t, wf_t, we_t);  \
+    wf_t = SHA256_EXPAND_S (wd_t, w8_t, w0_t, wf_t);  \
+  }
+
+  #define ROUND_STEP(i)                                                                   \
+  {                                                                                       \
+    SHA256_STEP_S (SHA256_F0o, SHA256_F1o, a, b, c, d, e, f, g, h, w0_t, k_sha256[i +  0]); \
+    SHA256_STEP_S (SHA256_F0o, SHA256_F1o, h, a, b, c, d, e, f, g, w1_t, k_sha256[i +  1]); \
+    SHA256_STEP_S (SHA256_F0o, SHA256_F1o, g, h, a, b, c, d, e, f, w2_t, k_sha256[i +  2]); \
+    SHA256_STEP_S (SHA256_F0o, SHA256_F1o, f, g, h, a, b, c, d, e, w3_t, k_sha256[i +  3]); \
+    SHA256_STEP_S (SHA256_F0o, SHA256_F1o, e, f, g, h, a, b, c, d, w4_t, k_sha256[i +  4]); \
+    SHA256_STEP_S (SHA256_F0o, SHA256_F1o, d, e, f, g, h, a, b, c, w5_t, k_sha256[i +  5]); \
+    SHA256_STEP_S (SHA256_F0o, SHA256_F1o, c, d, e, f, g, h, a, b, w6_t, k_sha256[i +  6]); \
+    SHA256_STEP_S (SHA256_F0o, SHA256_F1o, b, c, d, e, f, g, h, a, w7_t, k_sha256[i +  7]); \
+    SHA256_STEP_S (SHA256_F0o, SHA256_F1o, a, b, c, d, e, f, g, h, w8_t, k_sha256[i +  8]); \
+    SHA256_STEP_S (SHA256_F0o, SHA256_F1o, h, a, b, c, d, e, f, g, w9_t, k_sha256[i +  9]); \
+    SHA256_STEP_S (SHA256_F0o, SHA256_F1o, g, h, a, b, c, d, e, f, wa_t, k_sha256[i + 10]); \
+    SHA256_STEP_S (SHA256_F0o, SHA256_F1o, f, g, h, a, b, c, d, e, wb_t, k_sha256[i + 11]); \
+    SHA256_STEP_S (SHA256_F0o, SHA256_F1o, e, f, g, h, a, b, c, d, wc_t, k_sha256[i + 12]); \
+    SHA256_STEP_S (SHA256_F0o, SHA256_F1o, d, e, f, g, h, a, b, c, wd_t, k_sha256[i + 13]); \
+    SHA256_STEP_S (SHA256_F0o, SHA256_F1o, c, d, e, f, g, h, a, b, we_t, k_sha256[i + 14]); \
+    SHA256_STEP_S (SHA256_F0o, SHA256_F1o, b, c, d, e, f, g, h, a, wf_t, k_sha256[i + 15]); \
+  }
+
+  ROUND_STEP (0);
+
+  #ifdef _unroll
+  #pragma unroll
+  #endif
+  for (int i = 16; i < 64; i += 16)
+  {
+    ROUND_EXPAND (); ROUND_STEP (i);
+  }
+
+  digest[0] += a;
+  digest[1] += b;
+  digest[2] += c;
+  digest[3] += d;
+  digest[4] += e;
+  digest[5] += f;
+  digest[6] += g;
+  digest[7] += h;
+}
+
+void hmac_sha256_pad_S (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], u32 ipad[8], u32 opad[8])
+{
+  w0[0] = w0[0] ^ 0x36363636;
+  w0[1] = w0[1] ^ 0x36363636;
+  w0[2] = w0[2] ^ 0x36363636;
+  w0[3] = w0[3] ^ 0x36363636;
+  w1[0] = w1[0] ^ 0x36363636;
+  w1[1] = w1[1] ^ 0x36363636;
+  w1[2] = w1[2] ^ 0x36363636;
+  w1[3] = w1[3] ^ 0x36363636;
+  w2[0] = w2[0] ^ 0x36363636;
+  w2[1] = w2[1] ^ 0x36363636;
+  w2[2] = w2[2] ^ 0x36363636;
+  w2[3] = w2[3] ^ 0x36363636;
+  w3[0] = w3[0] ^ 0x36363636;
+  w3[1] = w3[1] ^ 0x36363636;
+  w3[2] = w3[2] ^ 0x36363636;
+  w3[3] = w3[3] ^ 0x36363636;
+
+  ipad[0] = SHA256M_A;
+  ipad[1] = SHA256M_B;
+  ipad[2] = SHA256M_C;
+  ipad[3] = SHA256M_D;
+  ipad[4] = SHA256M_E;
+  ipad[5] = SHA256M_F;
+  ipad[6] = SHA256M_G;
+  ipad[7] = SHA256M_H;
+
+  sha256_transform_S (w0, w1, w2, w3, ipad);
+
+  w0[0] = w0[0] ^ 0x6a6a6a6a;
+  w0[1] = w0[1] ^ 0x6a6a6a6a;
+  w0[2] = w0[2] ^ 0x6a6a6a6a;
+  w0[3] = w0[3] ^ 0x6a6a6a6a;
+  w1[0] = w1[0] ^ 0x6a6a6a6a;
+  w1[1] = w1[1] ^ 0x6a6a6a6a;
+  w1[2] = w1[2] ^ 0x6a6a6a6a;
+  w1[3] = w1[3] ^ 0x6a6a6a6a;
+  w2[0] = w2[0] ^ 0x6a6a6a6a;
+  w2[1] = w2[1] ^ 0x6a6a6a6a;
+  w2[2] = w2[2] ^ 0x6a6a6a6a;
+  w2[3] = w2[3] ^ 0x6a6a6a6a;
+  w3[0] = w3[0] ^ 0x6a6a6a6a;
+  w3[1] = w3[1] ^ 0x6a6a6a6a;
+  w3[2] = w3[2] ^ 0x6a6a6a6a;
+  w3[3] = w3[3] ^ 0x6a6a6a6a;
+
+  opad[0] = SHA256M_A;
+  opad[1] = SHA256M_B;
+  opad[2] = SHA256M_C;
+  opad[3] = SHA256M_D;
+  opad[4] = SHA256M_E;
+  opad[5] = SHA256M_F;
+  opad[6] = SHA256M_G;
+  opad[7] = SHA256M_H;
+
+  sha256_transform_S (w0, w1, w2, w3, opad);
+}
+
+void hmac_sha256_run_S (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], u32 ipad[8], u32 opad[8], u32 digest[8])
+{
+  digest[0] = ipad[0];
+  digest[1] = ipad[1];
+  digest[2] = ipad[2];
+  digest[3] = ipad[3];
+  digest[4] = ipad[4];
+  digest[5] = ipad[5];
+  digest[6] = ipad[6];
+  digest[7] = ipad[7];
+
+  sha256_transform_S (w0, w1, w2, w3, digest);
+
+  w0[0] = digest[0];
+  w0[1] = digest[1];
+  w0[2] = digest[2];
+  w0[3] = digest[3];
+  w1[0] = digest[4];
+  w1[1] = digest[5];
+  w1[2] = digest[6];
+  w1[3] = digest[7];
+  w2[0] = 0x80000000;
+  w2[1] = 0;
+  w2[2] = 0;
+  w2[3] = 0;
+  w3[0] = 0;
+  w3[1] = 0;
+  w3[2] = 0;
+  w3[3] = (64 + 32) * 8;
+
+  digest[0] = opad[0];
+  digest[1] = opad[1];
+  digest[2] = opad[2];
+  digest[3] = opad[3];
+  digest[4] = opad[4];
+  digest[5] = opad[5];
+  digest[6] = opad[6];
+  digest[7] = opad[7];
+
+  sha256_transform_S (w0, w1, w2, w3, digest);
+}
+
+void sha256_transform_V (const u32x w0[4], const u32x w1[4], const u32x w2[4], const u32x w3[4], u32x digest[8])
+{
+  u32x a = digest[0];
+  u32x b = digest[1];
+  u32x c = digest[2];
+  u32x d = digest[3];
+  u32x e = digest[4];
+  u32x f = digest[5];
+  u32x g = digest[6];
+  u32x h = digest[7];
+
+  u32x w0_t = w0[0];
+  u32x w1_t = w0[1];
+  u32x w2_t = w0[2];
+  u32x w3_t = w0[3];
+  u32x w4_t = w1[0];
+  u32x w5_t = w1[1];
+  u32x w6_t = w1[2];
+  u32x w7_t = w1[3];
+  u32x w8_t = w2[0];
+  u32x w9_t = w2[1];
+  u32x wa_t = w2[2];
+  u32x wb_t = w2[3];
+  u32x wc_t = w3[0];
+  u32x wd_t = w3[1];
+  u32x we_t = w3[2];
+  u32x wf_t = w3[3];
+
   #define ROUND_EXPAND()                            \
   {                                                 \
     w0_t = SHA256_EXPAND (we_t, w9_t, w1_t, w0_t);  \
@@ -129,7 +320,7 @@ void sha256_transform (const u32 w0[4], const u32 w1[4], const u32 w2[4], const
   digest[7] += h;
 }
 
-void hmac_sha256_pad (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], u32 ipad[8], u32 opad[8])
+void hmac_sha256_pad_V (u32x w0[4], u32x w1[4], u32x w2[4], u32x w3[4], u32x ipad[8], u32x opad[8])
 {
   w0[0] = w0[0] ^ 0x36363636;
   w0[1] = w0[1] ^ 0x36363636;
@@ -157,7 +348,7 @@ void hmac_sha256_pad (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], u32 ipad[8], u
   ipad[6] = SHA256M_G;
   ipad[7] = SHA256M_H;
 
-  sha256_transform (w0, w1, w2, w3, ipad);
+  sha256_transform_V (w0, w1, w2, w3, ipad);
 
   w0[0] = w0[0] ^ 0x6a6a6a6a;
   w0[1] = w0[1] ^ 0x6a6a6a6a;
@@ -185,10 +376,10 @@ void hmac_sha256_pad (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], u32 ipad[8], u
   opad[6] = SHA256M_G;
   opad[7] = SHA256M_H;
 
-  sha256_transform (w0, w1, w2, w3, opad);
+  sha256_transform_V (w0, w1, w2, w3, opad);
 }
 
-void hmac_sha256_run (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], u32 ipad[8], u32 opad[8], u32 digest[8])
+void hmac_sha256_run_V (u32x w0[4], u32x w1[4], u32x w2[4], u32x w3[4], u32x ipad[8], u32x opad[8], u32x digest[8])
 {
   digest[0] = ipad[0];
   digest[1] = ipad[1];
@@ -199,7 +390,7 @@ void hmac_sha256_run (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], u32 ipad[8], u
   digest[6] = ipad[6];
   digest[7] = ipad[7];
 
-  sha256_transform (w0, w1, w2, w3, digest);
+  sha256_transform_V (w0, w1, w2, w3, digest);
 
   w0[0] = digest[0];
   w0[1] = digest[1];
@@ -227,7 +418,7 @@ void hmac_sha256_run (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], u32 ipad[8], u
   digest[6] = opad[6];
   digest[7] = opad[7];
 
-  sha256_transform (w0, w1, w2, w3, digest);
+  sha256_transform_V (w0, w1, w2, w3, digest);
 }
 
 __kernel void m10900_init (__global pw_t *pws, __global kernel_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global pbkdf2_sha256_tmp_t *tmps, __global void *hooks, __global u32 *bitmaps_buf_s1_a, __global u32 *bitmaps_buf_s1_b, __global u32 *bitmaps_buf_s1_c, __global u32 *bitmaps_buf_s1_d, __global u32 *bitmaps_buf_s2_a, __global u32 *bitmaps_buf_s2_b, __global u32 *bitmaps_buf_s2_c, __global u32 *bitmaps_buf_s2_d, __global plain_t *plains_buf, __global digest_t *digests_buf, __global u32 *hashes_shown, __global salt_t *salt_bufs, __global pbkdf2_sha256_t *esalt_bufs, __global u32 *d_return_buf, __global u32 *d_scryptV_buf, const u32 bitmap_mask, const u32 bitmap_shift1, const u32 bitmap_shift2, const u32 salt_pos, const u32 loop_pos, const u32 loop_cnt, const u32 il_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
@@ -242,31 +433,31 @@ __kernel void m10900_init (__global pw_t *pws, __global kernel_rule_t *rules_buf
 
   u32 w0[4];
 
-  w0[0] = swap32 (pws[gid].i[ 0]);
-  w0[1] = swap32 (pws[gid].i[ 1]);
-  w0[2] = swap32 (pws[gid].i[ 2]);
-  w0[3] = swap32 (pws[gid].i[ 3]);
+  w0[0] = swap32_S (pws[gid].i[ 0]);
+  w0[1] = swap32_S (pws[gid].i[ 1]);
+  w0[2] = swap32_S (pws[gid].i[ 2]);
+  w0[3] = swap32_S (pws[gid].i[ 3]);
 
   u32 w1[4];
 
-  w1[0] = swap32 (pws[gid].i[ 4]);
-  w1[1] = swap32 (pws[gid].i[ 5]);
-  w1[2] = swap32 (pws[gid].i[ 6]);
-  w1[3] = swap32 (pws[gid].i[ 7]);
+  w1[0] = swap32_S (pws[gid].i[ 4]);
+  w1[1] = swap32_S (pws[gid].i[ 5]);
+  w1[2] = swap32_S (pws[gid].i[ 6]);
+  w1[3] = swap32_S (pws[gid].i[ 7]);
 
   u32 w2[4];
 
-  w2[0] = swap32 (pws[gid].i[ 8]);
-  w2[1] = swap32 (pws[gid].i[ 9]);
-  w2[2] = swap32 (pws[gid].i[10]);
-  w2[3] = swap32 (pws[gid].i[11]);
+  w2[0] = swap32_S (pws[gid].i[ 8]);
+  w2[1] = swap32_S (pws[gid].i[ 9]);
+  w2[2] = swap32_S (pws[gid].i[10]);
+  w2[3] = swap32_S (pws[gid].i[11]);
 
   u32 w3[4];
 
-  w3[0] = swap32 (pws[gid].i[12]);
-  w3[1] = swap32 (pws[gid].i[13]);
-  w3[2] = swap32 (pws[gid].i[14]);
-  w3[3] = swap32 (pws[gid].i[15]);
+  w3[0] = swap32_S (pws[gid].i[12]);
+  w3[1] = swap32_S (pws[gid].i[13]);
+  w3[2] = swap32_S (pws[gid].i[14]);
+  w3[3] = swap32_S (pws[gid].i[15]);
 
   /**
    * salt
@@ -279,27 +470,27 @@ __kernel void m10900_init (__global pw_t *pws, __global kernel_rule_t *rules_buf
   u32 esalt_buf2[4];
   u32 esalt_buf3[4];
 
-  esalt_buf0[0] = swap32 (esalt_bufs[salt_pos].salt_buf[ 0]);
-  esalt_buf0[1] = swap32 (esalt_bufs[salt_pos].salt_buf[ 1]);
-  esalt_buf0[2] = swap32 (esalt_bufs[salt_pos].salt_buf[ 2]);
-  esalt_buf0[3] = swap32 (esalt_bufs[salt_pos].salt_buf[ 3]);
-  esalt_buf1[0] = swap32 (esalt_bufs[salt_pos].salt_buf[ 4]);
-  esalt_buf1[1] = swap32 (esalt_bufs[salt_pos].salt_buf[ 5]);
-  esalt_buf1[2] = swap32 (esalt_bufs[salt_pos].salt_buf[ 6]);
-  esalt_buf1[3] = swap32 (esalt_bufs[salt_pos].salt_buf[ 7]);
-  esalt_buf2[0] = swap32 (esalt_bufs[salt_pos].salt_buf[ 8]);
-  esalt_buf2[1] = swap32 (esalt_bufs[salt_pos].salt_buf[ 9]);
-  esalt_buf2[2] = swap32 (esalt_bufs[salt_pos].salt_buf[10]);
-  esalt_buf2[3] = swap32 (esalt_bufs[salt_pos].salt_buf[11]);
-  esalt_buf3[0] = swap32 (esalt_bufs[salt_pos].salt_buf[12]);
-  esalt_buf3[1] = swap32 (esalt_bufs[salt_pos].salt_buf[13]);
+  esalt_buf0[0] = swap32_S (esalt_bufs[salt_pos].salt_buf[ 0]);
+  esalt_buf0[1] = swap32_S (esalt_bufs[salt_pos].salt_buf[ 1]);
+  esalt_buf0[2] = swap32_S (esalt_bufs[salt_pos].salt_buf[ 2]);
+  esalt_buf0[3] = swap32_S (esalt_bufs[salt_pos].salt_buf[ 3]);
+  esalt_buf1[0] = swap32_S (esalt_bufs[salt_pos].salt_buf[ 4]);
+  esalt_buf1[1] = swap32_S (esalt_bufs[salt_pos].salt_buf[ 5]);
+  esalt_buf1[2] = swap32_S (esalt_bufs[salt_pos].salt_buf[ 6]);
+  esalt_buf1[3] = swap32_S (esalt_bufs[salt_pos].salt_buf[ 7]);
+  esalt_buf2[0] = swap32_S (esalt_bufs[salt_pos].salt_buf[ 8]);
+  esalt_buf2[1] = swap32_S (esalt_bufs[salt_pos].salt_buf[ 9]);
+  esalt_buf2[2] = swap32_S (esalt_bufs[salt_pos].salt_buf[10]);
+  esalt_buf2[3] = swap32_S (esalt_bufs[salt_pos].salt_buf[11]);
+  esalt_buf3[0] = swap32_S (esalt_bufs[salt_pos].salt_buf[12]);
+  esalt_buf3[1] = swap32_S (esalt_bufs[salt_pos].salt_buf[13]);
   esalt_buf3[2] = 0;
   esalt_buf3[3] = (64 + salt_len + 4) * 8;
 
   u32 ipad[8];
   u32 opad[8];
 
-  hmac_sha256_pad (w0, w1, w2, w3, ipad, opad);
+  hmac_sha256_pad_S (w0, w1, w2, w3, ipad, opad);
 
   tmps[gid].ipad[0] = ipad[0];
   tmps[gid].ipad[1] = ipad[1];
@@ -323,7 +514,7 @@ __kernel void m10900_init (__global pw_t *pws, __global kernel_rule_t *rules_buf
   {
     u32 dgst[8];
 
-    hmac_sha256_run (esalt_buf0, esalt_buf1, esalt_buf2, esalt_buf3, ipad, opad, dgst);
+    hmac_sha256_run_S (esalt_buf0, esalt_buf1, esalt_buf2, esalt_buf3, ipad, opad, dgst);
 
     tmps[gid].dgst[i + 0] = dgst[0];
     tmps[gid].dgst[i + 1] = dgst[1];
@@ -351,58 +542,56 @@ __kernel void m10900_loop (__global pw_t *pws, __global kernel_rule_t *rules_buf
 
   if (gid >= gid_max) return;
 
-  u32 ipad[8];
-
-  ipad[0] = tmps[gid].ipad[0];
-  ipad[1] = tmps[gid].ipad[1];
-  ipad[2] = tmps[gid].ipad[2];
-  ipad[3] = tmps[gid].ipad[3];
-  ipad[4] = tmps[gid].ipad[4];
-  ipad[5] = tmps[gid].ipad[5];
-  ipad[6] = tmps[gid].ipad[6];
-  ipad[7] = tmps[gid].ipad[7];
-
-  u32 opad[8];
-
-  opad[0] = tmps[gid].opad[0];
-  opad[1] = tmps[gid].opad[1];
-  opad[2] = tmps[gid].opad[2];
-  opad[3] = tmps[gid].opad[3];
-  opad[4] = tmps[gid].opad[4];
-  opad[5] = tmps[gid].opad[5];
-  opad[6] = tmps[gid].opad[6];
-  opad[7] = tmps[gid].opad[7];
+  u32x ipad[8];
+  u32x opad[8];
+
+  ipad[0] = packv (tmps, ipad, gid, 0);
+  ipad[1] = packv (tmps, ipad, gid, 1);
+  ipad[2] = packv (tmps, ipad, gid, 2);
+  ipad[3] = packv (tmps, ipad, gid, 3);
+  ipad[4] = packv (tmps, ipad, gid, 4);
+  ipad[5] = packv (tmps, ipad, gid, 5);
+  ipad[6] = packv (tmps, ipad, gid, 6);
+  ipad[7] = packv (tmps, ipad, gid, 7);
+
+  opad[0] = packv (tmps, opad, gid, 0);
+  opad[1] = packv (tmps, opad, gid, 1);
+  opad[2] = packv (tmps, opad, gid, 2);
+  opad[3] = packv (tmps, opad, gid, 3);
+  opad[4] = packv (tmps, opad, gid, 4);
+  opad[5] = packv (tmps, opad, gid, 5);
+  opad[6] = packv (tmps, opad, gid, 6);
+  opad[7] = packv (tmps, opad, gid, 7);
 
   for (u32 i = 0; i < 8; i += 8)
   {
-    u32 dgst[8];
-
-    dgst[0] = tmps[gid].dgst[i + 0];
-    dgst[1] = tmps[gid].dgst[i + 1];
-    dgst[2] = tmps[gid].dgst[i + 2];
-    dgst[3] = tmps[gid].dgst[i + 3];
-    dgst[4] = tmps[gid].dgst[i + 4];
-    dgst[5] = tmps[gid].dgst[i + 5];
-    dgst[6] = tmps[gid].dgst[i + 6];
-    dgst[7] = tmps[gid].dgst[i + 7];
-
-    u32 out[8];
-
-    out[0] = tmps[gid].out[i + 0];
-    out[1] = tmps[gid].out[i + 1];
-    out[2] = tmps[gid].out[i + 2];
-    out[3] = tmps[gid].out[i + 3];
-    out[4] = tmps[gid].out[i + 4];
-    out[5] = tmps[gid].out[i + 5];
-    out[6] = tmps[gid].out[i + 6];
-    out[7] = tmps[gid].out[i + 7];
+    u32x dgst[8];
+    u32x out[8];
+
+    dgst[0] = packv (tmps, dgst, gid, 0);
+    dgst[1] = packv (tmps, dgst, gid, 1);
+    dgst[2] = packv (tmps, dgst, gid, 2);
+    dgst[3] = packv (tmps, dgst, gid, 3);
+    dgst[4] = packv (tmps, dgst, gid, 4);
+    dgst[5] = packv (tmps, dgst, gid, 5);
+    dgst[6] = packv (tmps, dgst, gid, 6);
+    dgst[7] = packv (tmps, dgst, gid, 7);
+
+    out[0] = packv (tmps, out, gid, 0);
+    out[1] = packv (tmps, out, gid, 1);
+    out[2] = packv (tmps, out, gid, 2);
+    out[3] = packv (tmps, out, gid, 3);
+    out[4] = packv (tmps, out, gid, 4);
+    out[5] = packv (tmps, out, gid, 5);
+    out[6] = packv (tmps, out, gid, 6);
+    out[7] = packv (tmps, out, gid, 7);
 
     for (u32 j = 0; j < loop_cnt; j++)
     {
-      u32 w0[4];
-      u32 w1[4];
-      u32 w2[4];
-      u32 w3[4];
+      u32x w0[4];
+      u32x w1[4];
+      u32x w2[4];
+      u32x w3[4];
 
       w0[0] = dgst[0];
       w0[1] = dgst[1];
@@ -421,7 +610,7 @@ __kernel void m10900_loop (__global pw_t *pws, __global kernel_rule_t *rules_buf
       w3[2] = 0;
       w3[3] = (64 + 32) * 8;
 
-      hmac_sha256_run (w0, w1, w2, w3, ipad, opad, dgst);
+      hmac_sha256_run_V (w0, w1, w2, w3, ipad, opad, dgst);
 
       out[0] ^= dgst[0];
       out[1] ^= dgst[1];
@@ -433,23 +622,23 @@ __kernel void m10900_loop (__global pw_t *pws, __global kernel_rule_t *rules_buf
       out[7] ^= dgst[7];
     }
 
-    tmps[gid].dgst[i + 0] = dgst[0];
-    tmps[gid].dgst[i + 1] = dgst[1];
-    tmps[gid].dgst[i + 2] = dgst[2];
-    tmps[gid].dgst[i + 3] = dgst[3];
-    tmps[gid].dgst[i + 4] = dgst[4];
-    tmps[gid].dgst[i + 5] = dgst[5];
-    tmps[gid].dgst[i + 6] = dgst[6];
-    tmps[gid].dgst[i + 7] = dgst[7];
-
-    tmps[gid].out[i + 0] = out[0];
-    tmps[gid].out[i + 1] = out[1];
-    tmps[gid].out[i + 2] = out[2];
-    tmps[gid].out[i + 3] = out[3];
-    tmps[gid].out[i + 4] = out[4];
-    tmps[gid].out[i + 5] = out[5];
-    tmps[gid].out[i + 6] = out[6];
-    tmps[gid].out[i + 7] = out[7];
+    unpackv (tmps, dgst, gid, 0, dgst[0]);
+    unpackv (tmps, dgst, gid, 1, dgst[1]);
+    unpackv (tmps, dgst, gid, 2, dgst[2]);
+    unpackv (tmps, dgst, gid, 3, dgst[3]);
+    unpackv (tmps, dgst, gid, 4, dgst[4]);
+    unpackv (tmps, dgst, gid, 5, dgst[5]);
+    unpackv (tmps, dgst, gid, 6, dgst[6]);
+    unpackv (tmps, dgst, gid, 7, dgst[7]);
+
+    unpackv (tmps, out, gid, 0, out[0]);
+    unpackv (tmps, out, gid, 1, out[1]);
+    unpackv (tmps, out, gid, 2, out[2]);
+    unpackv (tmps, out, gid, 3, out[3]);
+    unpackv (tmps, out, gid, 4, out[4]);
+    unpackv (tmps, out, gid, 5, out[5]);
+    unpackv (tmps, out, gid, 6, out[6]);
+    unpackv (tmps, out, gid, 7, out[7]);
   }
 }
 
diff --git a/OpenCL/m11900.cl b/OpenCL/m11900.cl
index 33bcc1bdc..53eec8f86 100644
--- a/OpenCL/m11900.cl
+++ b/OpenCL/m11900.cl
@@ -5,6 +5,8 @@
 
 #define _PBKDF2_MD5_
 
+#define NEW_SIMD_CODE
+
 #include "include/constants.h"
 #include "include/kernel_vendor.h"
 
@@ -14,14 +16,14 @@
 #define DGST_R3 3
 
 #include "include/kernel_functions.c"
-
 #include "OpenCL/types_ocl.c"
 #include "OpenCL/common.c"
+#include "OpenCL/simd.c"
 
 #define COMPARE_S "OpenCL/check_single_comp4.c"
 #define COMPARE_M "OpenCL/check_multi_comp4.c"
 
-void md5_transform (const u32 w0[4], const u32 w1[4], const u32 w2[4], const u32 w3[4], u32 digest[4])
+void md5_transform_S (const u32 w0[4], const u32 w1[4], const u32 w2[4], const u32 w3[4], u32 digest[4])
 {
   u32 a = digest[0];
   u32 b = digest[1];
@@ -45,6 +47,189 @@ void md5_transform (const u32 w0[4], const u32 w1[4], const u32 w2[4], const u32
   u32 we_t = w3[2];
   u32 wf_t = w3[3];
 
+  MD5_STEP_S (MD5_Fo, a, b, c, d, w0_t, MD5C00, MD5S00);
+  MD5_STEP_S (MD5_Fo, d, a, b, c, w1_t, MD5C01, MD5S01);
+  MD5_STEP_S (MD5_Fo, c, d, a, b, w2_t, MD5C02, MD5S02);
+  MD5_STEP_S (MD5_Fo, b, c, d, a, w3_t, MD5C03, MD5S03);
+  MD5_STEP_S (MD5_Fo, a, b, c, d, w4_t, MD5C04, MD5S00);
+  MD5_STEP_S (MD5_Fo, d, a, b, c, w5_t, MD5C05, MD5S01);
+  MD5_STEP_S (MD5_Fo, c, d, a, b, w6_t, MD5C06, MD5S02);
+  MD5_STEP_S (MD5_Fo, b, c, d, a, w7_t, MD5C07, MD5S03);
+  MD5_STEP_S (MD5_Fo, a, b, c, d, w8_t, MD5C08, MD5S00);
+  MD5_STEP_S (MD5_Fo, d, a, b, c, w9_t, MD5C09, MD5S01);
+  MD5_STEP_S (MD5_Fo, c, d, a, b, wa_t, MD5C0a, MD5S02);
+  MD5_STEP_S (MD5_Fo, b, c, d, a, wb_t, MD5C0b, MD5S03);
+  MD5_STEP_S (MD5_Fo, a, b, c, d, wc_t, MD5C0c, MD5S00);
+  MD5_STEP_S (MD5_Fo, d, a, b, c, wd_t, MD5C0d, MD5S01);
+  MD5_STEP_S (MD5_Fo, c, d, a, b, we_t, MD5C0e, MD5S02);
+  MD5_STEP_S (MD5_Fo, b, c, d, a, wf_t, MD5C0f, MD5S03);
+
+  MD5_STEP_S (MD5_Go, a, b, c, d, w1_t, MD5C10, MD5S10);
+  MD5_STEP_S (MD5_Go, d, a, b, c, w6_t, MD5C11, MD5S11);
+  MD5_STEP_S (MD5_Go, c, d, a, b, wb_t, MD5C12, MD5S12);
+  MD5_STEP_S (MD5_Go, b, c, d, a, w0_t, MD5C13, MD5S13);
+  MD5_STEP_S (MD5_Go, a, b, c, d, w5_t, MD5C14, MD5S10);
+  MD5_STEP_S (MD5_Go, d, a, b, c, wa_t, MD5C15, MD5S11);
+  MD5_STEP_S (MD5_Go, c, d, a, b, wf_t, MD5C16, MD5S12);
+  MD5_STEP_S (MD5_Go, b, c, d, a, w4_t, MD5C17, MD5S13);
+  MD5_STEP_S (MD5_Go, a, b, c, d, w9_t, MD5C18, MD5S10);
+  MD5_STEP_S (MD5_Go, d, a, b, c, we_t, MD5C19, MD5S11);
+  MD5_STEP_S (MD5_Go, c, d, a, b, w3_t, MD5C1a, MD5S12);
+  MD5_STEP_S (MD5_Go, b, c, d, a, w8_t, MD5C1b, MD5S13);
+  MD5_STEP_S (MD5_Go, a, b, c, d, wd_t, MD5C1c, MD5S10);
+  MD5_STEP_S (MD5_Go, d, a, b, c, w2_t, MD5C1d, MD5S11);
+  MD5_STEP_S (MD5_Go, c, d, a, b, w7_t, MD5C1e, MD5S12);
+  MD5_STEP_S (MD5_Go, b, c, d, a, wc_t, MD5C1f, MD5S13);
+
+  MD5_STEP_S (MD5_H , a, b, c, d, w5_t, MD5C20, MD5S20);
+  MD5_STEP_S (MD5_H , d, a, b, c, w8_t, MD5C21, MD5S21);
+  MD5_STEP_S (MD5_H , c, d, a, b, wb_t, MD5C22, MD5S22);
+  MD5_STEP_S (MD5_H , b, c, d, a, we_t, MD5C23, MD5S23);
+  MD5_STEP_S (MD5_H , a, b, c, d, w1_t, MD5C24, MD5S20);
+  MD5_STEP_S (MD5_H , d, a, b, c, w4_t, MD5C25, MD5S21);
+  MD5_STEP_S (MD5_H , c, d, a, b, w7_t, MD5C26, MD5S22);
+  MD5_STEP_S (MD5_H , b, c, d, a, wa_t, MD5C27, MD5S23);
+  MD5_STEP_S (MD5_H , a, b, c, d, wd_t, MD5C28, MD5S20);
+  MD5_STEP_S (MD5_H , d, a, b, c, w0_t, MD5C29, MD5S21);
+  MD5_STEP_S (MD5_H , c, d, a, b, w3_t, MD5C2a, MD5S22);
+  MD5_STEP_S (MD5_H , b, c, d, a, w6_t, MD5C2b, MD5S23);
+  MD5_STEP_S (MD5_H , a, b, c, d, w9_t, MD5C2c, MD5S20);
+  MD5_STEP_S (MD5_H , d, a, b, c, wc_t, MD5C2d, MD5S21);
+  MD5_STEP_S (MD5_H , c, d, a, b, wf_t, MD5C2e, MD5S22);
+  MD5_STEP_S (MD5_H , b, c, d, a, w2_t, MD5C2f, MD5S23);
+
+  MD5_STEP_S (MD5_I , a, b, c, d, w0_t, MD5C30, MD5S30);
+  MD5_STEP_S (MD5_I , d, a, b, c, w7_t, MD5C31, MD5S31);
+  MD5_STEP_S (MD5_I , c, d, a, b, we_t, MD5C32, MD5S32);
+  MD5_STEP_S (MD5_I , b, c, d, a, w5_t, MD5C33, MD5S33);
+  MD5_STEP_S (MD5_I , a, b, c, d, wc_t, MD5C34, MD5S30);
+  MD5_STEP_S (MD5_I , d, a, b, c, w3_t, MD5C35, MD5S31);
+  MD5_STEP_S (MD5_I , c, d, a, b, wa_t, MD5C36, MD5S32);
+  MD5_STEP_S (MD5_I , b, c, d, a, w1_t, MD5C37, MD5S33);
+  MD5_STEP_S (MD5_I , a, b, c, d, w8_t, MD5C38, MD5S30);
+  MD5_STEP_S (MD5_I , d, a, b, c, wf_t, MD5C39, MD5S31);
+  MD5_STEP_S (MD5_I , c, d, a, b, w6_t, MD5C3a, MD5S32);
+  MD5_STEP_S (MD5_I , b, c, d, a, wd_t, MD5C3b, MD5S33);
+  MD5_STEP_S (MD5_I , a, b, c, d, w4_t, MD5C3c, MD5S30);
+  MD5_STEP_S (MD5_I , d, a, b, c, wb_t, MD5C3d, MD5S31);
+  MD5_STEP_S (MD5_I , c, d, a, b, w2_t, MD5C3e, MD5S32);
+  MD5_STEP_S (MD5_I , b, c, d, a, w9_t, MD5C3f, MD5S33);
+
+  digest[0] += a;
+  digest[1] += b;
+  digest[2] += c;
+  digest[3] += d;
+}
+
+void hmac_md5_pad_S (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], u32 ipad[4], u32 opad[4])
+{
+  w0[0] = w0[0] ^ 0x36363636;
+  w0[1] = w0[1] ^ 0x36363636;
+  w0[2] = w0[2] ^ 0x36363636;
+  w0[3] = w0[3] ^ 0x36363636;
+  w1[0] = w1[0] ^ 0x36363636;
+  w1[1] = w1[1] ^ 0x36363636;
+  w1[2] = w1[2] ^ 0x36363636;
+  w1[3] = w1[3] ^ 0x36363636;
+  w2[0] = w2[0] ^ 0x36363636;
+  w2[1] = w2[1] ^ 0x36363636;
+  w2[2] = w2[2] ^ 0x36363636;
+  w2[3] = w2[3] ^ 0x36363636;
+  w3[0] = w3[0] ^ 0x36363636;
+  w3[1] = w3[1] ^ 0x36363636;
+  w3[2] = w3[2] ^ 0x36363636;
+  w3[3] = w3[3] ^ 0x36363636;
+
+  ipad[0] = MD5M_A;
+  ipad[1] = MD5M_B;
+  ipad[2] = MD5M_C;
+  ipad[3] = MD5M_D;
+
+  md5_transform_S (w0, w1, w2, w3, ipad);
+
+  w0[0] = w0[0] ^ 0x6a6a6a6a;
+  w0[1] = w0[1] ^ 0x6a6a6a6a;
+  w0[2] = w0[2] ^ 0x6a6a6a6a;
+  w0[3] = w0[3] ^ 0x6a6a6a6a;
+  w1[0] = w1[0] ^ 0x6a6a6a6a;
+  w1[1] = w1[1] ^ 0x6a6a6a6a;
+  w1[2] = w1[2] ^ 0x6a6a6a6a;
+  w1[3] = w1[3] ^ 0x6a6a6a6a;
+  w2[0] = w2[0] ^ 0x6a6a6a6a;
+  w2[1] = w2[1] ^ 0x6a6a6a6a;
+  w2[2] = w2[2] ^ 0x6a6a6a6a;
+  w2[3] = w2[3] ^ 0x6a6a6a6a;
+  w3[0] = w3[0] ^ 0x6a6a6a6a;
+  w3[1] = w3[1] ^ 0x6a6a6a6a;
+  w3[2] = w3[2] ^ 0x6a6a6a6a;
+  w3[3] = w3[3] ^ 0x6a6a6a6a;
+
+  opad[0] = MD5M_A;
+  opad[1] = MD5M_B;
+  opad[2] = MD5M_C;
+  opad[3] = MD5M_D;
+
+  md5_transform_S (w0, w1, w2, w3, opad);
+}
+
+void hmac_md5_run_S (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], u32 ipad[4], u32 opad[4], u32 digest[4])
+{
+  digest[0] = ipad[0];
+  digest[1] = ipad[1];
+  digest[2] = ipad[2];
+  digest[3] = ipad[3];
+
+  md5_transform_S (w0, w1, w2, w3, digest);
+
+  w0[0] = digest[0];
+  w0[1] = digest[1];
+  w0[2] = digest[2];
+  w0[3] = digest[3];
+  w1[0] = 0x80;
+  w1[1] = 0;
+  w1[2] = 0;
+  w1[3] = 0;
+  w2[0] = 0;
+  w2[1] = 0;
+  w2[2] = 0;
+  w2[3] = 0;
+  w3[0] = 0;
+  w3[1] = 0;
+  w3[2] = (64 + 16) * 8;
+  w3[3] = 0;
+
+  digest[0] = opad[0];
+  digest[1] = opad[1];
+  digest[2] = opad[2];
+  digest[3] = opad[3];
+
+  md5_transform_S (w0, w1, w2, w3, digest);
+}
+
+void md5_transform_V (const u32x w0[4], const u32x w1[4], const u32x w2[4], const u32x w3[4], u32x digest[4])
+{
+  u32x a = digest[0];
+  u32x b = digest[1];
+  u32x c = digest[2];
+  u32x d = digest[3];
+
+  u32x w0_t = w0[0];
+  u32x w1_t = w0[1];
+  u32x w2_t = w0[2];
+  u32x w3_t = w0[3];
+  u32x w4_t = w1[0];
+  u32x w5_t = w1[1];
+  u32x w6_t = w1[2];
+  u32x w7_t = w1[3];
+  u32x w8_t = w2[0];
+  u32x w9_t = w2[1];
+  u32x wa_t = w2[2];
+  u32x wb_t = w2[3];
+  u32x wc_t = w3[0];
+  u32x wd_t = w3[1];
+  u32x we_t = w3[2];
+  u32x wf_t = w3[3];
+
   MD5_STEP (MD5_Fo, a, b, c, d, w0_t, MD5C00, MD5S00);
   MD5_STEP (MD5_Fo, d, a, b, c, w1_t, MD5C01, MD5S01);
   MD5_STEP (MD5_Fo, c, d, a, b, w2_t, MD5C02, MD5S02);
@@ -119,7 +304,7 @@ void md5_transform (const u32 w0[4], const u32 w1[4], const u32 w2[4], const u32
   digest[3] += d;
 }
 
-void hmac_md5_pad (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], u32 ipad[4], u32 opad[4])
+void hmac_md5_pad_V (u32x w0[4], u32x w1[4], u32x w2[4], u32x w3[4], u32x ipad[4], u32x opad[4])
 {
   w0[0] = w0[0] ^ 0x36363636;
   w0[1] = w0[1] ^ 0x36363636;
@@ -143,7 +328,7 @@ void hmac_md5_pad (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], u32 ipad[4], u32
   ipad[2] = MD5M_C;
   ipad[3] = MD5M_D;
 
-  md5_transform (w0, w1, w2, w3, ipad);
+  md5_transform_V (w0, w1, w2, w3, ipad);
 
   w0[0] = w0[0] ^ 0x6a6a6a6a;
   w0[1] = w0[1] ^ 0x6a6a6a6a;
@@ -167,17 +352,17 @@ void hmac_md5_pad (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], u32 ipad[4], u32
   opad[2] = MD5M_C;
   opad[3] = MD5M_D;
 
-  md5_transform (w0, w1, w2, w3, opad);
+  md5_transform_V (w0, w1, w2, w3, opad);
 }
 
-void hmac_md5_run (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], u32 ipad[4], u32 opad[4], u32 digest[4])
+void hmac_md5_run_V (u32x w0[4], u32x w1[4], u32x w2[4], u32x w3[4], u32x ipad[4], u32x opad[4], u32x digest[4])
 {
   digest[0] = ipad[0];
   digest[1] = ipad[1];
   digest[2] = ipad[2];
   digest[3] = ipad[3];
 
-  md5_transform (w0, w1, w2, w3, digest);
+  md5_transform_V (w0, w1, w2, w3, digest);
 
   w0[0] = digest[0];
   w0[1] = digest[1];
@@ -201,7 +386,7 @@ void hmac_md5_run (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], u32 ipad[4], u32
   digest[2] = opad[2];
   digest[3] = opad[3];
 
-  md5_transform (w0, w1, w2, w3, digest);
+  md5_transform_V (w0, w1, w2, w3, digest);
 }
 
 __kernel void m11900_init (__global pw_t *pws, __global kernel_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global pbkdf2_md5_tmp_t *tmps, __global void *hooks, __global u32 *bitmaps_buf_s1_a, __global u32 *bitmaps_buf_s1_b, __global u32 *bitmaps_buf_s1_c, __global u32 *bitmaps_buf_s1_d, __global u32 *bitmaps_buf_s2_a, __global u32 *bitmaps_buf_s2_b, __global u32 *bitmaps_buf_s2_c, __global u32 *bitmaps_buf_s2_d, __global plain_t *plains_buf, __global digest_t *digests_buf, __global u32 *hashes_shown, __global salt_t *salt_bufs, __global pbkdf2_md5_t *esalt_bufs, __global u32 *d_return_buf, __global u32 *d_scryptV_buf, const u32 bitmap_mask, const u32 bitmap_shift1, const u32 bitmap_shift2, const u32 salt_pos, const u32 loop_pos, const u32 loop_cnt, const u32 il_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
@@ -273,7 +458,7 @@ __kernel void m11900_init (__global pw_t *pws, __global kernel_rule_t *rules_buf
   u32 ipad[4];
   u32 opad[4];
 
-  hmac_md5_pad (w0, w1, w2, w3, ipad, opad);
+  hmac_md5_pad_S (w0, w1, w2, w3, ipad, opad);
 
   tmps[gid].ipad[0] = ipad[0];
   tmps[gid].ipad[1] = ipad[1];
@@ -289,7 +474,7 @@ __kernel void m11900_init (__global pw_t *pws, __global kernel_rule_t *rules_buf
   {
     u32 dgst[4];
 
-    hmac_md5_run (esalt_buf0, esalt_buf1, esalt_buf2, esalt_buf3, ipad, opad, dgst);
+    hmac_md5_run_S (esalt_buf0, esalt_buf1, esalt_buf2, esalt_buf3, ipad, opad, dgst);
 
     tmps[gid].dgst[i + 0] = dgst[0];
     tmps[gid].dgst[i + 1] = dgst[1];
@@ -309,42 +494,40 @@ __kernel void m11900_loop (__global pw_t *pws, __global kernel_rule_t *rules_buf
 
   if (gid >= gid_max) return;
 
-  u32 ipad[4];
-
-  ipad[0] = tmps[gid].ipad[0];
-  ipad[1] = tmps[gid].ipad[1];
-  ipad[2] = tmps[gid].ipad[2];
-  ipad[3] = tmps[gid].ipad[3];
+  u32x ipad[4];
+  u32x opad[4];
 
-  u32 opad[4];
+  ipad[0] = packv (tmps, ipad, gid, 0);
+  ipad[1] = packv (tmps, ipad, gid, 1);
+  ipad[2] = packv (tmps, ipad, gid, 2);
+  ipad[3] = packv (tmps, ipad, gid, 3);
 
-  opad[0] = tmps[gid].opad[0];
-  opad[1] = tmps[gid].opad[1];
-  opad[2] = tmps[gid].opad[2];
-  opad[3] = tmps[gid].opad[3];
+  opad[0] = packv (tmps, opad, gid, 0);
+  opad[1] = packv (tmps, opad, gid, 1);
+  opad[2] = packv (tmps, opad, gid, 2);
+  opad[3] = packv (tmps, opad, gid, 3);
 
   for (u32 i = 0; i < 4; i += 4)
   {
-    u32 dgst[4];
-
-    dgst[0] = tmps[gid].dgst[i + 0];
-    dgst[1] = tmps[gid].dgst[i + 1];
-    dgst[2] = tmps[gid].dgst[i + 2];
-    dgst[3] = tmps[gid].dgst[i + 3];
+    u32x dgst[4];
+    u32x out[4];
 
-    u32 out[4];
+    dgst[0] = packv (tmps, dgst, gid, 0);
+    dgst[1] = packv (tmps, dgst, gid, 1);
+    dgst[2] = packv (tmps, dgst, gid, 2);
+    dgst[3] = packv (tmps, dgst, gid, 3);
 
-    out[0] = tmps[gid].out[i + 0];
-    out[1] = tmps[gid].out[i + 1];
-    out[2] = tmps[gid].out[i + 2];
-    out[3] = tmps[gid].out[i + 3];
+    out[0] = packv (tmps, out, gid, 0);
+    out[1] = packv (tmps, out, gid, 1);
+    out[2] = packv (tmps, out, gid, 2);
+    out[3] = packv (tmps, out, gid, 3);
 
     for (u32 j = 0; j < loop_cnt; j++)
     {
-      u32 w0[4];
-      u32 w1[4];
-      u32 w2[4];
-      u32 w3[4];
+      u32x w0[4];
+      u32x w1[4];
+      u32x w2[4];
+      u32x w3[4];
 
       w0[0] = dgst[0];
       w0[1] = dgst[1];
@@ -363,7 +546,7 @@ __kernel void m11900_loop (__global pw_t *pws, __global kernel_rule_t *rules_buf
       w3[2] = (64 + 16) * 8;
       w3[3] = 0;
 
-      hmac_md5_run (w0, w1, w2, w3, ipad, opad, dgst);
+      hmac_md5_run_V (w0, w1, w2, w3, ipad, opad, dgst);
 
       out[0] ^= dgst[0];
       out[1] ^= dgst[1];
@@ -371,15 +554,15 @@ __kernel void m11900_loop (__global pw_t *pws, __global kernel_rule_t *rules_buf
       out[3] ^= dgst[3];
     }
 
-    tmps[gid].dgst[i + 0] = dgst[0];
-    tmps[gid].dgst[i + 1] = dgst[1];
-    tmps[gid].dgst[i + 2] = dgst[2];
-    tmps[gid].dgst[i + 3] = dgst[3];
+    unpackv (tmps, dgst, gid, 0, dgst[0]);
+    unpackv (tmps, dgst, gid, 1, dgst[1]);
+    unpackv (tmps, dgst, gid, 2, dgst[2]);
+    unpackv (tmps, dgst, gid, 3, dgst[3]);
 
-    tmps[gid].out[i + 0] = out[0];
-    tmps[gid].out[i + 1] = out[1];
-    tmps[gid].out[i + 2] = out[2];
-    tmps[gid].out[i + 3] = out[3];
+    unpackv (tmps, out, gid, 0, out[0]);
+    unpackv (tmps, out, gid, 1, out[1]);
+    unpackv (tmps, out, gid, 2, out[2]);
+    unpackv (tmps, out, gid, 3, out[3]);
   }
 }
 
diff --git a/OpenCL/m12000.cl b/OpenCL/m12000.cl
index 00b0aac25..ffbcf867d 100644
--- a/OpenCL/m12000.cl
+++ b/OpenCL/m12000.cl
@@ -5,6 +5,8 @@
 
 #define _PBKDF2_SHA1_
 
+#define NEW_SIMD_CODE
+
 #include "include/constants.h"
 #include "include/kernel_vendor.h"
 
@@ -14,14 +16,14 @@
 #define DGST_R3 3
 
 #include "include/kernel_functions.c"
-
 #include "OpenCL/types_ocl.c"
 #include "OpenCL/common.c"
+#include "OpenCL/simd.c"
 
 #define COMPARE_S "OpenCL/check_single_comp4.c"
 #define COMPARE_M "OpenCL/check_multi_comp4.c"
 
-void sha1_transform (const u32 w0[4], const u32 w1[4], const u32 w2[4], const u32 w3[4], u32 digest[5])
+void sha1_transform_S (const u32 w0[4], const u32 w1[4], const u32 w2[4], const u32 w3[4], u32 digest[5])
 {
   u32 A = digest[0];
   u32 B = digest[1];
@@ -49,6 +51,223 @@ void sha1_transform (const u32 w0[4], const u32 w1[4], const u32 w2[4], const u3
   #undef K
   #define K SHA1C00
 
+  SHA1_STEP_S (SHA1_F0o, A, B, C, D, E, w0_t);
+  SHA1_STEP_S (SHA1_F0o, E, A, B, C, D, w1_t);
+  SHA1_STEP_S (SHA1_F0o, D, E, A, B, C, w2_t);
+  SHA1_STEP_S (SHA1_F0o, C, D, E, A, B, w3_t);
+  SHA1_STEP_S (SHA1_F0o, B, C, D, E, A, w4_t);
+  SHA1_STEP_S (SHA1_F0o, A, B, C, D, E, w5_t);
+  SHA1_STEP_S (SHA1_F0o, E, A, B, C, D, w6_t);
+  SHA1_STEP_S (SHA1_F0o, D, E, A, B, C, w7_t);
+  SHA1_STEP_S (SHA1_F0o, C, D, E, A, B, w8_t);
+  SHA1_STEP_S (SHA1_F0o, B, C, D, E, A, w9_t);
+  SHA1_STEP_S (SHA1_F0o, A, B, C, D, E, wa_t);
+  SHA1_STEP_S (SHA1_F0o, E, A, B, C, D, wb_t);
+  SHA1_STEP_S (SHA1_F0o, D, E, A, B, C, wc_t);
+  SHA1_STEP_S (SHA1_F0o, C, D, E, A, B, wd_t);
+  SHA1_STEP_S (SHA1_F0o, B, C, D, E, A, we_t);
+  SHA1_STEP_S (SHA1_F0o, A, B, C, D, E, wf_t);
+  w0_t = rotl32_S ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP_S (SHA1_F0o, E, A, B, C, D, w0_t);
+  w1_t = rotl32_S ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP_S (SHA1_F0o, D, E, A, B, C, w1_t);
+  w2_t = rotl32_S ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP_S (SHA1_F0o, C, D, E, A, B, w2_t);
+  w3_t = rotl32_S ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP_S (SHA1_F0o, B, C, D, E, A, w3_t);
+
+  #undef K
+  #define K SHA1C01
+
+  w4_t = rotl32_S ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP_S (SHA1_F1, A, B, C, D, E, w4_t);
+  w5_t = rotl32_S ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP_S (SHA1_F1, E, A, B, C, D, w5_t);
+  w6_t = rotl32_S ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP_S (SHA1_F1, D, E, A, B, C, w6_t);
+  w7_t = rotl32_S ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP_S (SHA1_F1, C, D, E, A, B, w7_t);
+  w8_t = rotl32_S ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP_S (SHA1_F1, B, C, D, E, A, w8_t);
+  w9_t = rotl32_S ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP_S (SHA1_F1, A, B, C, D, E, w9_t);
+  wa_t = rotl32_S ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP_S (SHA1_F1, E, A, B, C, D, wa_t);
+  wb_t = rotl32_S ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP_S (SHA1_F1, D, E, A, B, C, wb_t);
+  wc_t = rotl32_S ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP_S (SHA1_F1, C, D, E, A, B, wc_t);
+  wd_t = rotl32_S ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP_S (SHA1_F1, B, C, D, E, A, wd_t);
+  we_t = rotl32_S ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP_S (SHA1_F1, A, B, C, D, E, we_t);
+  wf_t = rotl32_S ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP_S (SHA1_F1, E, A, B, C, D, wf_t);
+  w0_t = rotl32_S ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP_S (SHA1_F1, D, E, A, B, C, w0_t);
+  w1_t = rotl32_S ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP_S (SHA1_F1, C, D, E, A, B, w1_t);
+  w2_t = rotl32_S ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP_S (SHA1_F1, B, C, D, E, A, w2_t);
+  w3_t = rotl32_S ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP_S (SHA1_F1, A, B, C, D, E, w3_t);
+  w4_t = rotl32_S ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP_S (SHA1_F1, E, A, B, C, D, w4_t);
+  w5_t = rotl32_S ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP_S (SHA1_F1, D, E, A, B, C, w5_t);
+  w6_t = rotl32_S ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP_S (SHA1_F1, C, D, E, A, B, w6_t);
+  w7_t = rotl32_S ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP_S (SHA1_F1, B, C, D, E, A, w7_t);
+
+  #undef K
+  #define K SHA1C02
+
+  w8_t = rotl32_S ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP_S (SHA1_F2o, A, B, C, D, E, w8_t);
+  w9_t = rotl32_S ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP_S (SHA1_F2o, E, A, B, C, D, w9_t);
+  wa_t = rotl32_S ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP_S (SHA1_F2o, D, E, A, B, C, wa_t);
+  wb_t = rotl32_S ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP_S (SHA1_F2o, C, D, E, A, B, wb_t);
+  wc_t = rotl32_S ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP_S (SHA1_F2o, B, C, D, E, A, wc_t);
+  wd_t = rotl32_S ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP_S (SHA1_F2o, A, B, C, D, E, wd_t);
+  we_t = rotl32_S ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP_S (SHA1_F2o, E, A, B, C, D, we_t);
+  wf_t = rotl32_S ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP_S (SHA1_F2o, D, E, A, B, C, wf_t);
+  w0_t = rotl32_S ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP_S (SHA1_F2o, C, D, E, A, B, w0_t);
+  w1_t = rotl32_S ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP_S (SHA1_F2o, B, C, D, E, A, w1_t);
+  w2_t = rotl32_S ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP_S (SHA1_F2o, A, B, C, D, E, w2_t);
+  w3_t = rotl32_S ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP_S (SHA1_F2o, E, A, B, C, D, w3_t);
+  w4_t = rotl32_S ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP_S (SHA1_F2o, D, E, A, B, C, w4_t);
+  w5_t = rotl32_S ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP_S (SHA1_F2o, C, D, E, A, B, w5_t);
+  w6_t = rotl32_S ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP_S (SHA1_F2o, B, C, D, E, A, w6_t);
+  w7_t = rotl32_S ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP_S (SHA1_F2o, A, B, C, D, E, w7_t);
+  w8_t = rotl32_S ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP_S (SHA1_F2o, E, A, B, C, D, w8_t);
+  w9_t = rotl32_S ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP_S (SHA1_F2o, D, E, A, B, C, w9_t);
+  wa_t = rotl32_S ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP_S (SHA1_F2o, C, D, E, A, B, wa_t);
+  wb_t = rotl32_S ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP_S (SHA1_F2o, B, C, D, E, A, wb_t);
+
+  #undef K
+  #define K SHA1C03
+
+  wc_t = rotl32_S ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP_S (SHA1_F1, A, B, C, D, E, wc_t);
+  wd_t = rotl32_S ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP_S (SHA1_F1, E, A, B, C, D, wd_t);
+  we_t = rotl32_S ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP_S (SHA1_F1, D, E, A, B, C, we_t);
+  wf_t = rotl32_S ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP_S (SHA1_F1, C, D, E, A, B, wf_t);
+  w0_t = rotl32_S ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP_S (SHA1_F1, B, C, D, E, A, w0_t);
+  w1_t = rotl32_S ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP_S (SHA1_F1, A, B, C, D, E, w1_t);
+  w2_t = rotl32_S ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP_S (SHA1_F1, E, A, B, C, D, w2_t);
+  w3_t = rotl32_S ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP_S (SHA1_F1, D, E, A, B, C, w3_t);
+  w4_t = rotl32_S ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP_S (SHA1_F1, C, D, E, A, B, w4_t);
+  w5_t = rotl32_S ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP_S (SHA1_F1, B, C, D, E, A, w5_t);
+  w6_t = rotl32_S ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP_S (SHA1_F1, A, B, C, D, E, w6_t);
+  w7_t = rotl32_S ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP_S (SHA1_F1, E, A, B, C, D, w7_t);
+  w8_t = rotl32_S ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP_S (SHA1_F1, D, E, A, B, C, w8_t);
+  w9_t = rotl32_S ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP_S (SHA1_F1, C, D, E, A, B, w9_t);
+  wa_t = rotl32_S ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP_S (SHA1_F1, B, C, D, E, A, wa_t);
+  wb_t = rotl32_S ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP_S (SHA1_F1, A, B, C, D, E, wb_t);
+  wc_t = rotl32_S ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP_S (SHA1_F1, E, A, B, C, D, wc_t);
+  wd_t = rotl32_S ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP_S (SHA1_F1, D, E, A, B, C, wd_t);
+  we_t = rotl32_S ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP_S (SHA1_F1, C, D, E, A, B, we_t);
+  wf_t = rotl32_S ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP_S (SHA1_F1, B, C, D, E, A, wf_t);
+
+  digest[0] += A;
+  digest[1] += B;
+  digest[2] += C;
+  digest[3] += D;
+  digest[4] += E;
+}
+
+void hmac_sha1_pad_S (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], u32 ipad[5], u32 opad[5])
+{
+  w0[0] = w0[0] ^ 0x36363636;
+  w0[1] = w0[1] ^ 0x36363636;
+  w0[2] = w0[2] ^ 0x36363636;
+  w0[3] = w0[3] ^ 0x36363636;
+  w1[0] = w1[0] ^ 0x36363636;
+  w1[1] = w1[1] ^ 0x36363636;
+  w1[2] = w1[2] ^ 0x36363636;
+  w1[3] = w1[3] ^ 0x36363636;
+  w2[0] = w2[0] ^ 0x36363636;
+  w2[1] = w2[1] ^ 0x36363636;
+  w2[2] = w2[2] ^ 0x36363636;
+  w2[3] = w2[3] ^ 0x36363636;
+  w3[0] = w3[0] ^ 0x36363636;
+  w3[1] = w3[1] ^ 0x36363636;
+  w3[2] = w3[2] ^ 0x36363636;
+  w3[3] = w3[3] ^ 0x36363636;
+
+  ipad[0] = SHA1M_A;
+  ipad[1] = SHA1M_B;
+  ipad[2] = SHA1M_C;
+  ipad[3] = SHA1M_D;
+  ipad[4] = SHA1M_E;
+
+  sha1_transform_S (w0, w1, w2, w3, ipad);
+
+  w0[0] = w0[0] ^ 0x6a6a6a6a;
+  w0[1] = w0[1] ^ 0x6a6a6a6a;
+  w0[2] = w0[2] ^ 0x6a6a6a6a;
+  w0[3] = w0[3] ^ 0x6a6a6a6a;
+  w1[0] = w1[0] ^ 0x6a6a6a6a;
+  w1[1] = w1[1] ^ 0x6a6a6a6a;
+  w1[2] = w1[2] ^ 0x6a6a6a6a;
+  w1[3] = w1[3] ^ 0x6a6a6a6a;
+  w2[0] = w2[0] ^ 0x6a6a6a6a;
+  w2[1] = w2[1] ^ 0x6a6a6a6a;
+  w2[2] = w2[2] ^ 0x6a6a6a6a;
+  w2[3] = w2[3] ^ 0x6a6a6a6a;
+  w3[0] = w3[0] ^ 0x6a6a6a6a;
+  w3[1] = w3[1] ^ 0x6a6a6a6a;
+  w3[2] = w3[2] ^ 0x6a6a6a6a;
+  w3[3] = w3[3] ^ 0x6a6a6a6a;
+
+  opad[0] = SHA1M_A;
+  opad[1] = SHA1M_B;
+  opad[2] = SHA1M_C;
+  opad[3] = SHA1M_D;
+  opad[4] = SHA1M_E;
+
+  sha1_transform_S (w0, w1, w2, w3, opad);
+}
+
+void hmac_sha1_run_S (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], u32 ipad[5], u32 opad[5], u32 digest[5])
+{
+  digest[0] = ipad[0];
+  digest[1] = ipad[1];
+  digest[2] = ipad[2];
+  digest[3] = ipad[3];
+  digest[4] = ipad[4];
+
+  sha1_transform_S (w0, w1, w2, w3, digest);
+
+  w0[0] = digest[0];
+  w0[1] = digest[1];
+  w0[2] = digest[2];
+  w0[3] = digest[3];
+  w1[0] = digest[4];
+  w1[1] = 0x80000000;
+  w1[2] = 0;
+  w1[3] = 0;
+  w2[0] = 0;
+  w2[1] = 0;
+  w2[2] = 0;
+  w2[3] = 0;
+  w3[0] = 0;
+  w3[1] = 0;
+  w3[2] = 0;
+  w3[3] = (64 + 20) * 8;
+
+  digest[0] = opad[0];
+  digest[1] = opad[1];
+  digest[2] = opad[2];
+  digest[3] = opad[3];
+  digest[4] = opad[4];
+
+  sha1_transform_S (w0, w1, w2, w3, digest);
+}
+
+void sha1_transform_V (const u32x w0[4], const u32x w1[4], const u32x w2[4], const u32x w3[4], u32x digest[5])
+{
+  u32x A = digest[0];
+  u32x B = digest[1];
+  u32x C = digest[2];
+  u32x D = digest[3];
+  u32x E = digest[4];
+
+  u32x w0_t = w0[0];
+  u32x w1_t = w0[1];
+  u32x w2_t = w0[2];
+  u32x w3_t = w0[3];
+  u32x w4_t = w1[0];
+  u32x w5_t = w1[1];
+  u32x w6_t = w1[2];
+  u32x w7_t = w1[3];
+  u32x w8_t = w2[0];
+  u32x w9_t = w2[1];
+  u32x wa_t = w2[2];
+  u32x wb_t = w2[3];
+  u32x wc_t = w3[0];
+  u32x wd_t = w3[1];
+  u32x we_t = w3[2];
+  u32x wf_t = w3[3];
+
+  #undef K
+  #define K SHA1C00
+
   SHA1_STEP (SHA1_F0o, A, B, C, D, E, w0_t);
   SHA1_STEP (SHA1_F0o, E, A, B, C, D, w1_t);
   SHA1_STEP (SHA1_F0o, D, E, A, B, C, w2_t);
@@ -149,7 +368,7 @@ void sha1_transform (const u32 w0[4], const u32 w1[4], const u32 w2[4], const u3
   digest[4] += E;
 }
 
-void hmac_sha1_pad (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], u32 ipad[5], u32 opad[5])
+void hmac_sha1_pad_V (u32x w0[4], u32x w1[4], u32x w2[4], u32x w3[4], u32x ipad[5], u32x opad[5])
 {
   w0[0] = w0[0] ^ 0x36363636;
   w0[1] = w0[1] ^ 0x36363636;
@@ -174,7 +393,7 @@ void hmac_sha1_pad (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], u32 ipad[5], u32
   ipad[3] = SHA1M_D;
   ipad[4] = SHA1M_E;
 
-  sha1_transform (w0, w1, w2, w3, ipad);
+  sha1_transform_V (w0, w1, w2, w3, ipad);
 
   w0[0] = w0[0] ^ 0x6a6a6a6a;
   w0[1] = w0[1] ^ 0x6a6a6a6a;
@@ -199,10 +418,10 @@ void hmac_sha1_pad (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], u32 ipad[5], u32
   opad[3] = SHA1M_D;
   opad[4] = SHA1M_E;
 
-  sha1_transform (w0, w1, w2, w3, opad);
+  sha1_transform_V (w0, w1, w2, w3, opad);
 }
 
-void hmac_sha1_run (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], u32 ipad[5], u32 opad[5], u32 digest[5])
+void hmac_sha1_run_V (u32x w0[4], u32x w1[4], u32x w2[4], u32x w3[4], u32x ipad[5], u32x opad[5], u32x digest[5])
 {
   digest[0] = ipad[0];
   digest[1] = ipad[1];
@@ -210,7 +429,7 @@ void hmac_sha1_run (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], u32 ipad[5], u32
   digest[3] = ipad[3];
   digest[4] = ipad[4];
 
-  sha1_transform (w0, w1, w2, w3, digest);
+  sha1_transform_V (w0, w1, w2, w3, digest);
 
   w0[0] = digest[0];
   w0[1] = digest[1];
@@ -235,7 +454,7 @@ void hmac_sha1_run (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], u32 ipad[5], u32
   digest[3] = opad[3];
   digest[4] = opad[4];
 
-  sha1_transform (w0, w1, w2, w3, digest);
+  sha1_transform_V (w0, w1, w2, w3, digest);
 }
 
 __kernel void m12000_init (__global pw_t *pws, __global kernel_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global pbkdf2_sha1_tmp_t *tmps, __global void *hooks, __global u32 *bitmaps_buf_s1_a, __global u32 *bitmaps_buf_s1_b, __global u32 *bitmaps_buf_s1_c, __global u32 *bitmaps_buf_s1_d, __global u32 *bitmaps_buf_s2_a, __global u32 *bitmaps_buf_s2_b, __global u32 *bitmaps_buf_s2_c, __global u32 *bitmaps_buf_s2_d, __global plain_t *plains_buf, __global digest_t *digests_buf, __global u32 *hashes_shown, __global salt_t *salt_bufs, __global pbkdf2_sha1_t *esalt_bufs, __global u32 *d_return_buf, __global u32 *d_scryptV_buf, const u32 bitmap_mask, const u32 bitmap_shift1, const u32 bitmap_shift2, const u32 salt_pos, const u32 loop_pos, const u32 loop_cnt, const u32 il_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
@@ -250,31 +469,31 @@ __kernel void m12000_init (__global pw_t *pws, __global kernel_rule_t *rules_buf
 
   u32 w0[4];
 
-  w0[0] = swap32 (pws[gid].i[ 0]);
-  w0[1] = swap32 (pws[gid].i[ 1]);
-  w0[2] = swap32 (pws[gid].i[ 2]);
-  w0[3] = swap32 (pws[gid].i[ 3]);
+  w0[0] = swap32_S (pws[gid].i[ 0]);
+  w0[1] = swap32_S (pws[gid].i[ 1]);
+  w0[2] = swap32_S (pws[gid].i[ 2]);
+  w0[3] = swap32_S (pws[gid].i[ 3]);
 
   u32 w1[4];
 
-  w1[0] = swap32 (pws[gid].i[ 4]);
-  w1[1] = swap32 (pws[gid].i[ 5]);
-  w1[2] = swap32 (pws[gid].i[ 6]);
-  w1[3] = swap32 (pws[gid].i[ 7]);
+  w1[0] = swap32_S (pws[gid].i[ 4]);
+  w1[1] = swap32_S (pws[gid].i[ 5]);
+  w1[2] = swap32_S (pws[gid].i[ 6]);
+  w1[3] = swap32_S (pws[gid].i[ 7]);
 
   u32 w2[4];
 
-  w2[0] = swap32 (pws[gid].i[ 8]);
-  w2[1] = swap32 (pws[gid].i[ 9]);
-  w2[2] = swap32 (pws[gid].i[10]);
-  w2[3] = swap32 (pws[gid].i[11]);
+  w2[0] = swap32_S (pws[gid].i[ 8]);
+  w2[1] = swap32_S (pws[gid].i[ 9]);
+  w2[2] = swap32_S (pws[gid].i[10]);
+  w2[3] = swap32_S (pws[gid].i[11]);
 
   u32 w3[4];
 
-  w3[0] = swap32 (pws[gid].i[12]);
-  w3[1] = swap32 (pws[gid].i[13]);
-  w3[2] = swap32 (pws[gid].i[14]);
-  w3[3] = swap32 (pws[gid].i[15]);
+  w3[0] = swap32_S (pws[gid].i[12]);
+  w3[1] = swap32_S (pws[gid].i[13]);
+  w3[2] = swap32_S (pws[gid].i[14]);
+  w3[3] = swap32_S (pws[gid].i[15]);
 
   /**
    * salt
@@ -287,27 +506,27 @@ __kernel void m12000_init (__global pw_t *pws, __global kernel_rule_t *rules_buf
   u32 esalt_buf2[4];
   u32 esalt_buf3[4];
 
-  esalt_buf0[0] = swap32 (esalt_bufs[salt_pos].salt_buf[ 0]);
-  esalt_buf0[1] = swap32 (esalt_bufs[salt_pos].salt_buf[ 1]);
-  esalt_buf0[2] = swap32 (esalt_bufs[salt_pos].salt_buf[ 2]);
-  esalt_buf0[3] = swap32 (esalt_bufs[salt_pos].salt_buf[ 3]);
-  esalt_buf1[0] = swap32 (esalt_bufs[salt_pos].salt_buf[ 4]);
-  esalt_buf1[1] = swap32 (esalt_bufs[salt_pos].salt_buf[ 5]);
-  esalt_buf1[2] = swap32 (esalt_bufs[salt_pos].salt_buf[ 6]);
-  esalt_buf1[3] = swap32 (esalt_bufs[salt_pos].salt_buf[ 7]);
-  esalt_buf2[0] = swap32 (esalt_bufs[salt_pos].salt_buf[ 8]);
-  esalt_buf2[1] = swap32 (esalt_bufs[salt_pos].salt_buf[ 9]);
-  esalt_buf2[2] = swap32 (esalt_bufs[salt_pos].salt_buf[10]);
-  esalt_buf2[3] = swap32 (esalt_bufs[salt_pos].salt_buf[11]);
-  esalt_buf3[0] = swap32 (esalt_bufs[salt_pos].salt_buf[12]);
-  esalt_buf3[1] = swap32 (esalt_bufs[salt_pos].salt_buf[13]);
+  esalt_buf0[0] = swap32_S (esalt_bufs[salt_pos].salt_buf[ 0]);
+  esalt_buf0[1] = swap32_S (esalt_bufs[salt_pos].salt_buf[ 1]);
+  esalt_buf0[2] = swap32_S (esalt_bufs[salt_pos].salt_buf[ 2]);
+  esalt_buf0[3] = swap32_S (esalt_bufs[salt_pos].salt_buf[ 3]);
+  esalt_buf1[0] = swap32_S (esalt_bufs[salt_pos].salt_buf[ 4]);
+  esalt_buf1[1] = swap32_S (esalt_bufs[salt_pos].salt_buf[ 5]);
+  esalt_buf1[2] = swap32_S (esalt_bufs[salt_pos].salt_buf[ 6]);
+  esalt_buf1[3] = swap32_S (esalt_bufs[salt_pos].salt_buf[ 7]);
+  esalt_buf2[0] = swap32_S (esalt_bufs[salt_pos].salt_buf[ 8]);
+  esalt_buf2[1] = swap32_S (esalt_bufs[salt_pos].salt_buf[ 9]);
+  esalt_buf2[2] = swap32_S (esalt_bufs[salt_pos].salt_buf[10]);
+  esalt_buf2[3] = swap32_S (esalt_bufs[salt_pos].salt_buf[11]);
+  esalt_buf3[0] = swap32_S (esalt_bufs[salt_pos].salt_buf[12]);
+  esalt_buf3[1] = swap32_S (esalt_bufs[salt_pos].salt_buf[13]);
   esalt_buf3[2] = 0;
   esalt_buf3[3] = (64 + salt_len + 4) * 8;
 
   u32 ipad[5];
   u32 opad[5];
 
-  hmac_sha1_pad (w0, w1, w2, w3, ipad, opad);
+  hmac_sha1_pad_S (w0, w1, w2, w3, ipad, opad);
 
   tmps[gid].ipad[0] = ipad[0];
   tmps[gid].ipad[1] = ipad[1];
@@ -325,7 +544,7 @@ __kernel void m12000_init (__global pw_t *pws, __global kernel_rule_t *rules_buf
   {
     u32 dgst[5];
 
-    hmac_sha1_run (esalt_buf0, esalt_buf1, esalt_buf2, esalt_buf3, ipad, opad, dgst);
+    hmac_sha1_run_S (esalt_buf0, esalt_buf1, esalt_buf2, esalt_buf3, ipad, opad, dgst);
 
     tmps[gid].dgst[i + 0] = dgst[0];
     tmps[gid].dgst[i + 1] = dgst[1];
@@ -347,46 +566,44 @@ __kernel void m12000_loop (__global pw_t *pws, __global kernel_rule_t *rules_buf
 
   if (gid >= gid_max) return;
 
-  u32 ipad[5];
-
-  ipad[0] = tmps[gid].ipad[0];
-  ipad[1] = tmps[gid].ipad[1];
-  ipad[2] = tmps[gid].ipad[2];
-  ipad[3] = tmps[gid].ipad[3];
-  ipad[4] = tmps[gid].ipad[4];
+  u32x ipad[5];
+  u32x opad[5];
 
-  u32 opad[5];
+  ipad[0] = packv (tmps, ipad, gid, 0);
+  ipad[1] = packv (tmps, ipad, gid, 1);
+  ipad[2] = packv (tmps, ipad, gid, 2);
+  ipad[3] = packv (tmps, ipad, gid, 3);
+  ipad[4] = packv (tmps, ipad, gid, 4);
 
-  opad[0] = tmps[gid].opad[0];
-  opad[1] = tmps[gid].opad[1];
-  opad[2] = tmps[gid].opad[2];
-  opad[3] = tmps[gid].opad[3];
-  opad[4] = tmps[gid].opad[4];
+  opad[0] = packv (tmps, opad, gid, 0);
+  opad[1] = packv (tmps, opad, gid, 1);
+  opad[2] = packv (tmps, opad, gid, 2);
+  opad[3] = packv (tmps, opad, gid, 3);
+  opad[4] = packv (tmps, opad, gid, 4);
 
   for (u32 i = 0; i < 5; i += 5)
   {
-    u32 dgst[5];
+    u32x dgst[5];
+    u32x out[5];
 
-    dgst[0] = tmps[gid].dgst[i + 0];
-    dgst[1] = tmps[gid].dgst[i + 1];
-    dgst[2] = tmps[gid].dgst[i + 2];
-    dgst[3] = tmps[gid].dgst[i + 3];
-    dgst[4] = tmps[gid].dgst[i + 4];
+    dgst[0] = packv (tmps, dgst, gid, 0);
+    dgst[1] = packv (tmps, dgst, gid, 1);
+    dgst[2] = packv (tmps, dgst, gid, 2);
+    dgst[3] = packv (tmps, dgst, gid, 3);
+    dgst[4] = packv (tmps, dgst, gid, 4);
 
-    u32 out[5];
-
-    out[0] = tmps[gid].out[i + 0];
-    out[1] = tmps[gid].out[i + 1];
-    out[2] = tmps[gid].out[i + 2];
-    out[3] = tmps[gid].out[i + 3];
-    out[4] = tmps[gid].out[i + 4];
+    out[0] = packv (tmps, out, gid, 0);
+    out[1] = packv (tmps, out, gid, 1);
+    out[2] = packv (tmps, out, gid, 2);
+    out[3] = packv (tmps, out, gid, 3);
+    out[4] = packv (tmps, out, gid, 4);
 
     for (u32 j = 0; j < loop_cnt; j++)
     {
-      u32 w0[4];
-      u32 w1[4];
-      u32 w2[4];
-      u32 w3[4];
+      u32x w0[4];
+      u32x w1[4];
+      u32x w2[4];
+      u32x w3[4];
 
       w0[0] = dgst[0];
       w0[1] = dgst[1];
@@ -405,7 +622,7 @@ __kernel void m12000_loop (__global pw_t *pws, __global kernel_rule_t *rules_buf
       w3[2] = 0;
       w3[3] = (64 + 20) * 8;
 
-      hmac_sha1_run (w0, w1, w2, w3, ipad, opad, dgst);
+      hmac_sha1_run_V (w0, w1, w2, w3, ipad, opad, dgst);
 
       out[0] ^= dgst[0];
       out[1] ^= dgst[1];
@@ -414,17 +631,17 @@ __kernel void m12000_loop (__global pw_t *pws, __global kernel_rule_t *rules_buf
       out[4] ^= dgst[4];
     }
 
-    tmps[gid].dgst[i + 0] = dgst[0];
-    tmps[gid].dgst[i + 1] = dgst[1];
-    tmps[gid].dgst[i + 2] = dgst[2];
-    tmps[gid].dgst[i + 3] = dgst[3];
-    tmps[gid].dgst[i + 4] = dgst[4];
-
-    tmps[gid].out[i + 0] = out[0];
-    tmps[gid].out[i + 1] = out[1];
-    tmps[gid].out[i + 2] = out[2];
-    tmps[gid].out[i + 3] = out[3];
-    tmps[gid].out[i + 4] = out[4];
+    unpackv (tmps, dgst, gid, 0, dgst[0]);
+    unpackv (tmps, dgst, gid, 1, dgst[1]);
+    unpackv (tmps, dgst, gid, 2, dgst[2]);
+    unpackv (tmps, dgst, gid, 3, dgst[3]);
+    unpackv (tmps, dgst, gid, 4, dgst[4]);
+
+    unpackv (tmps, out, gid, 0, out[0]);
+    unpackv (tmps, out, gid, 1, out[1]);
+    unpackv (tmps, out, gid, 2, out[2]);
+    unpackv (tmps, out, gid, 3, out[3]);
+    unpackv (tmps, out, gid, 4, out[4]);
   }
 }
 
diff --git a/OpenCL/simd.c b/OpenCL/simd.c
index 865819dc0..9e87b2c6a 100644
--- a/OpenCL/simd.c
+++ b/OpenCL/simd.c
@@ -1251,6 +1251,30 @@ inline u32x ix_create_combt (__global comb_t *combs_buf, const u32 il_pos, const
 #define packvf(arr,var,gid) (u32x) ((arr)[((gid) * 16) + 0].var, (arr)[((gid) * 16) + 1].var, (arr)[((gid) * 16) + 2].var, (arr)[((gid) * 16) + 3].var, (arr)[((gid) * 16) + 4].var, (arr)[((gid) * 16) + 5].var, (arr)[((gid) * 16) + 6].var, (arr)[((gid) * 16) + 7].var, (arr)[((gid) * 16) + 8].var, (arr)[((gid) * 16) + 9].var, (arr)[((gid) * 16) + 10].var, (arr)[((gid) * 16) + 11].var, (arr)[((gid) * 16) + 12].var, (arr)[((gid) * 16) + 13].var, (arr)[((gid) * 16) + 14].var, (arr)[((gid) * 16) + 15].var)
 #endif
 
+#if   VECT_SIZE == 1
+#define pack64v(arr,var,gid,idx) (u64x) ((arr)[((gid) *  1) + 0].var[(idx)])
+#elif VECT_SIZE == 2
+#define pack64v(arr,var,gid,idx) (u64x) ((arr)[((gid) *  2) + 0].var[(idx)], (arr)[((gid) *  2) + 1].var[(idx)])
+#elif VECT_SIZE == 4
+#define pack64v(arr,var,gid,idx) (u64x) ((arr)[((gid) *  4) + 0].var[(idx)], (arr)[((gid) *  4) + 1].var[(idx)], (arr)[((gid) *  4) + 2].var[(idx)], (arr)[((gid) *  4) + 3].var[(idx)])
+#elif VECT_SIZE == 8
+#define pack64v(arr,var,gid,idx) (u64x) ((arr)[((gid) *  8) + 0].var[(idx)], (arr)[((gid) *  8) + 1].var[(idx)], (arr)[((gid) *  8) + 2].var[(idx)], (arr)[((gid) *  8) + 3].var[(idx)], (arr)[((gid) *  8) + 4].var[(idx)], (arr)[((gid) *  8) + 5].var[(idx)], (arr)[((gid) *  8) + 6].var[(idx)], (arr)[((gid) *  8) + 7].var[(idx)])
+#elif VECT_SIZE == 16
+#define pack64v(arr,var,gid,idx) (u64x) ((arr)[((gid) * 16) + 0].var[(idx)], (arr)[((gid) * 16) + 1].var[(idx)], (arr)[((gid) * 16) + 2].var[(idx)], (arr)[((gid) * 16) + 3].var[(idx)], (arr)[((gid) * 16) + 4].var[(idx)], (arr)[((gid) * 16) + 5].var[(idx)], (arr)[((gid) * 16) + 6].var[(idx)], (arr)[((gid) * 16) + 7].var[(idx)], (arr)[((gid) * 16) + 8].var[(idx)], (arr)[((gid) * 16) + 9].var[(idx)], (arr)[((gid) * 16) + 10].var[(idx)], (arr)[((gid) * 16) + 11].var[(idx)], (arr)[((gid) * 16) + 12].var[(idx)], (arr)[((gid) * 16) + 13].var[(idx)], (arr)[((gid) * 16) + 14].var[(idx)], (arr)[((gid) * 16) + 15].var[(idx)])
+#endif
+
+#if   VECT_SIZE == 1
+#define pack64vf(arr,var,gid) (u64x) ((arr)[((gid) *  1) + 0].var)
+#elif VECT_SIZE == 2
+#define pack64vf(arr,var,gid) (u64x) ((arr)[((gid) *  2) + 0].var, (arr)[((gid) *  2) + 1].var)
+#elif VECT_SIZE == 4
+#define pack64vf(arr,var,gid) (u64x) ((arr)[((gid) *  4) + 0].var, (arr)[((gid) *  4) + 1].var, (arr)[((gid) *  4) + 2].var, (arr)[((gid) *  4) + 3].var)
+#elif VECT_SIZE == 8
+#define pack64vf(arr,var,gid) (u64x) ((arr)[((gid) *  8) + 0].var, (arr)[((gid) *  8) + 1].var, (arr)[((gid) *  8) + 2].var, (arr)[((gid) *  8) + 3].var, (arr)[((gid) *  8) + 4].var, (arr)[((gid) *  8) + 5].var, (arr)[((gid) *  8) + 6].var, (arr)[((gid) *  8) + 7].var)
+#elif VECT_SIZE == 16
+#define pack64vf(arr,var,gid) (u64x) ((arr)[((gid) * 16) + 0].var, (arr)[((gid) * 16) + 1].var, (arr)[((gid) * 16) + 2].var, (arr)[((gid) * 16) + 3].var, (arr)[((gid) * 16) + 4].var, (arr)[((gid) * 16) + 5].var, (arr)[((gid) * 16) + 6].var, (arr)[((gid) * 16) + 7].var, (arr)[((gid) * 16) + 8].var, (arr)[((gid) * 16) + 9].var, (arr)[((gid) * 16) + 10].var, (arr)[((gid) * 16) + 11].var, (arr)[((gid) * 16) + 12].var, (arr)[((gid) * 16) + 13].var, (arr)[((gid) * 16) + 14].var, (arr)[((gid) * 16) + 15].var)
+#endif
+
 #if   VECT_SIZE == 1
 #define unpackv(arr,var,gid,idx,val) (arr)[((gid) *  1) + 0].var[(idx)] = val;
 #elif VECT_SIZE == 2
diff --git a/include/kernel_functions.c b/include/kernel_functions.c
index 0c70f5eff..0ccbb2534 100644
--- a/include/kernel_functions.c
+++ b/include/kernel_functions.c
@@ -197,6 +197,11 @@
 
 #define SHIFT_RIGHT_32(x,n) ((x) >> (n))
 
+#define SHA256_S0_S(x) (rotl32_S ((x), 25u) ^ rotl32_S ((x), 14u) ^ SHIFT_RIGHT_32 ((x),  3u))
+#define SHA256_S1_S(x) (rotl32_S ((x), 15u) ^ rotl32_S ((x), 13u) ^ SHIFT_RIGHT_32 ((x), 10u))
+#define SHA256_S2_S(x) (rotl32_S ((x), 30u) ^ rotl32_S ((x), 19u) ^ rotl32_S ((x), 10u))
+#define SHA256_S3_S(x) (rotl32_S ((x), 26u) ^ rotl32_S ((x), 21u) ^ rotl32_S ((x),  7u))
+
 #define SHA256_S0(x) (rotl32 ((x), 25u) ^ rotl32 ((x), 14u) ^ SHIFT_RIGHT_32 ((x),  3u))
 #define SHA256_S1(x) (rotl32 ((x), 15u) ^ rotl32 ((x), 13u) ^ SHIFT_RIGHT_32 ((x), 10u))
 #define SHA256_S2(x) (rotl32 ((x), 30u) ^ rotl32 ((x), 19u) ^ rotl32 ((x), 10u))
@@ -223,15 +228,28 @@
 #define SHA256_F1o(x,y,z) (SHA256_F1 ((x), (y), (z)))
 #endif
 
-#define SHA256_STEP(F0,F1,a,b,c,d,e,f,g,h,x,K)  \
-{                                               \
-  h += K;                                       \
-  h += x;                                       \
-  h += SHA256_S3 (e);                           \
-  h += F1 (e,f,g);                              \
-  d += h;                                       \
-  h += SHA256_S2 (a);                           \
-  h += F0 (a,b,c);                              \
+#define SHA256_STEP_S(F0,F1,a,b,c,d,e,f,g,h,x,K)  \
+{                                                 \
+  h += K;                                         \
+  h += x;                                         \
+  h += SHA256_S3_S (e);                           \
+  h += F1 (e,f,g);                                \
+  d += h;                                         \
+  h += SHA256_S2_S (a);                           \
+  h += F0 (a,b,c);                                \
+}
+
+#define SHA256_EXPAND_S(x,y,z,w) (SHA256_S1_S (x) + y + SHA256_S0_S (z) + w)
+
+#define SHA256_STEP(F0,F1,a,b,c,d,e,f,g,h,x,K)    \
+{                                                 \
+  h += K;                                         \
+  h += x;                                         \
+  h += SHA256_S3 (e);                             \
+  h += F1 (e,f,g);                                \
+  d += h;                                         \
+  h += SHA256_S2 (a);                             \
+  h += F0 (a,b,c);                                \
 }
 
 #define SHA256_EXPAND(x,y,z,w) (SHA256_S1 (x) + y + SHA256_S0 (z) + w)
@@ -283,6 +301,11 @@
 
 #define SHIFT_RIGHT_64(x,n) ((x) >> (n))
 
+#define SHA512_S0_S(x) (rotr64_S ((x), 28) ^ rotr64_S ((x), 34) ^ rotr64_S ((x), 39))
+#define SHA512_S1_S(x) (rotr64_S ((x), 14) ^ rotr64_S ((x), 18) ^ rotr64_S ((x), 41))
+#define SHA512_S2_S(x) (rotr64_S ((x),  1) ^ rotr64_S ((x),  8) ^ SHIFT_RIGHT_64 ((x), 7))
+#define SHA512_S3_S(x) (rotr64_S ((x), 19) ^ rotr64_S ((x), 61) ^ SHIFT_RIGHT_64 ((x), 6))
+
 #define SHA512_S0(x) (rotr64 ((x), 28) ^ rotr64 ((x), 34) ^ rotr64 ((x), 39))
 #define SHA512_S1(x) (rotr64 ((x), 14) ^ rotr64 ((x), 18) ^ rotr64 ((x), 41))
 #define SHA512_S2(x) (rotr64 ((x),  1) ^ rotr64 ((x),  8) ^ SHIFT_RIGHT_64 ((x), 7))
@@ -306,6 +329,19 @@
 #define SHA512_F1o(x,y,z) (SHA512_F1 ((x), (y), (z)))
 #endif
 
+#define SHA512_STEP_S(F0,F1,a,b,c,d,e,f,g,h,x,K)  \
+{                                                 \
+  h += K;                                         \
+  h += x;                                         \
+  h += SHA512_S1_S (e);                           \
+  h += F0 (e, f, g);                              \
+  d += h;                                         \
+  h += SHA512_S0_S (a);                           \
+  h += F1 (a, b, c);                              \
+}
+
+#define SHA512_EXPAND_S(x,y,z,w) (SHA512_S3_S (x) + y + SHA512_S2_S (z) + w)
+
 #define SHA512_STEP(F0,F1,a,b,c,d,e,f,g,h,x,K)  \
 {                                               \
   h += K;                                       \
diff --git a/src/hashcat.c b/src/hashcat.c
index 220dff0be..1e7393e99 100644
--- a/src/hashcat.c
+++ b/src/hashcat.c
@@ -8992,7 +8992,8 @@ int main (int argc, char **argv)
                    parse_func  = sha512osx_parse_hash;
                    sort_by_digest = sort_by_digest_8_16;
                    opti_type   = OPTI_TYPE_ZERO_BYTE
-                               | OPTI_TYPE_USES_BITS_64;
+                               | OPTI_TYPE_USES_BITS_64
+                               | OPTI_TYPE_SLOW_HASH_SIMD;
                    dgst_pos0   = 0;
                    dgst_pos1   = 1;
                    dgst_pos2   = 2;
@@ -9008,7 +9009,8 @@ int main (int argc, char **argv)
                    parse_func  = sha512grub_parse_hash;
                    sort_by_digest = sort_by_digest_8_16;
                    opti_type   = OPTI_TYPE_ZERO_BYTE
-                               | OPTI_TYPE_USES_BITS_64;
+                               | OPTI_TYPE_USES_BITS_64
+                               | OPTI_TYPE_SLOW_HASH_SIMD;
                    dgst_pos0   = 0;
                    dgst_pos1   = 1;
                    dgst_pos2   = 2;
@@ -9354,7 +9356,8 @@ int main (int argc, char **argv)
                    dgst_size   = DGST_SIZE_4_32;
                    parse_func  = cisco8_parse_hash;
                    sort_by_digest = sort_by_digest_4_32;
-                   opti_type   = OPTI_TYPE_ZERO_BYTE;
+                   opti_type   = OPTI_TYPE_ZERO_BYTE
+                               | OPTI_TYPE_SLOW_HASH_SIMD;
                    dgst_pos0   = 0;
                    dgst_pos1   = 1;
                    dgst_pos2   = 2;
@@ -9561,7 +9564,8 @@ int main (int argc, char **argv)
                    dgst_size   = DGST_SIZE_4_32;
                    parse_func  = djangopbkdf2_parse_hash;
                    sort_by_digest = sort_by_digest_4_32;
-                   opti_type   = OPTI_TYPE_ZERO_BYTE;
+                   opti_type   = OPTI_TYPE_ZERO_BYTE
+                               | OPTI_TYPE_SLOW_HASH_SIMD;
                    dgst_pos0   = 0;
                    dgst_pos1   = 1;
                    dgst_pos2   = 2;
@@ -9757,7 +9761,8 @@ int main (int argc, char **argv)
                    dgst_size   = DGST_SIZE_4_32;
                    parse_func  = pbkdf2_sha256_parse_hash;
                    sort_by_digest = sort_by_digest_4_32;
-                   opti_type   = OPTI_TYPE_ZERO_BYTE;
+                   opti_type   = OPTI_TYPE_ZERO_BYTE
+                               | OPTI_TYPE_SLOW_HASH_SIMD;
                    dgst_pos0   = 0;
                    dgst_pos1   = 1;
                    dgst_pos2   = 2;
@@ -9929,7 +9934,8 @@ int main (int argc, char **argv)
                    dgst_size   = DGST_SIZE_4_32;
                    parse_func  = pbkdf2_md5_parse_hash;
                    sort_by_digest = sort_by_digest_4_32;
-                   opti_type   = OPTI_TYPE_ZERO_BYTE;
+                   opti_type   = OPTI_TYPE_ZERO_BYTE
+                               | OPTI_TYPE_SLOW_HASH_SIMD;
                    dgst_pos0   = 0;
                    dgst_pos1   = 1;
                    dgst_pos2   = 2;
@@ -9946,7 +9952,8 @@ int main (int argc, char **argv)
                    dgst_size   = DGST_SIZE_4_32;
                    parse_func  = pbkdf2_sha1_parse_hash;
                    sort_by_digest = sort_by_digest_4_32;
-                   opti_type   = OPTI_TYPE_ZERO_BYTE;
+                   opti_type   = OPTI_TYPE_ZERO_BYTE
+                               | OPTI_TYPE_SLOW_HASH_SIMD;
                    dgst_pos0   = 0;
                    dgst_pos1   = 1;
                    dgst_pos2   = 2;
@@ -9964,7 +9971,8 @@ int main (int argc, char **argv)
                    parse_func  = pbkdf2_sha512_parse_hash;
                    sort_by_digest = sort_by_digest_8_16;
                    opti_type   = OPTI_TYPE_ZERO_BYTE
-                               | OPTI_TYPE_USES_BITS_64;
+                               | OPTI_TYPE_USES_BITS_64
+                               | OPTI_TYPE_SLOW_HASH_SIMD;
                    dgst_pos0   = 0;
                    dgst_pos1   = 1;
                    dgst_pos2   = 2;