X-Git-Url: https://www.flypig.org.uk/git/?a=blobdiff_plain;f=OpenCL%2Fm08400_a1.cl;h=af903d96cfc7cfb4e753779bab5f44510c66a824;hb=161a6eb4bc643d8e636e96eda613f5137d30da59;hp=c5bef54d047abc551805d3e5d3359860e9cc3f10;hpb=6893a9ad857a1d4fe63f1a92365802f5d9d3b39e;p=hashcat.git

diff --git a/OpenCL/m08400_a1.cl b/OpenCL/m08400_a1.cl
index c5bef54..af903d9 100644
--- a/OpenCL/m08400_a1.cl
+++ b/OpenCL/m08400_a1.cl
@@ -5,47 +5,51 @@
 
 #define _SHA1_
 
-#include "include/constants.h"
-#include "include/kernel_vendor.h"
-
-#define DGST_R0 3
-#define DGST_R1 4
-#define DGST_R2 2
-#define DGST_R3 1
-
-#include "include/kernel_functions.c"
-#include "types_ocl.c"
-#include "common.c"
-
-#define COMPARE_S "check_single_comp4.c"
-#define COMPARE_M "check_multi_comp4.c"
-
-#define uint_to_hex_lower8_le(i) l_bin2asc[(i)]
-
-static void sha1_transform (const u32 w0[4], const u32 w1[4], const u32 w2[4], const u32 w3[4], u32 digest[5])
+#define NEW_SIMD_CODE
+
+#include "inc_vendor.cl"
+#include "inc_hash_constants.h"
+#include "inc_hash_functions.cl"
+#include "inc_types.cl"
+#include "inc_common.cl"
+#include "inc_simd.cl"
+
+#if   VECT_SIZE == 1
+#define uint_to_hex_lower8_le(i) (u32x) (l_bin2asc[(i)])
+#elif VECT_SIZE == 2
+#define uint_to_hex_lower8_le(i) (u32x) (l_bin2asc[(i).s0], l_bin2asc[(i).s1])
+#elif VECT_SIZE == 4
+#define uint_to_hex_lower8_le(i) (u32x) (l_bin2asc[(i).s0], l_bin2asc[(i).s1], l_bin2asc[(i).s2], l_bin2asc[(i).s3])
+#elif VECT_SIZE == 8
+#define uint_to_hex_lower8_le(i) (u32x) (l_bin2asc[(i).s0], l_bin2asc[(i).s1], l_bin2asc[(i).s2], l_bin2asc[(i).s3], l_bin2asc[(i).s4], l_bin2asc[(i).s5], l_bin2asc[(i).s6], l_bin2asc[(i).s7])
+#elif VECT_SIZE == 16
+#define uint_to_hex_lower8_le(i) (u32x) (l_bin2asc[(i).s0], l_bin2asc[(i).s1], l_bin2asc[(i).s2], l_bin2asc[(i).s3], l_bin2asc[(i).s4], l_bin2asc[(i).s5], l_bin2asc[(i).s6], l_bin2asc[(i).s7], l_bin2asc[(i).s8], l_bin2asc[(i).s9], l_bin2asc[(i).sa], l_bin2asc[(i).sb], l_bin2asc[(i).sc], l_bin2asc[(i).sd], l_bin2asc[(i).se], l_bin2asc[(i).sf])
+#endif
+
+void sha1_transform (const u32x w0[4], const u32x w1[4], const u32x w2[4], const u32x w3[4], u32x digest[5])
 {
-  u32 A = digest[0];
-  u32 B = digest[1];
-  u32 C = digest[2];
-  u32 D = digest[3];
-  u32 E = digest[4];
-
-  u32 w0_t = w0[0];
-  u32 w1_t = w0[1];
-  u32 w2_t = w0[2];
-  u32 w3_t = w0[3];
-  u32 w4_t = w1[0];
-  u32 w5_t = w1[1];
-  u32 w6_t = w1[2];
-  u32 w7_t = w1[3];
-  u32 w8_t = w2[0];
-  u32 w9_t = w2[1];
-  u32 wa_t = w2[2];
-  u32 wb_t = w2[3];
-  u32 wc_t = w3[0];
-  u32 wd_t = w3[1];
-  u32 we_t = w3[2];
-  u32 wf_t = w3[3];
+  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
@@ -150,114 +154,71 @@ static void sha1_transform (const u32 w0[4], const u32 w1[4], const u32 w2[4], c
   digest[4] += E;
 }
 
-__kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m08400_m04 (__global pw_t *pws, __global gpu_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global void *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 void *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 combs_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
+__kernel void m08400_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global void *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 void *esalt_bufs, __global u32 *d_return_buf, __global u32 *d_scryptV0_buf, __global u32 *d_scryptV1_buf, __global u32 *d_scryptV2_buf, __global u32 *d_scryptV3_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)
 {
   /**
    * modifier
    */
 
+  const u32 gid = get_global_id (0);
   const u32 lid = get_local_id (0);
+  const u32 lsz = get_local_size (0);
 
   /**
-   * base
+   * shared
    */
 
-  const u32 gid = get_global_id (0);
-
-  u32 wordl0[4];
-
-  wordl0[0] = pws[gid].i[ 0];
-  wordl0[1] = pws[gid].i[ 1];
-  wordl0[2] = pws[gid].i[ 2];
-  wordl0[3] = pws[gid].i[ 3];
-
-  u32 wordl1[4];
-
-  wordl1[0] = pws[gid].i[ 4];
-  wordl1[1] = pws[gid].i[ 5];
-  wordl1[2] = pws[gid].i[ 6];
-  wordl1[3] = pws[gid].i[ 7];
-
-  u32 wordl2[4];
-
-  wordl2[0] = 0;
-  wordl2[1] = 0;
-  wordl2[2] = 0;
-  wordl2[3] = 0;
-
-  u32 wordl3[4];
-
-  wordl3[0] = 0;
-  wordl3[1] = 0;
-  wordl3[2] = 0;
-  wordl3[3] = 0;
-
-  const u32 pw_l_len = pws[gid].pw_len;
+  __local u32 l_bin2asc[256];
 
-  if (combs_mode == COMBINATOR_MODE_BASE_RIGHT)
+  for (u32 i = lid; i < 256; i += lsz)
   {
-    append_0x80_2x4 (wordl0, wordl1, pw_l_len);
+    const u32 i0 = (i >> 0) & 15;
+    const u32 i1 = (i >> 4) & 15;
 
-    switch_buffer_by_offset (wordl0, wordl1, wordl2, wordl3, combs_buf[0].pw_len);
+    l_bin2asc[i] = ((i0 < 10) ? '0' + i0 : 'a' - 10 + i0) << 0
+                 | ((i1 < 10) ? '0' + i1 : 'a' - 10 + i1) << 8;
   }
 
+  barrier (CLK_LOCAL_MEM_FENCE);
+
+  if (gid >= gid_max) return;
+
   /**
-   * bin2asc table
+   * base
    */
 
-  __local u32 l_bin2asc[256];
+  u32 pw_buf0[4];
+  u32 pw_buf1[4];
 
-  const u32 lid4 = lid * 4;
-
-  const u32 lid40 = lid4 + 0;
-  const u32 lid41 = lid4 + 1;
-  const u32 lid42 = lid4 + 2;
-  const u32 lid43 = lid4 + 3;
-
-  const u32 v400 = (lid40 >> 0) & 15;
-  const u32 v401 = (lid40 >> 4) & 15;
-  const u32 v410 = (lid41 >> 0) & 15;
-  const u32 v411 = (lid41 >> 4) & 15;
-  const u32 v420 = (lid42 >> 0) & 15;
-  const u32 v421 = (lid42 >> 4) & 15;
-  const u32 v430 = (lid43 >> 0) & 15;
-  const u32 v431 = (lid43 >> 4) & 15;
-
-  l_bin2asc[lid40] = ((v400 < 10) ? '0' + v400 : 'a' - 10 + v400) << 0
-                   | ((v401 < 10) ? '0' + v401 : 'a' - 10 + v401) << 8;
-  l_bin2asc[lid41] = ((v410 < 10) ? '0' + v410 : 'a' - 10 + v410) << 0
-                   | ((v411 < 10) ? '0' + v411 : 'a' - 10 + v411) << 8;
-  l_bin2asc[lid42] = ((v420 < 10) ? '0' + v420 : 'a' - 10 + v420) << 0
-                   | ((v421 < 10) ? '0' + v421 : 'a' - 10 + v421) << 8;
-  l_bin2asc[lid43] = ((v430 < 10) ? '0' + v430 : 'a' - 10 + v430) << 0
-                   | ((v431 < 10) ? '0' + v431 : 'a' - 10 + v431) << 8;
+  pw_buf0[0] = pws[gid].i[0];
+  pw_buf0[1] = pws[gid].i[1];
+  pw_buf0[2] = pws[gid].i[2];
+  pw_buf0[3] = pws[gid].i[3];
+  pw_buf1[0] = pws[gid].i[4];
+  pw_buf1[1] = pws[gid].i[5];
+  pw_buf1[2] = pws[gid].i[6];
+  pw_buf1[3] = pws[gid].i[7];
 
-  barrier (CLK_LOCAL_MEM_FENCE);
-
-  if (gid >= gid_max) return;
+  const u32 pw_l_len = pws[gid].pw_len;
 
   /**
    * salt
    */
 
   u32 salt_buf0[4];
-
-  salt_buf0[0] = swap32 (salt_bufs[salt_pos].salt_buf[ 0]);
-  salt_buf0[1] = swap32 (salt_bufs[salt_pos].salt_buf[ 1]);
-  salt_buf0[2] = swap32 (salt_bufs[salt_pos].salt_buf[ 2]);
-  salt_buf0[3] = swap32 (salt_bufs[salt_pos].salt_buf[ 3]);
-
   u32 salt_buf1[4];
-
-  salt_buf1[0] = swap32 (salt_bufs[salt_pos].salt_buf[ 4]);
-  salt_buf1[1] = swap32 (salt_bufs[salt_pos].salt_buf[ 5]);
-  salt_buf1[2] = swap32 (salt_bufs[salt_pos].salt_buf[ 6]);
-  salt_buf1[3] = swap32 (salt_bufs[salt_pos].salt_buf[ 7]);
-
   u32 salt_buf2[4];
 
-  salt_buf2[0] = swap32 (salt_bufs[salt_pos].salt_buf[ 8]);
-  salt_buf2[1] = swap32 (salt_bufs[salt_pos].salt_buf[ 9]);
+  salt_buf0[0] = swap32_S (salt_bufs[salt_pos].salt_buf[ 0]);
+  salt_buf0[1] = swap32_S (salt_bufs[salt_pos].salt_buf[ 1]);
+  salt_buf0[2] = swap32_S (salt_bufs[salt_pos].salt_buf[ 2]);
+  salt_buf0[3] = swap32_S (salt_bufs[salt_pos].salt_buf[ 3]);
+  salt_buf1[0] = swap32_S (salt_bufs[salt_pos].salt_buf[ 4]);
+  salt_buf1[1] = swap32_S (salt_bufs[salt_pos].salt_buf[ 5]);
+  salt_buf1[2] = swap32_S (salt_bufs[salt_pos].salt_buf[ 6]);
+  salt_buf1[3] = swap32_S (salt_bufs[salt_pos].salt_buf[ 7]);
+  salt_buf2[0] = swap32_S (salt_bufs[salt_pos].salt_buf[ 8]);
+  salt_buf2[1] = swap32_S (salt_bufs[salt_pos].salt_buf[ 9]);
   salt_buf2[2] = 0;
   salt_buf2[3] = 0;
 
@@ -267,104 +228,97 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m08400_m04 (__glo
    * loop
    */
 
-  for (u32 il_pos = 0; il_pos < combs_cnt; il_pos++)
+  for (u32 il_pos = 0; il_pos < il_cnt; il_pos += VECT_SIZE)
   {
-    const u32 pw_r_len = combs_buf[il_pos].pw_len;
-
-    const u32 pw_len = pw_l_len + pw_r_len;
-
-    u32 wordr0[4];
-
-    wordr0[0] = combs_buf[il_pos].i[0];
-    wordr0[1] = combs_buf[il_pos].i[1];
-    wordr0[2] = combs_buf[il_pos].i[2];
-    wordr0[3] = combs_buf[il_pos].i[3];
-
-    u32 wordr1[4];
-
-    wordr1[0] = combs_buf[il_pos].i[4];
-    wordr1[1] = combs_buf[il_pos].i[5];
-    wordr1[2] = combs_buf[il_pos].i[6];
-    wordr1[3] = combs_buf[il_pos].i[7];
-
-    u32 wordr2[4];
-
-    wordr2[0] = 0;
-    wordr2[1] = 0;
-    wordr2[2] = 0;
-    wordr2[3] = 0;
-
-    u32 wordr3[4];
-
-    wordr3[0] = 0;
-    wordr3[1] = 0;
-    wordr3[2] = 0;
-    wordr3[3] = 0;
+    const u32x pw_r_len = pwlenx_create_combt (combs_buf, il_pos);
+
+    const u32x pw_len = pw_l_len + pw_r_len;
+
+    /**
+     * concat password candidate
+     */
+
+    u32x wordl0[4] = { 0 };
+    u32x wordl1[4] = { 0 };
+    u32x wordl2[4] = { 0 };
+    u32x wordl3[4] = { 0 };
+
+    wordl0[0] = pw_buf0[0];
+    wordl0[1] = pw_buf0[1];
+    wordl0[2] = pw_buf0[2];
+    wordl0[3] = pw_buf0[3];
+    wordl1[0] = pw_buf1[0];
+    wordl1[1] = pw_buf1[1];
+    wordl1[2] = pw_buf1[2];
+    wordl1[3] = pw_buf1[3];
+
+    u32x wordr0[4] = { 0 };
+    u32x wordr1[4] = { 0 };
+    u32x wordr2[4] = { 0 };
+    u32x wordr3[4] = { 0 };
+
+    wordr0[0] = ix_create_combt (combs_buf, il_pos, 0);
+    wordr0[1] = ix_create_combt (combs_buf, il_pos, 1);
+    wordr0[2] = ix_create_combt (combs_buf, il_pos, 2);
+    wordr0[3] = ix_create_combt (combs_buf, il_pos, 3);
+    wordr1[0] = ix_create_combt (combs_buf, il_pos, 4);
+    wordr1[1] = ix_create_combt (combs_buf, il_pos, 5);
+    wordr1[2] = ix_create_combt (combs_buf, il_pos, 6);
+    wordr1[3] = ix_create_combt (combs_buf, il_pos, 7);
 
     if (combs_mode == COMBINATOR_MODE_BASE_LEFT)
     {
-      append_0x80_2x4 (wordr0, wordr1, pw_r_len);
-
-      switch_buffer_by_offset (wordr0, wordr1, wordr2, wordr3, pw_l_len);
+      switch_buffer_by_offset_le_VV (wordr0, wordr1, wordr2, wordr3, pw_l_len);
+    }
+    else
+    {
+      switch_buffer_by_offset_le_VV (wordl0, wordl1, wordl2, wordl3, pw_r_len);
     }
 
-    u32 w0[4];
+    u32x w0[4];
+    u32x w1[4];
+    u32x w2[4];
+    u32x w3[4];
 
     w0[0] = wordl0[0] | wordr0[0];
     w0[1] = wordl0[1] | wordr0[1];
     w0[2] = wordl0[2] | wordr0[2];
     w0[3] = wordl0[3] | wordr0[3];
-
-    u32 w1[4];
-
     w1[0] = wordl1[0] | wordr1[0];
     w1[1] = wordl1[1] | wordr1[1];
     w1[2] = wordl1[2] | wordr1[2];
     w1[3] = wordl1[3] | wordr1[3];
-
-    u32 w2[4];
-
     w2[0] = wordl2[0] | wordr2[0];
     w2[1] = wordl2[1] | wordr2[1];
     w2[2] = wordl2[2] | wordr2[2];
     w2[3] = wordl2[3] | wordr2[3];
-
-    u32 w3[4];
-
     w3[0] = wordl3[0] | wordr3[0];
     w3[1] = wordl3[1] | wordr3[1];
     w3[2] = wordl3[2] | wordr3[2];
     w3[3] = wordl3[3] | wordr3[3];
 
-    u32 w0_t[4];
-
-    w0_t[0] = swap32 (w0[0]);
-    w0_t[1] = swap32 (w0[1]);
-    w0_t[2] = swap32 (w0[2]);
-    w0_t[3] = swap32 (w0[3]);
-
-    u32 w1_t[4];
-
-    w1_t[0] = swap32 (w1[0]);
-    w1_t[1] = swap32 (w1[1]);
-    w1_t[2] = swap32 (w1[2]);
-    w1_t[3] = swap32 (w1[3]);
-
-    u32 w2_t[4];
-
-    w2_t[0] = swap32 (w2[0]);
-    w2_t[1] = swap32 (w2[1]);
-    w2_t[2] = swap32 (w2[2]);
-    w2_t[3] = swap32 (w2[3]);
-
-    u32 w3_t[4];
-
-    w3_t[0] = swap32 (w3[0]);
-    w3_t[1] = swap32 (w3[1]);
-    w3_t[2] = 0;
-    w3_t[3] = pw_len * 8;
-
-    u32 digest[5];
+    /**
+     * SHA1
+     */
+
+    w0[0] = swap32 (w0[0]);
+    w0[1] = swap32 (w0[1]);
+    w0[2] = swap32 (w0[2]);
+    w0[3] = swap32 (w0[3]);
+    w1[0] = swap32 (w1[0]);
+    w1[1] = swap32 (w1[1]);
+    w1[2] = swap32 (w1[2]);
+    w1[3] = swap32 (w1[3]);
+    w2[0] = swap32 (w2[0]);
+    w2[1] = swap32 (w2[1]);
+    w2[2] = swap32 (w2[2]);
+    w2[3] = swap32 (w2[3]);
+    w3[0] = swap32 (w3[0]);
+    w3[1] = swap32 (w3[1]);
+    w3[2] = 0;
+    w3[3] = pw_len * 8;
+
+    u32x digest[5];
 
     digest[0] = SHA1M_A;
     digest[1] = SHA1M_B;
@@ -372,13 +326,13 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m08400_m04 (__glo
     digest[3] = SHA1M_D;
     digest[4] = SHA1M_E;
 
-    sha1_transform (w0_t, w1_t, w2_t, w3_t, digest);
+    sha1_transform (w0, w1, w2, w3, digest);
 
-    u32 a;
-    u32 b;
-    u32 c;
-    u32 d;
-    u32 e;
+    u32x a;
+    u32x b;
+    u32x c;
+    u32x d;
+    u32x e;
 
     a = digest[0];
     b = digest[1];
@@ -386,28 +340,28 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m08400_m04 (__glo
     d = digest[3];
     e = digest[4];
 
-    w0_t[0] = salt_buf0[0];
-    w0_t[1] = salt_buf0[1];
-    w0_t[2] = salt_buf0[2];
-    w0_t[3] = salt_buf0[3];
-    w1_t[0] = salt_buf1[0];
-    w1_t[1] = salt_buf1[1];
-    w1_t[2] = salt_buf1[2];
-    w1_t[3] = salt_buf1[3];
-    w2_t[0] = salt_buf2[0];
-    w2_t[1] = salt_buf2[1];
-    w2_t[2] = uint_to_hex_lower8_le ((a >> 16) & 255) <<  0
-            | uint_to_hex_lower8_le ((a >> 24) & 255) << 16;
-    w2_t[3] = uint_to_hex_lower8_le ((a >>  0) & 255) <<  0
-            | uint_to_hex_lower8_le ((a >>  8) & 255) << 16;
-    w3_t[0] = uint_to_hex_lower8_le ((b >> 16) & 255) <<  0
-            | uint_to_hex_lower8_le ((b >> 24) & 255) << 16;
-    w3_t[1] = uint_to_hex_lower8_le ((b >>  0) & 255) <<  0
-            | uint_to_hex_lower8_le ((b >>  8) & 255) << 16;
-    w3_t[2] = uint_to_hex_lower8_le ((c >> 16) & 255) <<  0
-            | uint_to_hex_lower8_le ((c >> 24) & 255) << 16;
-    w3_t[3] = uint_to_hex_lower8_le ((c >>  0) & 255) <<  0
-            | uint_to_hex_lower8_le ((c >>  8) & 255) << 16;
+    w0[0] = salt_buf0[0];
+    w0[1] = salt_buf0[1];
+    w0[2] = salt_buf0[2];
+    w0[3] = salt_buf0[3];
+    w1[0] = salt_buf1[0];
+    w1[1] = salt_buf1[1];
+    w1[2] = salt_buf1[2];
+    w1[3] = salt_buf1[3];
+    w2[0] = salt_buf2[0];
+    w2[1] = salt_buf2[1];
+    w2[2] = uint_to_hex_lower8_le ((a >> 16) & 255) <<  0
+          | uint_to_hex_lower8_le ((a >> 24) & 255) << 16;
+    w2[3] = uint_to_hex_lower8_le ((a >>  0) & 255) <<  0
+          | uint_to_hex_lower8_le ((a >>  8) & 255) << 16;
+    w3[0] = uint_to_hex_lower8_le ((b >> 16) & 255) <<  0
+          | uint_to_hex_lower8_le ((b >> 24) & 255) << 16;
+    w3[1] = uint_to_hex_lower8_le ((b >>  0) & 255) <<  0
+          | uint_to_hex_lower8_le ((b >>  8) & 255) << 16;
+    w3[2] = uint_to_hex_lower8_le ((c >> 16) & 255) <<  0
+          | uint_to_hex_lower8_le ((c >> 24) & 255) << 16;
+    w3[3] = uint_to_hex_lower8_le ((c >>  0) & 255) <<  0
+          | uint_to_hex_lower8_le ((c >>  8) & 255) << 16;
 
     digest[0] = SHA1M_A;
     digest[1] = SHA1M_B;
@@ -415,30 +369,30 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m08400_m04 (__glo
     digest[3] = SHA1M_D;
     digest[4] = SHA1M_E;
 
-    sha1_transform (w0_t, w1_t, w2_t, w3_t, digest);
-
-    w0_t[0] = uint_to_hex_lower8_le ((d >> 16) & 255) <<  0
-            | uint_to_hex_lower8_le ((d >> 24) & 255) << 16;
-    w0_t[1] = uint_to_hex_lower8_le ((d >>  0) & 255) <<  0
-            | uint_to_hex_lower8_le ((d >>  8) & 255) << 16;
-    w0_t[2] = uint_to_hex_lower8_le ((e >> 16) & 255) <<  0
-            | uint_to_hex_lower8_le ((e >> 24) & 255) << 16;
-    w0_t[3] = uint_to_hex_lower8_le ((e >>  0) & 255) <<  0
-            | uint_to_hex_lower8_le ((e >>  8) & 255) << 16;
-    w1_t[0] = 0x80000000;
-    w1_t[1] = 0;
-    w1_t[2] = 0;
-    w1_t[3] = 0;
-    w2_t[0] = 0;
-    w2_t[1] = 0;
-    w2_t[2] = 0;
-    w2_t[3] = 0;
-    w3_t[0] = 0;
-    w3_t[1] = 0;
-    w3_t[2] = 0;
-    w3_t[3] = 80 * 8;
-
-    sha1_transform (w0_t, w1_t, w2_t, w3_t, digest);
+    sha1_transform (w0, w1, w2, w3, digest);
+
+    w0[0] = uint_to_hex_lower8_le ((d >> 16) & 255) <<  0
+          | uint_to_hex_lower8_le ((d >> 24) & 255) << 16;
+    w0[1] = uint_to_hex_lower8_le ((d >>  0) & 255) <<  0
+          | uint_to_hex_lower8_le ((d >>  8) & 255) << 16;
+    w0[2] = uint_to_hex_lower8_le ((e >> 16) & 255) <<  0
+          | uint_to_hex_lower8_le ((e >> 24) & 255) << 16;
+    w0[3] = uint_to_hex_lower8_le ((e >>  0) & 255) <<  0
+          | uint_to_hex_lower8_le ((e >>  8) & 255) << 16;
+    w1[0] = 0x80000000;
+    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] = 0;
+    w3[3] = 80 * 8;
+
+    sha1_transform (w0, w1, w2, w3, digest);
 
     a = digest[0];
     b = digest[1];
@@ -446,28 +400,28 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m08400_m04 (__glo
     d = digest[3];
     e = digest[4];
 
-    w0_t[0] = salt_buf0[0];
-    w0_t[1] = salt_buf0[1];
-    w0_t[2] = salt_buf0[2];
-    w0_t[3] = salt_buf0[3];
-    w1_t[0] = salt_buf1[0];
-    w1_t[1] = salt_buf1[1];
-    w1_t[2] = salt_buf1[2];
-    w1_t[3] = salt_buf1[3];
-    w2_t[0] = salt_buf2[0];
-    w2_t[1] = salt_buf2[1];
-    w2_t[2] = uint_to_hex_lower8_le ((a >> 16) & 255) <<  0
-            | uint_to_hex_lower8_le ((a >> 24) & 255) << 16;
-    w2_t[3] = uint_to_hex_lower8_le ((a >>  0) & 255) <<  0
-            | uint_to_hex_lower8_le ((a >>  8) & 255) << 16;
-    w3_t[0] = uint_to_hex_lower8_le ((b >> 16) & 255) <<  0
-            | uint_to_hex_lower8_le ((b >> 24) & 255) << 16;
-    w3_t[1] = uint_to_hex_lower8_le ((b >>  0) & 255) <<  0
-            | uint_to_hex_lower8_le ((b >>  8) & 255) << 16;
-    w3_t[2] = uint_to_hex_lower8_le ((c >> 16) & 255) <<  0
-            | uint_to_hex_lower8_le ((c >> 24) & 255) << 16;
-    w3_t[3] = uint_to_hex_lower8_le ((c >>  0) & 255) <<  0
-            | uint_to_hex_lower8_le ((c >>  8) & 255) << 16;
+    w0[0] = salt_buf0[0];
+    w0[1] = salt_buf0[1];
+    w0[2] = salt_buf0[2];
+    w0[3] = salt_buf0[3];
+    w1[0] = salt_buf1[0];
+    w1[1] = salt_buf1[1];
+    w1[2] = salt_buf1[2];
+    w1[3] = salt_buf1[3];
+    w2[0] = salt_buf2[0];
+    w2[1] = salt_buf2[1];
+    w2[2] = uint_to_hex_lower8_le ((a >> 16) & 255) <<  0
+          | uint_to_hex_lower8_le ((a >> 24) & 255) << 16;
+    w2[3] = uint_to_hex_lower8_le ((a >>  0) & 255) <<  0
+          | uint_to_hex_lower8_le ((a >>  8) & 255) << 16;
+    w3[0] = uint_to_hex_lower8_le ((b >> 16) & 255) <<  0
+          | uint_to_hex_lower8_le ((b >> 24) & 255) << 16;
+    w3[1] = uint_to_hex_lower8_le ((b >>  0) & 255) <<  0
+          | uint_to_hex_lower8_le ((b >>  8) & 255) << 16;
+    w3[2] = uint_to_hex_lower8_le ((c >> 16) & 255) <<  0
+          | uint_to_hex_lower8_le ((c >> 24) & 255) << 16;
+    w3[3] = uint_to_hex_lower8_le ((c >>  0) & 255) <<  0
+          | uint_to_hex_lower8_le ((c >>  8) & 255) << 16;
 
     digest[0] = SHA1M_A;
     digest[1] = SHA1M_B;
@@ -475,156 +429,108 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m08400_m04 (__glo
     digest[3] = SHA1M_D;
     digest[4] = SHA1M_E;
 
-    sha1_transform (w0_t, w1_t, w2_t, w3_t, digest);
-
-    w0_t[0] = uint_to_hex_lower8_le ((d >> 16) & 255) <<  0
-            | uint_to_hex_lower8_le ((d >> 24) & 255) << 16;
-    w0_t[1] = uint_to_hex_lower8_le ((d >>  0) & 255) <<  0
-            | uint_to_hex_lower8_le ((d >>  8) & 255) << 16;
-    w0_t[2] = uint_to_hex_lower8_le ((e >> 16) & 255) <<  0
-            | uint_to_hex_lower8_le ((e >> 24) & 255) << 16;
-    w0_t[3] = uint_to_hex_lower8_le ((e >>  0) & 255) <<  0
-            | uint_to_hex_lower8_le ((e >>  8) & 255) << 16;
-    w1_t[0] = 0x80000000;
-    w1_t[1] = 0;
-    w1_t[2] = 0;
-    w1_t[3] = 0;
-    w2_t[0] = 0;
-    w2_t[1] = 0;
-    w2_t[2] = 0;
-    w2_t[3] = 0;
-    w3_t[0] = 0;
-    w3_t[1] = 0;
-    w3_t[2] = 0;
-    w3_t[3] = 80 * 8;
-
-    sha1_transform (w0_t, w1_t, w2_t, w3_t, digest);
-
-    const u32 r0 = digest[3];
-    const u32 r1 = digest[4];
-    const u32 r2 = digest[2];
-    const u32 r3 = digest[1];
-
-    #include COMPARE_M
+    sha1_transform (w0, w1, w2, w3, digest);
+
+    w0[0] = uint_to_hex_lower8_le ((d >> 16) & 255) <<  0
+          | uint_to_hex_lower8_le ((d >> 24) & 255) << 16;
+    w0[1] = uint_to_hex_lower8_le ((d >>  0) & 255) <<  0
+          | uint_to_hex_lower8_le ((d >>  8) & 255) << 16;
+    w0[2] = uint_to_hex_lower8_le ((e >> 16) & 255) <<  0
+          | uint_to_hex_lower8_le ((e >> 24) & 255) << 16;
+    w0[3] = uint_to_hex_lower8_le ((e >>  0) & 255) <<  0
+          | uint_to_hex_lower8_le ((e >>  8) & 255) << 16;
+    w1[0] = 0x80000000;
+    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] = 0;
+    w3[3] = 80 * 8;
+
+    sha1_transform (w0, w1, w2, w3, digest);
+
+    COMPARE_M_SIMD (digest[3], digest[4], digest[2], digest[1]);
   }
 }
 
-__kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m08400_m08 (__global pw_t *pws, __global gpu_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global void *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 void *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 combs_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
+__kernel void m08400_m08 (__global pw_t *pws, __global kernel_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global void *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 void *esalt_bufs, __global u32 *d_return_buf, __global u32 *d_scryptV0_buf, __global u32 *d_scryptV1_buf, __global u32 *d_scryptV2_buf, __global u32 *d_scryptV3_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)
 {
 }
 
-__kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m08400_m16 (__global pw_t *pws, __global gpu_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global void *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 void *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 combs_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
+__kernel void m08400_m16 (__global pw_t *pws, __global kernel_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global void *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 void *esalt_bufs, __global u32 *d_return_buf, __global u32 *d_scryptV0_buf, __global u32 *d_scryptV1_buf, __global u32 *d_scryptV2_buf, __global u32 *d_scryptV3_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)
 {
 }
 
-__kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m08400_s04 (__global pw_t *pws, __global gpu_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global void *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 void *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 combs_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
+__kernel void m08400_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global void *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 void *esalt_bufs, __global u32 *d_return_buf, __global u32 *d_scryptV0_buf, __global u32 *d_scryptV1_buf, __global u32 *d_scryptV2_buf, __global u32 *d_scryptV3_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)
 {
   /**
    * modifier
    */
 
+  const u32 gid = get_global_id (0);
   const u32 lid = get_local_id (0);
+  const u32 lsz = get_local_size (0);
 
   /**
-   * base
+   * shared
    */
 
-  const u32 gid = get_global_id (0);
-
-  u32 wordl0[4];
-
-  wordl0[0] = pws[gid].i[ 0];
-  wordl0[1] = pws[gid].i[ 1];
-  wordl0[2] = pws[gid].i[ 2];
-  wordl0[3] = pws[gid].i[ 3];
-
-  u32 wordl1[4];
-
-  wordl1[0] = pws[gid].i[ 4];
-  wordl1[1] = pws[gid].i[ 5];
-  wordl1[2] = pws[gid].i[ 6];
-  wordl1[3] = pws[gid].i[ 7];
-
-  u32 wordl2[4];
-
-  wordl2[0] = 0;
-  wordl2[1] = 0;
-  wordl2[2] = 0;
-  wordl2[3] = 0;
-
-  u32 wordl3[4];
-
-  wordl3[0] = 0;
-  wordl3[1] = 0;
-  wordl3[2] = 0;
-  wordl3[3] = 0;
-
-  const u32 pw_l_len = pws[gid].pw_len;
+  __local u32 l_bin2asc[256];
 
-  if (combs_mode == COMBINATOR_MODE_BASE_RIGHT)
+  for (u32 i = lid; i < 256; i += lsz)
   {
-    append_0x80_2x4 (wordl0, wordl1, pw_l_len);
+    const u32 i0 = (i >> 0) & 15;
+    const u32 i1 = (i >> 4) & 15;
 
-    switch_buffer_by_offset (wordl0, wordl1, wordl2, wordl3, combs_buf[0].pw_len);
+    l_bin2asc[i] = ((i0 < 10) ? '0' + i0 : 'a' - 10 + i0) << 0
+                 | ((i1 < 10) ? '0' + i1 : 'a' - 10 + i1) << 8;
   }
 
+  barrier (CLK_LOCAL_MEM_FENCE);
+
+  if (gid >= gid_max) return;
+
   /**
-   * bin2asc table
+   * base
    */
 
-  __local u32 l_bin2asc[256];
-
-  const u32 lid4 = lid * 4;
-
-  const u32 lid40 = lid4 + 0;
-  const u32 lid41 = lid4 + 1;
-  const u32 lid42 = lid4 + 2;
-  const u32 lid43 = lid4 + 3;
-
-  const u32 v400 = (lid40 >> 0) & 15;
-  const u32 v401 = (lid40 >> 4) & 15;
-  const u32 v410 = (lid41 >> 0) & 15;
-  const u32 v411 = (lid41 >> 4) & 15;
-  const u32 v420 = (lid42 >> 0) & 15;
-  const u32 v421 = (lid42 >> 4) & 15;
-  const u32 v430 = (lid43 >> 0) & 15;
-  const u32 v431 = (lid43 >> 4) & 15;
-
-  l_bin2asc[lid40] = ((v400 < 10) ? '0' + v400 : 'a' - 10 + v400) << 0
-                   | ((v401 < 10) ? '0' + v401 : 'a' - 10 + v401) << 8;
-  l_bin2asc[lid41] = ((v410 < 10) ? '0' + v410 : 'a' - 10 + v410) << 0
-                   | ((v411 < 10) ? '0' + v411 : 'a' - 10 + v411) << 8;
-  l_bin2asc[lid42] = ((v420 < 10) ? '0' + v420 : 'a' - 10 + v420) << 0
-                   | ((v421 < 10) ? '0' + v421 : 'a' - 10 + v421) << 8;
-  l_bin2asc[lid43] = ((v430 < 10) ? '0' + v430 : 'a' - 10 + v430) << 0
-                   | ((v431 < 10) ? '0' + v431 : 'a' - 10 + v431) << 8;
+  u32 pw_buf0[4];
+  u32 pw_buf1[4];
 
-  barrier (CLK_LOCAL_MEM_FENCE);
+  pw_buf0[0] = pws[gid].i[0];
+  pw_buf0[1] = pws[gid].i[1];
+  pw_buf0[2] = pws[gid].i[2];
+  pw_buf0[3] = pws[gid].i[3];
+  pw_buf1[0] = pws[gid].i[4];
+  pw_buf1[1] = pws[gid].i[5];
+  pw_buf1[2] = pws[gid].i[6];
+  pw_buf1[3] = pws[gid].i[7];
 
-  if (gid >= gid_max) return;
+  const u32 pw_l_len = pws[gid].pw_len;
 
   /**
    * salt
    */
 
   u32 salt_buf0[4];
-
-  salt_buf0[0] = swap32 (salt_bufs[salt_pos].salt_buf[ 0]);
-  salt_buf0[1] = swap32 (salt_bufs[salt_pos].salt_buf[ 1]);
-  salt_buf0[2] = swap32 (salt_bufs[salt_pos].salt_buf[ 2]);
-  salt_buf0[3] = swap32 (salt_bufs[salt_pos].salt_buf[ 3]);
-
   u32 salt_buf1[4];
-
-  salt_buf1[0] = swap32 (salt_bufs[salt_pos].salt_buf[ 4]);
-  salt_buf1[1] = swap32 (salt_bufs[salt_pos].salt_buf[ 5]);
-  salt_buf1[2] = swap32 (salt_bufs[salt_pos].salt_buf[ 6]);
-  salt_buf1[3] = swap32 (salt_bufs[salt_pos].salt_buf[ 7]);
-
   u32 salt_buf2[4];
 
-  salt_buf2[0] = swap32 (salt_bufs[salt_pos].salt_buf[ 8]);
-  salt_buf2[1] = swap32 (salt_bufs[salt_pos].salt_buf[ 9]);
+  salt_buf0[0] = swap32_S (salt_bufs[salt_pos].salt_buf[ 0]);
+  salt_buf0[1] = swap32_S (salt_bufs[salt_pos].salt_buf[ 1]);
+  salt_buf0[2] = swap32_S (salt_bufs[salt_pos].salt_buf[ 2]);
+  salt_buf0[3] = swap32_S (salt_bufs[salt_pos].salt_buf[ 3]);
+  salt_buf1[0] = swap32_S (salt_bufs[salt_pos].salt_buf[ 4]);
+  salt_buf1[1] = swap32_S (salt_bufs[salt_pos].salt_buf[ 5]);
+  salt_buf1[2] = swap32_S (salt_bufs[salt_pos].salt_buf[ 6]);
+  salt_buf1[3] = swap32_S (salt_bufs[salt_pos].salt_buf[ 7]);
+  salt_buf2[0] = swap32_S (salt_bufs[salt_pos].salt_buf[ 8]);
+  salt_buf2[1] = swap32_S (salt_bufs[salt_pos].salt_buf[ 9]);
   salt_buf2[2] = 0;
   salt_buf2[3] = 0;
 
@@ -646,104 +552,97 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m08400_s04 (__glo
    * loop
    */
 
-  for (u32 il_pos = 0; il_pos < combs_cnt; il_pos++)
+  for (u32 il_pos = 0; il_pos < il_cnt; il_pos += VECT_SIZE)
   {
-    const u32 pw_r_len = combs_buf[il_pos].pw_len;
-
-    const u32 pw_len = pw_l_len + pw_r_len;
-
-    u32 wordr0[4];
-
-    wordr0[0] = combs_buf[il_pos].i[0];
-    wordr0[1] = combs_buf[il_pos].i[1];
-    wordr0[2] = combs_buf[il_pos].i[2];
-    wordr0[3] = combs_buf[il_pos].i[3];
-
-    u32 wordr1[4];
-
-    wordr1[0] = combs_buf[il_pos].i[4];
-    wordr1[1] = combs_buf[il_pos].i[5];
-    wordr1[2] = combs_buf[il_pos].i[6];
-    wordr1[3] = combs_buf[il_pos].i[7];
-
-    u32 wordr2[4];
-
-    wordr2[0] = 0;
-    wordr2[1] = 0;
-    wordr2[2] = 0;
-    wordr2[3] = 0;
-
-    u32 wordr3[4];
-
-    wordr3[0] = 0;
-    wordr3[1] = 0;
-    wordr3[2] = 0;
-    wordr3[3] = 0;
+    const u32x pw_r_len = pwlenx_create_combt (combs_buf, il_pos);
+
+    const u32x pw_len = pw_l_len + pw_r_len;
+
+    /**
+     * concat password candidate
+     */
+
+    u32x wordl0[4] = { 0 };
+    u32x wordl1[4] = { 0 };
+    u32x wordl2[4] = { 0 };
+    u32x wordl3[4] = { 0 };
+
+    wordl0[0] = pw_buf0[0];
+    wordl0[1] = pw_buf0[1];
+    wordl0[2] = pw_buf0[2];
+    wordl0[3] = pw_buf0[3];
+    wordl1[0] = pw_buf1[0];
+    wordl1[1] = pw_buf1[1];
+    wordl1[2] = pw_buf1[2];
+    wordl1[3] = pw_buf1[3];
+
+    u32x wordr0[4] = { 0 };
+    u32x wordr1[4] = { 0 };
+    u32x wordr2[4] = { 0 };
+    u32x wordr3[4] = { 0 };
+
+    wordr0[0] = ix_create_combt (combs_buf, il_pos, 0);
+    wordr0[1] = ix_create_combt (combs_buf, il_pos, 1);
+    wordr0[2] = ix_create_combt (combs_buf, il_pos, 2);
+    wordr0[3] = ix_create_combt (combs_buf, il_pos, 3);
+    wordr1[0] = ix_create_combt (combs_buf, il_pos, 4);
+    wordr1[1] = ix_create_combt (combs_buf, il_pos, 5);
+    wordr1[2] = ix_create_combt (combs_buf, il_pos, 6);
+    wordr1[3] = ix_create_combt (combs_buf, il_pos, 7);
 
     if (combs_mode == COMBINATOR_MODE_BASE_LEFT)
     {
-      append_0x80_2x4 (wordr0, wordr1, pw_r_len);
-
-      switch_buffer_by_offset (wordr0, wordr1, wordr2, wordr3, pw_l_len);
+      switch_buffer_by_offset_le_VV (wordr0, wordr1, wordr2, wordr3, pw_l_len);
+    }
+    else
+    {
+      switch_buffer_by_offset_le_VV (wordl0, wordl1, wordl2, wordl3, pw_r_len);
     }
 
-    u32 w0[4];
+    u32x w0[4];
+    u32x w1[4];
+    u32x w2[4];
+    u32x w3[4];
 
     w0[0] = wordl0[0] | wordr0[0];
     w0[1] = wordl0[1] | wordr0[1];
     w0[2] = wordl0[2] | wordr0[2];
     w0[3] = wordl0[3] | wordr0[3];
-
-    u32 w1[4];
-
     w1[0] = wordl1[0] | wordr1[0];
     w1[1] = wordl1[1] | wordr1[1];
     w1[2] = wordl1[2] | wordr1[2];
     w1[3] = wordl1[3] | wordr1[3];
-
-    u32 w2[4];
-
     w2[0] = wordl2[0] | wordr2[0];
     w2[1] = wordl2[1] | wordr2[1];
     w2[2] = wordl2[2] | wordr2[2];
     w2[3] = wordl2[3] | wordr2[3];
-
-    u32 w3[4];
-
     w3[0] = wordl3[0] | wordr3[0];
     w3[1] = wordl3[1] | wordr3[1];
     w3[2] = wordl3[2] | wordr3[2];
     w3[3] = wordl3[3] | wordr3[3];
 
-    u32 w0_t[4];
-
-    w0_t[0] = swap32 (w0[0]);
-    w0_t[1] = swap32 (w0[1]);
-    w0_t[2] = swap32 (w0[2]);
-    w0_t[3] = swap32 (w0[3]);
-
-    u32 w1_t[4];
-
-    w1_t[0] = swap32 (w1[0]);
-    w1_t[1] = swap32 (w1[1]);
-    w1_t[2] = swap32 (w1[2]);
-    w1_t[3] = swap32 (w1[3]);
-
-    u32 w2_t[4];
-
-    w2_t[0] = swap32 (w2[0]);
-    w2_t[1] = swap32 (w2[1]);
-    w2_t[2] = swap32 (w2[2]);
-    w2_t[3] = swap32 (w2[3]);
-
-    u32 w3_t[4];
-
-    w3_t[0] = swap32 (w3[0]);
-    w3_t[1] = swap32 (w3[1]);
-    w3_t[2] = 0;
-    w3_t[3] = pw_len * 8;
-
-    u32 digest[5];
+    /**
+     * SHA1
+     */
+
+    w0[0] = swap32 (w0[0]);
+    w0[1] = swap32 (w0[1]);
+    w0[2] = swap32 (w0[2]);
+    w0[3] = swap32 (w0[3]);
+    w1[0] = swap32 (w1[0]);
+    w1[1] = swap32 (w1[1]);
+    w1[2] = swap32 (w1[2]);
+    w1[3] = swap32 (w1[3]);
+    w2[0] = swap32 (w2[0]);
+    w2[1] = swap32 (w2[1]);
+    w2[2] = swap32 (w2[2]);
+    w2[3] = swap32 (w2[3]);
+    w3[0] = swap32 (w3[0]);
+    w3[1] = swap32 (w3[1]);
+    w3[2] = 0;
+    w3[3] = pw_len * 8;
+
+    u32x digest[5];
 
     digest[0] = SHA1M_A;
     digest[1] = SHA1M_B;
@@ -751,13 +650,13 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m08400_s04 (__glo
     digest[3] = SHA1M_D;
     digest[4] = SHA1M_E;
 
-    sha1_transform (w0_t, w1_t, w2_t, w3_t, digest);
+    sha1_transform (w0, w1, w2, w3, digest);
 
-    u32 a;
-    u32 b;
-    u32 c;
-    u32 d;
-    u32 e;
+    u32x a;
+    u32x b;
+    u32x c;
+    u32x d;
+    u32x e;
 
     a = digest[0];
     b = digest[1];
@@ -765,28 +664,28 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m08400_s04 (__glo
     d = digest[3];
     e = digest[4];
 
-    w0_t[0] = salt_buf0[0];
-    w0_t[1] = salt_buf0[1];
-    w0_t[2] = salt_buf0[2];
-    w0_t[3] = salt_buf0[3];
-    w1_t[0] = salt_buf1[0];
-    w1_t[1] = salt_buf1[1];
-    w1_t[2] = salt_buf1[2];
-    w1_t[3] = salt_buf1[3];
-    w2_t[0] = salt_buf2[0];
-    w2_t[1] = salt_buf2[1];
-    w2_t[2] = uint_to_hex_lower8_le ((a >> 16) & 255) <<  0
-            | uint_to_hex_lower8_le ((a >> 24) & 255) << 16;
-    w2_t[3] = uint_to_hex_lower8_le ((a >>  0) & 255) <<  0
-            | uint_to_hex_lower8_le ((a >>  8) & 255) << 16;
-    w3_t[0] = uint_to_hex_lower8_le ((b >> 16) & 255) <<  0
-            | uint_to_hex_lower8_le ((b >> 24) & 255) << 16;
-    w3_t[1] = uint_to_hex_lower8_le ((b >>  0) & 255) <<  0
-            | uint_to_hex_lower8_le ((b >>  8) & 255) << 16;
-    w3_t[2] = uint_to_hex_lower8_le ((c >> 16) & 255) <<  0
-            | uint_to_hex_lower8_le ((c >> 24) & 255) << 16;
-    w3_t[3] = uint_to_hex_lower8_le ((c >>  0) & 255) <<  0
-            | uint_to_hex_lower8_le ((c >>  8) & 255) << 16;
+    w0[0] = salt_buf0[0];
+    w0[1] = salt_buf0[1];
+    w0[2] = salt_buf0[2];
+    w0[3] = salt_buf0[3];
+    w1[0] = salt_buf1[0];
+    w1[1] = salt_buf1[1];
+    w1[2] = salt_buf1[2];
+    w1[3] = salt_buf1[3];
+    w2[0] = salt_buf2[0];
+    w2[1] = salt_buf2[1];
+    w2[2] = uint_to_hex_lower8_le ((a >> 16) & 255) <<  0
+          | uint_to_hex_lower8_le ((a >> 24) & 255) << 16;
+    w2[3] = uint_to_hex_lower8_le ((a >>  0) & 255) <<  0
+          | uint_to_hex_lower8_le ((a >>  8) & 255) << 16;
+    w3[0] = uint_to_hex_lower8_le ((b >> 16) & 255) <<  0
+          | uint_to_hex_lower8_le ((b >> 24) & 255) << 16;
+    w3[1] = uint_to_hex_lower8_le ((b >>  0) & 255) <<  0
+          | uint_to_hex_lower8_le ((b >>  8) & 255) << 16;
+    w3[2] = uint_to_hex_lower8_le ((c >> 16) & 255) <<  0
+          | uint_to_hex_lower8_le ((c >> 24) & 255) << 16;
+    w3[3] = uint_to_hex_lower8_le ((c >>  0) & 255) <<  0
+          | uint_to_hex_lower8_le ((c >>  8) & 255) << 16;
 
     digest[0] = SHA1M_A;
     digest[1] = SHA1M_B;
@@ -794,30 +693,30 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m08400_s04 (__glo
     digest[3] = SHA1M_D;
     digest[4] = SHA1M_E;
 
-    sha1_transform (w0_t, w1_t, w2_t, w3_t, digest);
-
-    w0_t[0] = uint_to_hex_lower8_le ((d >> 16) & 255) <<  0
-            | uint_to_hex_lower8_le ((d >> 24) & 255) << 16;
-    w0_t[1] = uint_to_hex_lower8_le ((d >>  0) & 255) <<  0
-            | uint_to_hex_lower8_le ((d >>  8) & 255) << 16;
-    w0_t[2] = uint_to_hex_lower8_le ((e >> 16) & 255) <<  0
-            | uint_to_hex_lower8_le ((e >> 24) & 255) << 16;
-    w0_t[3] = uint_to_hex_lower8_le ((e >>  0) & 255) <<  0
-            | uint_to_hex_lower8_le ((e >>  8) & 255) << 16;
-    w1_t[0] = 0x80000000;
-    w1_t[1] = 0;
-    w1_t[2] = 0;
-    w1_t[3] = 0;
-    w2_t[0] = 0;
-    w2_t[1] = 0;
-    w2_t[2] = 0;
-    w2_t[3] = 0;
-    w3_t[0] = 0;
-    w3_t[1] = 0;
-    w3_t[2] = 0;
-    w3_t[3] = (salt_len + 40) * 8;
-
-    sha1_transform (w0_t, w1_t, w2_t, w3_t, digest);
+    sha1_transform (w0, w1, w2, w3, digest);
+
+    w0[0] = uint_to_hex_lower8_le ((d >> 16) & 255) <<  0
+          | uint_to_hex_lower8_le ((d >> 24) & 255) << 16;
+    w0[1] = uint_to_hex_lower8_le ((d >>  0) & 255) <<  0
+          | uint_to_hex_lower8_le ((d >>  8) & 255) << 16;
+    w0[2] = uint_to_hex_lower8_le ((e >> 16) & 255) <<  0
+          | uint_to_hex_lower8_le ((e >> 24) & 255) << 16;
+    w0[3] = uint_to_hex_lower8_le ((e >>  0) & 255) <<  0
+          | uint_to_hex_lower8_le ((e >>  8) & 255) << 16;
+    w1[0] = 0x80000000;
+    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] = 0;
+    w3[3] = 80 * 8;
+
+    sha1_transform (w0, w1, w2, w3, digest);
 
     a = digest[0];
     b = digest[1];
@@ -825,28 +724,28 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m08400_s04 (__glo
     d = digest[3];
     e = digest[4];
 
-    w0_t[0] = salt_buf0[0];
-    w0_t[1] = salt_buf0[1];
-    w0_t[2] = salt_buf0[2];
-    w0_t[3] = salt_buf0[3];
-    w1_t[0] = salt_buf1[0];
-    w1_t[1] = salt_buf1[1];
-    w1_t[2] = salt_buf1[2];
-    w1_t[3] = salt_buf1[3];
-    w2_t[0] = salt_buf2[0];
-    w2_t[1] = salt_buf2[1];
-    w2_t[2] = uint_to_hex_lower8_le ((a >> 16) & 255) <<  0
-            | uint_to_hex_lower8_le ((a >> 24) & 255) << 16;
-    w2_t[3] = uint_to_hex_lower8_le ((a >>  0) & 255) <<  0
-            | uint_to_hex_lower8_le ((a >>  8) & 255) << 16;
-    w3_t[0] = uint_to_hex_lower8_le ((b >> 16) & 255) <<  0
-            | uint_to_hex_lower8_le ((b >> 24) & 255) << 16;
-    w3_t[1] = uint_to_hex_lower8_le ((b >>  0) & 255) <<  0
-            | uint_to_hex_lower8_le ((b >>  8) & 255) << 16;
-    w3_t[2] = uint_to_hex_lower8_le ((c >> 16) & 255) <<  0
-            | uint_to_hex_lower8_le ((c >> 24) & 255) << 16;
-    w3_t[3] = uint_to_hex_lower8_le ((c >>  0) & 255) <<  0
-            | uint_to_hex_lower8_le ((c >>  8) & 255) << 16;
+    w0[0] = salt_buf0[0];
+    w0[1] = salt_buf0[1];
+    w0[2] = salt_buf0[2];
+    w0[3] = salt_buf0[3];
+    w1[0] = salt_buf1[0];
+    w1[1] = salt_buf1[1];
+    w1[2] = salt_buf1[2];
+    w1[3] = salt_buf1[3];
+    w2[0] = salt_buf2[0];
+    w2[1] = salt_buf2[1];
+    w2[2] = uint_to_hex_lower8_le ((a >> 16) & 255) <<  0
+          | uint_to_hex_lower8_le ((a >> 24) & 255) << 16;
+    w2[3] = uint_to_hex_lower8_le ((a >>  0) & 255) <<  0
+          | uint_to_hex_lower8_le ((a >>  8) & 255) << 16;
+    w3[0] = uint_to_hex_lower8_le ((b >> 16) & 255) <<  0
+          | uint_to_hex_lower8_le ((b >> 24) & 255) << 16;
+    w3[1] = uint_to_hex_lower8_le ((b >>  0) & 255) <<  0
+          | uint_to_hex_lower8_le ((b >>  8) & 255) << 16;
+    w3[2] = uint_to_hex_lower8_le ((c >> 16) & 255) <<  0
+          | uint_to_hex_lower8_le ((c >> 24) & 255) << 16;
+    w3[3] = uint_to_hex_lower8_le ((c >>  0) & 255) <<  0
+          | uint_to_hex_lower8_le ((c >>  8) & 255) << 16;
 
     digest[0] = SHA1M_A;
     digest[1] = SHA1M_B;
@@ -854,44 +753,39 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m08400_s04 (__glo
     digest[3] = SHA1M_D;
     digest[4] = SHA1M_E;
 
-    sha1_transform (w0_t, w1_t, w2_t, w3_t, digest);
-
-    w0_t[0] = uint_to_hex_lower8_le ((d >> 16) & 255) <<  0
-            | uint_to_hex_lower8_le ((d >> 24) & 255) << 16;
-    w0_t[1] = uint_to_hex_lower8_le ((d >>  0) & 255) <<  0
-            | uint_to_hex_lower8_le ((d >>  8) & 255) << 16;
-    w0_t[2] = uint_to_hex_lower8_le ((e >> 16) & 255) <<  0
-            | uint_to_hex_lower8_le ((e >> 24) & 255) << 16;
-    w0_t[3] = uint_to_hex_lower8_le ((e >>  0) & 255) <<  0
-            | uint_to_hex_lower8_le ((e >>  8) & 255) << 16;
-    w1_t[0] = 0x80000000;
-    w1_t[1] = 0;
-    w1_t[2] = 0;
-    w1_t[3] = 0;
-    w2_t[0] = 0;
-    w2_t[1] = 0;
-    w2_t[2] = 0;
-    w2_t[3] = 0;
-    w3_t[0] = 0;
-    w3_t[1] = 0;
-    w3_t[2] = 0;
-    w3_t[3] = (salt_len + 40) * 8;
-
-    sha1_transform (w0_t, w1_t, w2_t, w3_t, digest);
-
-    const u32 r0 = digest[3];
-    const u32 r1 = digest[4];
-    const u32 r2 = digest[2];
-    const u32 r3 = digest[1];
-
-    #include COMPARE_S
+    sha1_transform (w0, w1, w2, w3, digest);
+
+    w0[0] = uint_to_hex_lower8_le ((d >> 16) & 255) <<  0
+          | uint_to_hex_lower8_le ((d >> 24) & 255) << 16;
+    w0[1] = uint_to_hex_lower8_le ((d >>  0) & 255) <<  0
+          | uint_to_hex_lower8_le ((d >>  8) & 255) << 16;
+    w0[2] = uint_to_hex_lower8_le ((e >> 16) & 255) <<  0
+          | uint_to_hex_lower8_le ((e >> 24) & 255) << 16;
+    w0[3] = uint_to_hex_lower8_le ((e >>  0) & 255) <<  0
+          | uint_to_hex_lower8_le ((e >>  8) & 255) << 16;
+    w1[0] = 0x80000000;
+    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] = 0;
+    w3[3] = 80 * 8;
+
+    sha1_transform (w0, w1, w2, w3, digest);
+
+    COMPARE_S_SIMD (digest[3], digest[4], digest[2], digest[1]);
   }
 }
 
-__kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m08400_s08 (__global pw_t *pws, __global gpu_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global void *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 void *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 combs_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
+__kernel void m08400_s08 (__global pw_t *pws, __global kernel_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global void *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 void *esalt_bufs, __global u32 *d_return_buf, __global u32 *d_scryptV0_buf, __global u32 *d_scryptV1_buf, __global u32 *d_scryptV2_buf, __global u32 *d_scryptV3_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)
 {
 }
 
-__kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m08400_s16 (__global pw_t *pws, __global gpu_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global void *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 void *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 combs_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
+__kernel void m08400_s16 (__global pw_t *pws, __global kernel_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global void *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 void *esalt_bufs, __global u32 *d_return_buf, __global u32 *d_scryptV0_buf, __global u32 *d_scryptV1_buf, __global u32 *d_scryptV2_buf, __global u32 *d_scryptV3_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)
 {
 }