X-Git-Url: https://www.flypig.org.uk/git/?a=blobdiff_plain;f=OpenCL%2Fm04900_a0.cl;h=0646c591bef8e0a317c551b42460d8ac08867e5e;hb=161a6eb4bc643d8e636e96eda613f5137d30da59;hp=e53112e90365392cd9f42d5c7a0088c8b50c62bc;hpb=b57a143e42048eb8c247427ebae3a3fe5de79c3c;p=hashcat.git

diff --git a/OpenCL/m04900_a0.cl b/OpenCL/m04900_a0.cl
index e53112e..0646c59 100644
--- a/OpenCL/m04900_a0.cl
+++ b/OpenCL/m04900_a0.cl
@@ -5,24 +5,18 @@
 
 #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 "OpenCL/types_ocl.c"
-#include "OpenCL/common.c"
-#include "include/rp_kernel.h"
-#include "OpenCL/rp.c"
-
-#define COMPARE_S "OpenCL/check_single_comp4.c"
-#define COMPARE_M "OpenCL/check_multi_comp4.c"
-
-__kernel void m04900_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_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 rules_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
+#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_rp.h"
+#include "inc_rp.cl"
+#include "inc_simd.cl"
+
+__kernel void m04900_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
@@ -39,18 +33,16 @@ __kernel void m04900_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
   if (gid >= gid_max) return;
 
   u32 pw_buf0[4];
-
-  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];
-
   u32 pw_buf1[4];
 
-  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];
+  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];
 
   const u32 pw_len = pws[gid].pw_len;
 
@@ -59,32 +51,26 @@ __kernel void m04900_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
    */
 
   u32 salt_buf0[4];
+  u32 salt_buf1[4];
+  u32 salt_buf2[4];
+  u32 salt_buf3[4];
 
   salt_buf0[0] = salt_bufs[salt_pos].salt_buf[ 0];
   salt_buf0[1] = salt_bufs[salt_pos].salt_buf[ 1];
   salt_buf0[2] = salt_bufs[salt_pos].salt_buf[ 2];
   salt_buf0[3] = salt_bufs[salt_pos].salt_buf[ 3];
-
-  u32 salt_buf1[4];
-
   salt_buf1[0] = salt_bufs[salt_pos].salt_buf[ 4];
   salt_buf1[1] = salt_bufs[salt_pos].salt_buf[ 5];
   salt_buf1[2] = salt_bufs[salt_pos].salt_buf[ 6];
   salt_buf1[3] = salt_bufs[salt_pos].salt_buf[ 7];
-
-  u32 salt_buf2[4];
-
-  salt_buf2[0] = 0;
-  salt_buf2[1] = 0;
-  salt_buf2[2] = 0;
-  salt_buf2[3] = 0;
-
-  u32 salt_buf3[4];
-
-  salt_buf3[0] = 0;
-  salt_buf3[1] = 0;
-  salt_buf3[2] = 0;
-  salt_buf3[3] = 0;
+  salt_buf2[0] = salt_bufs[salt_pos].salt_buf[ 8];
+  salt_buf2[1] = salt_bufs[salt_pos].salt_buf[ 9];
+  salt_buf2[2] = salt_bufs[salt_pos].salt_buf[10];
+  salt_buf2[3] = salt_bufs[salt_pos].salt_buf[11];
+  salt_buf3[0] = salt_bufs[salt_pos].salt_buf[12];
+  salt_buf3[1] = salt_bufs[salt_pos].salt_buf[13];
+  salt_buf3[2] = salt_bufs[salt_pos].salt_buf[14];
+  salt_buf3[3] = salt_bufs[salt_pos].salt_buf[15];
 
   const u32 salt_len = salt_bufs[salt_pos].salt_len;
 
@@ -92,257 +78,225 @@ __kernel void m04900_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
    * loop
    */
 
-  for (u32 il_pos = 0; il_pos < rules_cnt; il_pos++)
+  for (u32 il_pos = 0; il_pos < il_cnt; il_pos += VECT_SIZE)
   {
-    u32 w0_t[4];
-
-    w0_t[0] = pw_buf0[0];
-    w0_t[1] = pw_buf0[1];
-    w0_t[2] = pw_buf0[2];
-    w0_t[3] = pw_buf0[3];
-
-    u32 w1_t[4];
+    u32x w0[4] = { 0 };
+    u32x w1[4] = { 0 };
+    u32x w2[4] = { 0 };
+    u32x w3[4] = { 0 };
 
-    w1_t[0] = pw_buf1[0];
-    w1_t[1] = pw_buf1[1];
-    w1_t[2] = pw_buf1[2];
-    w1_t[3] = pw_buf1[3];
-
-    u32 w2_t[4];
-
-    w2_t[0] = 0;
-    w2_t[1] = 0;
-    w2_t[2] = 0;
-    w2_t[3] = 0;
-
-    u32 w3_t[4];
-
-    w3_t[0] = 0;
-    w3_t[1] = 0;
-    w3_t[2] = 0;
-    w3_t[3] = 0;
-
-    const u32 out_len = apply_rules (rules_buf[il_pos].cmds, w0_t, w1_t, pw_len);
+    const u32x out_len = apply_rules_vect (pw_buf0, pw_buf1, pw_len, rules_buf, il_pos, w0, w1);
 
     /**
      * prepend salt
      */
 
-    switch_buffer_by_offset_le (w0_t, w1_t, w2_t, w3_t, salt_len);
-
-    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] |= salt_buf2[2];
-    w2_t[3] |= salt_buf2[3];
-    w3_t[0] |= salt_buf3[0];
-    w3_t[1] |= salt_buf3[1];
-    w3_t[2] |= salt_buf3[2];
-    w3_t[3] |= salt_buf3[3];
+    const u32x out_salt_len = out_len + salt_len;
+
+    switch_buffer_by_offset_le_VV (w0, w1, w2, w3, salt_len);
+
+    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] |= salt_buf2[2];
+    w2[3] |= salt_buf2[3];
+    w3[0] |= salt_buf3[0];
+    w3[1] |= salt_buf3[1];
+    w3[2] |= salt_buf3[2];
+    w3[3] |= salt_buf3[3];
 
     /**
      * append salt
      */
 
-    u32 s0[4];
+    u32x s0[4];
+    u32x s1[4];
+    u32x s2[4];
+    u32x s3[4];
 
     s0[0] = salt_buf0[0];
     s0[1] = salt_buf0[1];
     s0[2] = salt_buf0[2];
     s0[3] = salt_buf0[3];
-
-    u32 s1[4];
-
     s1[0] = salt_buf1[0];
     s1[1] = salt_buf1[1];
     s1[2] = salt_buf1[2];
     s1[3] = salt_buf1[3];
-
-    u32 s2[4];
-
-    s2[0] = 0;
-    s2[1] = 0;
-    s2[2] = 0;
-    s2[3] = 0;
-
-    u32 s3[4];
-
-    s3[0] = 0;
-    s3[1] = 0;
-    s3[2] = 0;
-    s3[3] = 0;
-
-    switch_buffer_by_offset_le (s0, s1, s2, s3, salt_len + out_len);
-
-    w0_t[0] |= s0[0];
-    w0_t[1] |= s0[1];
-    w0_t[2] |= s0[2];
-    w0_t[3] |= s0[3];
-    w1_t[0] |= s1[0];
-    w1_t[1] |= s1[1];
-    w1_t[2] |= s1[2];
-    w1_t[3] |= s1[3];
-    w2_t[0] |= s2[0];
-    w2_t[1] |= s2[1];
-    w2_t[2] |= s2[2];
-    w2_t[3] |= s2[3];
-    w3_t[0] |= s3[0];
-    w3_t[1] |= s3[1];
-    w3_t[2] |= s3[2];
-    w3_t[3] |= s3[3];
-
-    const u32 pw_salt_len = salt_len + out_len + salt_len;
-
-    append_0x80_4x4 (w0_t, w1_t, w2_t, w3_t, pw_salt_len);
-
-    u32 w0 = swap32 (w0_t[0]);
-    u32 w1 = swap32 (w0_t[1]);
-    u32 w2 = swap32 (w0_t[2]);
-    u32 w3 = swap32 (w0_t[3]);
-    u32 w4 = swap32 (w1_t[0]);
-    u32 w5 = swap32 (w1_t[1]);
-    u32 w6 = swap32 (w1_t[2]);
-    u32 w7 = swap32 (w1_t[3]);
-    u32 w8 = swap32 (w2_t[0]);
-    u32 w9 = swap32 (w2_t[1]);
-    u32 wa = swap32 (w2_t[2]);
-    u32 wb = swap32 (w2_t[3]);
-    u32 wc = swap32 (w3_t[0]);
-    u32 wd = swap32 (w3_t[1]);
-    u32 we = 0;
-    u32 wf = pw_salt_len * 8;
+    s2[0] = salt_buf2[0];
+    s2[1] = salt_buf2[1];
+    s2[2] = salt_buf2[2];
+    s2[3] = salt_buf2[3];
+    s3[0] = salt_buf3[0];
+    s3[1] = salt_buf3[1];
+    s3[2] = salt_buf3[2];
+    s3[3] = salt_buf3[3];
+
+    switch_buffer_by_offset_le_VV (s0, s1, s2, s3, out_salt_len);
+
+    w0[0] |= s0[0];
+    w0[1] |= s0[1];
+    w0[2] |= s0[2];
+    w0[3] |= s0[3];
+    w1[0] |= s1[0];
+    w1[1] |= s1[1];
+    w1[2] |= s1[2];
+    w1[3] |= s1[3];
+    w2[0] |= s2[0];
+    w2[1] |= s2[1];
+    w2[2] |= s2[2];
+    w2[3] |= s2[3];
+    w3[0] |= s3[0];
+    w3[1] |= s3[1];
+    w3[2] |= s3[2];
+    w3[3] |= s3[3];
+
+    const u32x salt_out_salt_len = salt_len + out_len + salt_len;
+
+    append_0x80_4x4_VV (w0, w1, w2, w3, salt_out_salt_len);
 
     /**
      * sha1
      */
 
-    u32 a = SHA1M_A;
-    u32 b = SHA1M_B;
-    u32 c = SHA1M_C;
-    u32 d = SHA1M_D;
-    u32 e = SHA1M_E;
+    u32x w0_t = swap32 (w0[0]);
+    u32x w1_t = swap32 (w0[1]);
+    u32x w2_t = swap32 (w0[2]);
+    u32x w3_t = swap32 (w0[3]);
+    u32x w4_t = swap32 (w1[0]);
+    u32x w5_t = swap32 (w1[1]);
+    u32x w6_t = swap32 (w1[2]);
+    u32x w7_t = swap32 (w1[3]);
+    u32x w8_t = swap32 (w2[0]);
+    u32x w9_t = swap32 (w2[1]);
+    u32x wa_t = swap32 (w2[2]);
+    u32x wb_t = swap32 (w2[3]);
+    u32x wc_t = swap32 (w3[0]);
+    u32x wd_t = swap32 (w3[1]);
+    u32x we_t = 0;
+    u32x wf_t = salt_out_salt_len * 8;
+
+    u32x a = SHA1M_A;
+    u32x b = SHA1M_B;
+    u32x c = SHA1M_C;
+    u32x d = SHA1M_D;
+    u32x e = SHA1M_E;
 
     #undef K
     #define K SHA1C00
 
-    SHA1_STEP (SHA1_F0o, a, b, c, d, e, w0);
-    SHA1_STEP (SHA1_F0o, e, a, b, c, d, w1);
-    SHA1_STEP (SHA1_F0o, d, e, a, b, c, w2);
-    SHA1_STEP (SHA1_F0o, c, d, e, a, b, w3);
-    SHA1_STEP (SHA1_F0o, b, c, d, e, a, w4);
-    SHA1_STEP (SHA1_F0o, a, b, c, d, e, w5);
-    SHA1_STEP (SHA1_F0o, e, a, b, c, d, w6);
-    SHA1_STEP (SHA1_F0o, d, e, a, b, c, w7);
-    SHA1_STEP (SHA1_F0o, c, d, e, a, b, w8);
-    SHA1_STEP (SHA1_F0o, b, c, d, e, a, w9);
-    SHA1_STEP (SHA1_F0o, a, b, c, d, e, wa);
-    SHA1_STEP (SHA1_F0o, e, a, b, c, d, wb);
-    SHA1_STEP (SHA1_F0o, d, e, a, b, c, wc);
-    SHA1_STEP (SHA1_F0o, c, d, e, a, b, wd);
-    SHA1_STEP (SHA1_F0o, b, c, d, e, a, we);
-    SHA1_STEP (SHA1_F0o, a, b, c, d, e, wf);
-    w0 = rotl32 ((wd ^ w8 ^ w2 ^ w0), 1u); SHA1_STEP (SHA1_F0o, e, a, b, c, d, w0);
-    w1 = rotl32 ((we ^ w9 ^ w3 ^ w1), 1u); SHA1_STEP (SHA1_F0o, d, e, a, b, c, w1);
-    w2 = rotl32 ((wf ^ wa ^ w4 ^ w2), 1u); SHA1_STEP (SHA1_F0o, c, d, e, a, b, w2);
-    w3 = rotl32 ((w0 ^ wb ^ w5 ^ w3), 1u); SHA1_STEP (SHA1_F0o, b, c, d, e, a, w3);
+    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);
+    SHA1_STEP (SHA1_F0o, c, d, e, a, b, w3_t);
+    SHA1_STEP (SHA1_F0o, b, c, d, e, a, w4_t);
+    SHA1_STEP (SHA1_F0o, a, b, c, d, e, w5_t);
+    SHA1_STEP (SHA1_F0o, e, a, b, c, d, w6_t);
+    SHA1_STEP (SHA1_F0o, d, e, a, b, c, w7_t);
+    SHA1_STEP (SHA1_F0o, c, d, e, a, b, w8_t);
+    SHA1_STEP (SHA1_F0o, b, c, d, e, a, w9_t);
+    SHA1_STEP (SHA1_F0o, a, b, c, d, e, wa_t);
+    SHA1_STEP (SHA1_F0o, e, a, b, c, d, wb_t);
+    SHA1_STEP (SHA1_F0o, d, e, a, b, c, wc_t);
+    SHA1_STEP (SHA1_F0o, c, d, e, a, b, wd_t);
+    SHA1_STEP (SHA1_F0o, b, c, d, e, a, we_t);
+    SHA1_STEP (SHA1_F0o, a, b, c, d, e, wf_t);
+    w0_t = rotl32 ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP (SHA1_F0o, e, a, b, c, d, w0_t);
+    w1_t = rotl32 ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP (SHA1_F0o, d, e, a, b, c, w1_t);
+    w2_t = rotl32 ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP (SHA1_F0o, c, d, e, a, b, w2_t);
+    w3_t = rotl32 ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP (SHA1_F0o, b, c, d, e, a, w3_t);
 
     #undef K
     #define K SHA1C01
 
-    w4 = rotl32 ((w1 ^ wc ^ w6 ^ w4), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, w4);
-    w5 = rotl32 ((w2 ^ wd ^ w7 ^ w5), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, w5);
-    w6 = rotl32 ((w3 ^ we ^ w8 ^ w6), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, w6);
-    w7 = rotl32 ((w4 ^ wf ^ w9 ^ w7), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, w7);
-    w8 = rotl32 ((w5 ^ w0 ^ wa ^ w8), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, w8);
-    w9 = rotl32 ((w6 ^ w1 ^ wb ^ w9), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, w9);
-    wa = rotl32 ((w7 ^ w2 ^ wc ^ wa), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, wa);
-    wb = rotl32 ((w8 ^ w3 ^ wd ^ wb), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, wb);
-    wc = rotl32 ((w9 ^ w4 ^ we ^ wc), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, wc);
-    wd = rotl32 ((wa ^ w5 ^ wf ^ wd), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, wd);
-    we = rotl32 ((wb ^ w6 ^ w0 ^ we), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, we);
-    wf = rotl32 ((wc ^ w7 ^ w1 ^ wf), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, wf);
-    w0 = rotl32 ((wd ^ w8 ^ w2 ^ w0), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, w0);
-    w1 = rotl32 ((we ^ w9 ^ w3 ^ w1), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, w1);
-    w2 = rotl32 ((wf ^ wa ^ w4 ^ w2), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, w2);
-    w3 = rotl32 ((w0 ^ wb ^ w5 ^ w3), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, w3);
-    w4 = rotl32 ((w1 ^ wc ^ w6 ^ w4), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, w4);
-    w5 = rotl32 ((w2 ^ wd ^ w7 ^ w5), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, w5);
-    w6 = rotl32 ((w3 ^ we ^ w8 ^ w6), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, w6);
-    w7 = rotl32 ((w4 ^ wf ^ w9 ^ w7), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, w7);
+    w4_t = rotl32 ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, w4_t);
+    w5_t = rotl32 ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, w5_t);
+    w6_t = rotl32 ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, w6_t);
+    w7_t = rotl32 ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, w7_t);
+    w8_t = rotl32 ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, w8_t);
+    w9_t = rotl32 ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, w9_t);
+    wa_t = rotl32 ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, wa_t);
+    wb_t = rotl32 ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, wb_t);
+    wc_t = rotl32 ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, wc_t);
+    wd_t = rotl32 ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, wd_t);
+    we_t = rotl32 ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, we_t);
+    wf_t = rotl32 ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, wf_t);
+    w0_t = rotl32 ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, w0_t);
+    w1_t = rotl32 ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, w1_t);
+    w2_t = rotl32 ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, w2_t);
+    w3_t = rotl32 ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, w3_t);
+    w4_t = rotl32 ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, w4_t);
+    w5_t = rotl32 ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, w5_t);
+    w6_t = rotl32 ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, w6_t);
+    w7_t = rotl32 ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, w7_t);
 
     #undef K
     #define K SHA1C02
 
-    w8 = rotl32 ((w5 ^ w0 ^ wa ^ w8), 1u); SHA1_STEP (SHA1_F2o, a, b, c, d, e, w8);
-    w9 = rotl32 ((w6 ^ w1 ^ wb ^ w9), 1u); SHA1_STEP (SHA1_F2o, e, a, b, c, d, w9);
-    wa = rotl32 ((w7 ^ w2 ^ wc ^ wa), 1u); SHA1_STEP (SHA1_F2o, d, e, a, b, c, wa);
-    wb = rotl32 ((w8 ^ w3 ^ wd ^ wb), 1u); SHA1_STEP (SHA1_F2o, c, d, e, a, b, wb);
-    wc = rotl32 ((w9 ^ w4 ^ we ^ wc), 1u); SHA1_STEP (SHA1_F2o, b, c, d, e, a, wc);
-    wd = rotl32 ((wa ^ w5 ^ wf ^ wd), 1u); SHA1_STEP (SHA1_F2o, a, b, c, d, e, wd);
-    we = rotl32 ((wb ^ w6 ^ w0 ^ we), 1u); SHA1_STEP (SHA1_F2o, e, a, b, c, d, we);
-    wf = rotl32 ((wc ^ w7 ^ w1 ^ wf), 1u); SHA1_STEP (SHA1_F2o, d, e, a, b, c, wf);
-    w0 = rotl32 ((wd ^ w8 ^ w2 ^ w0), 1u); SHA1_STEP (SHA1_F2o, c, d, e, a, b, w0);
-    w1 = rotl32 ((we ^ w9 ^ w3 ^ w1), 1u); SHA1_STEP (SHA1_F2o, b, c, d, e, a, w1);
-    w2 = rotl32 ((wf ^ wa ^ w4 ^ w2), 1u); SHA1_STEP (SHA1_F2o, a, b, c, d, e, w2);
-    w3 = rotl32 ((w0 ^ wb ^ w5 ^ w3), 1u); SHA1_STEP (SHA1_F2o, e, a, b, c, d, w3);
-    w4 = rotl32 ((w1 ^ wc ^ w6 ^ w4), 1u); SHA1_STEP (SHA1_F2o, d, e, a, b, c, w4);
-    w5 = rotl32 ((w2 ^ wd ^ w7 ^ w5), 1u); SHA1_STEP (SHA1_F2o, c, d, e, a, b, w5);
-    w6 = rotl32 ((w3 ^ we ^ w8 ^ w6), 1u); SHA1_STEP (SHA1_F2o, b, c, d, e, a, w6);
-    w7 = rotl32 ((w4 ^ wf ^ w9 ^ w7), 1u); SHA1_STEP (SHA1_F2o, a, b, c, d, e, w7);
-    w8 = rotl32 ((w5 ^ w0 ^ wa ^ w8), 1u); SHA1_STEP (SHA1_F2o, e, a, b, c, d, w8);
-    w9 = rotl32 ((w6 ^ w1 ^ wb ^ w9), 1u); SHA1_STEP (SHA1_F2o, d, e, a, b, c, w9);
-    wa = rotl32 ((w7 ^ w2 ^ wc ^ wa), 1u); SHA1_STEP (SHA1_F2o, c, d, e, a, b, wa);
-    wb = rotl32 ((w8 ^ w3 ^ wd ^ wb), 1u); SHA1_STEP (SHA1_F2o, b, c, d, e, a, wb);
+    w8_t = rotl32 ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP (SHA1_F2o, a, b, c, d, e, w8_t);
+    w9_t = rotl32 ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP (SHA1_F2o, e, a, b, c, d, w9_t);
+    wa_t = rotl32 ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP (SHA1_F2o, d, e, a, b, c, wa_t);
+    wb_t = rotl32 ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP (SHA1_F2o, c, d, e, a, b, wb_t);
+    wc_t = rotl32 ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP (SHA1_F2o, b, c, d, e, a, wc_t);
+    wd_t = rotl32 ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP (SHA1_F2o, a, b, c, d, e, wd_t);
+    we_t = rotl32 ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP (SHA1_F2o, e, a, b, c, d, we_t);
+    wf_t = rotl32 ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP (SHA1_F2o, d, e, a, b, c, wf_t);
+    w0_t = rotl32 ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP (SHA1_F2o, c, d, e, a, b, w0_t);
+    w1_t = rotl32 ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP (SHA1_F2o, b, c, d, e, a, w1_t);
+    w2_t = rotl32 ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP (SHA1_F2o, a, b, c, d, e, w2_t);
+    w3_t = rotl32 ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP (SHA1_F2o, e, a, b, c, d, w3_t);
+    w4_t = rotl32 ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP (SHA1_F2o, d, e, a, b, c, w4_t);
+    w5_t = rotl32 ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP (SHA1_F2o, c, d, e, a, b, w5_t);
+    w6_t = rotl32 ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP (SHA1_F2o, b, c, d, e, a, w6_t);
+    w7_t = rotl32 ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP (SHA1_F2o, a, b, c, d, e, w7_t);
+    w8_t = rotl32 ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP (SHA1_F2o, e, a, b, c, d, w8_t);
+    w9_t = rotl32 ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP (SHA1_F2o, d, e, a, b, c, w9_t);
+    wa_t = rotl32 ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP (SHA1_F2o, c, d, e, a, b, wa_t);
+    wb_t = rotl32 ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP (SHA1_F2o, b, c, d, e, a, wb_t);
 
     #undef K
     #define K SHA1C03
 
-    wc = rotl32 ((w9 ^ w4 ^ we ^ wc), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, wc);
-    wd = rotl32 ((wa ^ w5 ^ wf ^ wd), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, wd);
-    we = rotl32 ((wb ^ w6 ^ w0 ^ we), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, we);
-    wf = rotl32 ((wc ^ w7 ^ w1 ^ wf), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, wf);
-    w0 = rotl32 ((wd ^ w8 ^ w2 ^ w0), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, w0);
-    w1 = rotl32 ((we ^ w9 ^ w3 ^ w1), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, w1);
-    w2 = rotl32 ((wf ^ wa ^ w4 ^ w2), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, w2);
-    w3 = rotl32 ((w0 ^ wb ^ w5 ^ w3), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, w3);
-    w4 = rotl32 ((w1 ^ wc ^ w6 ^ w4), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, w4);
-    w5 = rotl32 ((w2 ^ wd ^ w7 ^ w5), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, w5);
-    w6 = rotl32 ((w3 ^ we ^ w8 ^ w6), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, w6);
-    w7 = rotl32 ((w4 ^ wf ^ w9 ^ w7), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, w7);
-    w8 = rotl32 ((w5 ^ w0 ^ wa ^ w8), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, w8);
-    w9 = rotl32 ((w6 ^ w1 ^ wb ^ w9), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, w9);
-    wa = rotl32 ((w7 ^ w2 ^ wc ^ wa), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, wa);
-    wb = rotl32 ((w8 ^ w3 ^ wd ^ wb), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, wb);
-    wc = rotl32 ((w9 ^ w4 ^ we ^ wc), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, wc);
-    wd = rotl32 ((wa ^ w5 ^ wf ^ wd), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, wd);
-    we = rotl32 ((wb ^ w6 ^ w0 ^ we), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, we);
-    wf = rotl32 ((wc ^ w7 ^ w1 ^ wf), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, wf);
-
-    const u32 r0 = d;
-    const u32 r1 = e;
-    const u32 r2 = c;
-    const u32 r3 = b;
-
-    #include COMPARE_M
+    wc_t = rotl32 ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, wc_t);
+    wd_t = rotl32 ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, wd_t);
+    we_t = rotl32 ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, we_t);
+    wf_t = rotl32 ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, wf_t);
+    w0_t = rotl32 ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, w0_t);
+    w1_t = rotl32 ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, w1_t);
+    w2_t = rotl32 ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, w2_t);
+    w3_t = rotl32 ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, w3_t);
+    w4_t = rotl32 ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, w4_t);
+    w5_t = rotl32 ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, w5_t);
+    w6_t = rotl32 ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, w6_t);
+    w7_t = rotl32 ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, w7_t);
+    w8_t = rotl32 ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, w8_t);
+    w9_t = rotl32 ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, w9_t);
+    wa_t = rotl32 ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, wa_t);
+    wb_t = rotl32 ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, wb_t);
+    wc_t = rotl32 ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, wc_t);
+    wd_t = rotl32 ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, wd_t);
+    we_t = rotl32 ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, we_t);
+    wf_t = rotl32 ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, wf_t);
+
+    COMPARE_M_SIMD (d, e, c, b);
   }
 }
 
-__kernel void m04900_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_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 m04900_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 m04900_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_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 m04900_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 m04900_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_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 rules_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
+__kernel void m04900_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
@@ -359,18 +313,16 @@ __kernel void m04900_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
   if (gid >= gid_max) return;
 
   u32 pw_buf0[4];
-
-  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];
-
   u32 pw_buf1[4];
 
-  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];
+  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];
 
   const u32 pw_len = pws[gid].pw_len;
 
@@ -379,32 +331,26 @@ __kernel void m04900_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
    */
 
   u32 salt_buf0[4];
+  u32 salt_buf1[4];
+  u32 salt_buf2[4];
+  u32 salt_buf3[4];
 
   salt_buf0[0] = salt_bufs[salt_pos].salt_buf[ 0];
   salt_buf0[1] = salt_bufs[salt_pos].salt_buf[ 1];
   salt_buf0[2] = salt_bufs[salt_pos].salt_buf[ 2];
   salt_buf0[3] = salt_bufs[salt_pos].salt_buf[ 3];
-
-  u32 salt_buf1[4];
-
   salt_buf1[0] = salt_bufs[salt_pos].salt_buf[ 4];
   salt_buf1[1] = salt_bufs[salt_pos].salt_buf[ 5];
   salt_buf1[2] = salt_bufs[salt_pos].salt_buf[ 6];
   salt_buf1[3] = salt_bufs[salt_pos].salt_buf[ 7];
-
-  u32 salt_buf2[4];
-
-  salt_buf2[0] = 0;
-  salt_buf2[1] = 0;
-  salt_buf2[2] = 0;
-  salt_buf2[3] = 0;
-
-  u32 salt_buf3[4];
-
-  salt_buf3[0] = 0;
-  salt_buf3[1] = 0;
-  salt_buf3[2] = 0;
-  salt_buf3[3] = 0;
+  salt_buf2[0] = salt_bufs[salt_pos].salt_buf[ 8];
+  salt_buf2[1] = salt_bufs[salt_pos].salt_buf[ 9];
+  salt_buf2[2] = salt_bufs[salt_pos].salt_buf[10];
+  salt_buf2[3] = salt_bufs[salt_pos].salt_buf[11];
+  salt_buf3[0] = salt_bufs[salt_pos].salt_buf[12];
+  salt_buf3[1] = salt_bufs[salt_pos].salt_buf[13];
+  salt_buf3[2] = salt_bufs[salt_pos].salt_buf[14];
+  salt_buf3[3] = salt_bufs[salt_pos].salt_buf[15];
 
   const u32 salt_len = salt_bufs[salt_pos].salt_len;
 
@@ -424,261 +370,229 @@ __kernel void m04900_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
    * reverse
    */
 
-  const u32 e_rev = rotl32 (search[1], 2u);
+  const u32 e_rev = rotl32_S (search[1], 2u);
 
   /**
    * loop
    */
 
-  for (u32 il_pos = 0; il_pos < rules_cnt; il_pos++)
+  for (u32 il_pos = 0; il_pos < il_cnt; il_pos += VECT_SIZE)
   {
-    u32 w0_t[4];
-
-    w0_t[0] = pw_buf0[0];
-    w0_t[1] = pw_buf0[1];
-    w0_t[2] = pw_buf0[2];
-    w0_t[3] = pw_buf0[3];
-
-    u32 w1_t[4];
-
-    w1_t[0] = pw_buf1[0];
-    w1_t[1] = pw_buf1[1];
-    w1_t[2] = pw_buf1[2];
-    w1_t[3] = pw_buf1[3];
-
-    u32 w2_t[4];
+    u32x w0[4] = { 0 };
+    u32x w1[4] = { 0 };
+    u32x w2[4] = { 0 };
+    u32x w3[4] = { 0 };
 
-    w2_t[0] = 0;
-    w2_t[1] = 0;
-    w2_t[2] = 0;
-    w2_t[3] = 0;
-
-    u32 w3_t[4];
-
-    w3_t[0] = 0;
-    w3_t[1] = 0;
-    w3_t[2] = 0;
-    w3_t[3] = 0;
-
-    const u32 out_len = apply_rules (rules_buf[il_pos].cmds, w0_t, w1_t, pw_len);
+    const u32x out_len = apply_rules_vect (pw_buf0, pw_buf1, pw_len, rules_buf, il_pos, w0, w1);
 
     /**
      * prepend salt
      */
 
-    switch_buffer_by_offset_le (w0_t, w1_t, w2_t, w3_t, salt_len);
-
-    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] |= salt_buf2[2];
-    w2_t[3] |= salt_buf2[3];
-    w3_t[0] |= salt_buf3[0];
-    w3_t[1] |= salt_buf3[1];
-    w3_t[2] |= salt_buf3[2];
-    w3_t[3] |= salt_buf3[3];
+    const u32x out_salt_len = out_len + salt_len;
+
+    switch_buffer_by_offset_le_VV (w0, w1, w2, w3, salt_len);
+
+    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] |= salt_buf2[2];
+    w2[3] |= salt_buf2[3];
+    w3[0] |= salt_buf3[0];
+    w3[1] |= salt_buf3[1];
+    w3[2] |= salt_buf3[2];
+    w3[3] |= salt_buf3[3];
 
     /**
      * append salt
      */
 
-    u32 s0[4];
+    u32x s0[4];
+    u32x s1[4];
+    u32x s2[4];
+    u32x s3[4];
 
     s0[0] = salt_buf0[0];
     s0[1] = salt_buf0[1];
     s0[2] = salt_buf0[2];
     s0[3] = salt_buf0[3];
-
-    u32 s1[4];
-
     s1[0] = salt_buf1[0];
     s1[1] = salt_buf1[1];
     s1[2] = salt_buf1[2];
     s1[3] = salt_buf1[3];
-
-    u32 s2[4];
-
-    s2[0] = 0;
-    s2[1] = 0;
-    s2[2] = 0;
-    s2[3] = 0;
-
-    u32 s3[4];
-
-    s3[0] = 0;
-    s3[1] = 0;
-    s3[2] = 0;
-    s3[3] = 0;
-
-    switch_buffer_by_offset_le (s0, s1, s2, s3, salt_len + out_len);
-
-    w0_t[0] |= s0[0];
-    w0_t[1] |= s0[1];
-    w0_t[2] |= s0[2];
-    w0_t[3] |= s0[3];
-    w1_t[0] |= s1[0];
-    w1_t[1] |= s1[1];
-    w1_t[2] |= s1[2];
-    w1_t[3] |= s1[3];
-    w2_t[0] |= s2[0];
-    w2_t[1] |= s2[1];
-    w2_t[2] |= s2[2];
-    w2_t[3] |= s2[3];
-    w3_t[0] |= s3[0];
-    w3_t[1] |= s3[1];
-    w3_t[2] |= s3[2];
-    w3_t[3] |= s3[3];
-
-    const u32 pw_salt_len = salt_len + out_len + salt_len;
-
-    append_0x80_4x4 (w0_t, w1_t, w2_t, w3_t, pw_salt_len);
-
-    u32 w0 = swap32 (w0_t[0]);
-    u32 w1 = swap32 (w0_t[1]);
-    u32 w2 = swap32 (w0_t[2]);
-    u32 w3 = swap32 (w0_t[3]);
-    u32 w4 = swap32 (w1_t[0]);
-    u32 w5 = swap32 (w1_t[1]);
-    u32 w6 = swap32 (w1_t[2]);
-    u32 w7 = swap32 (w1_t[3]);
-    u32 w8 = swap32 (w2_t[0]);
-    u32 w9 = swap32 (w2_t[1]);
-    u32 wa = swap32 (w2_t[2]);
-    u32 wb = swap32 (w2_t[3]);
-    u32 wc = swap32 (w3_t[0]);
-    u32 wd = swap32 (w3_t[1]);
-    u32 we = 0;
-    u32 wf = pw_salt_len * 8;
+    s2[0] = salt_buf2[0];
+    s2[1] = salt_buf2[1];
+    s2[2] = salt_buf2[2];
+    s2[3] = salt_buf2[3];
+    s3[0] = salt_buf3[0];
+    s3[1] = salt_buf3[1];
+    s3[2] = salt_buf3[2];
+    s3[3] = salt_buf3[3];
+
+    switch_buffer_by_offset_le_VV (s0, s1, s2, s3, out_salt_len);
+
+    w0[0] |= s0[0];
+    w0[1] |= s0[1];
+    w0[2] |= s0[2];
+    w0[3] |= s0[3];
+    w1[0] |= s1[0];
+    w1[1] |= s1[1];
+    w1[2] |= s1[2];
+    w1[3] |= s1[3];
+    w2[0] |= s2[0];
+    w2[1] |= s2[1];
+    w2[2] |= s2[2];
+    w2[3] |= s2[3];
+    w3[0] |= s3[0];
+    w3[1] |= s3[1];
+    w3[2] |= s3[2];
+    w3[3] |= s3[3];
+
+    const u32x salt_out_salt_len = salt_len + out_len + salt_len;
+
+    append_0x80_4x4_VV (w0, w1, w2, w3, salt_out_salt_len);
 
     /**
      * sha1
      */
 
-    u32 a = SHA1M_A;
-    u32 b = SHA1M_B;
-    u32 c = SHA1M_C;
-    u32 d = SHA1M_D;
-    u32 e = SHA1M_E;
+    u32x w0_t = swap32 (w0[0]);
+    u32x w1_t = swap32 (w0[1]);
+    u32x w2_t = swap32 (w0[2]);
+    u32x w3_t = swap32 (w0[3]);
+    u32x w4_t = swap32 (w1[0]);
+    u32x w5_t = swap32 (w1[1]);
+    u32x w6_t = swap32 (w1[2]);
+    u32x w7_t = swap32 (w1[3]);
+    u32x w8_t = swap32 (w2[0]);
+    u32x w9_t = swap32 (w2[1]);
+    u32x wa_t = swap32 (w2[2]);
+    u32x wb_t = swap32 (w2[3]);
+    u32x wc_t = swap32 (w3[0]);
+    u32x wd_t = swap32 (w3[1]);
+    u32x we_t = 0;
+    u32x wf_t = salt_out_salt_len * 8;
+
+    u32x a = SHA1M_A;
+    u32x b = SHA1M_B;
+    u32x c = SHA1M_C;
+    u32x d = SHA1M_D;
+    u32x e = SHA1M_E;
 
     #undef K
     #define K SHA1C00
 
-    SHA1_STEP (SHA1_F0o, a, b, c, d, e, w0);
-    SHA1_STEP (SHA1_F0o, e, a, b, c, d, w1);
-    SHA1_STEP (SHA1_F0o, d, e, a, b, c, w2);
-    SHA1_STEP (SHA1_F0o, c, d, e, a, b, w3);
-    SHA1_STEP (SHA1_F0o, b, c, d, e, a, w4);
-    SHA1_STEP (SHA1_F0o, a, b, c, d, e, w5);
-    SHA1_STEP (SHA1_F0o, e, a, b, c, d, w6);
-    SHA1_STEP (SHA1_F0o, d, e, a, b, c, w7);
-    SHA1_STEP (SHA1_F0o, c, d, e, a, b, w8);
-    SHA1_STEP (SHA1_F0o, b, c, d, e, a, w9);
-    SHA1_STEP (SHA1_F0o, a, b, c, d, e, wa);
-    SHA1_STEP (SHA1_F0o, e, a, b, c, d, wb);
-    SHA1_STEP (SHA1_F0o, d, e, a, b, c, wc);
-    SHA1_STEP (SHA1_F0o, c, d, e, a, b, wd);
-    SHA1_STEP (SHA1_F0o, b, c, d, e, a, we);
-    SHA1_STEP (SHA1_F0o, a, b, c, d, e, wf);
-    w0 = rotl32 ((wd ^ w8 ^ w2 ^ w0), 1u); SHA1_STEP (SHA1_F0o, e, a, b, c, d, w0);
-    w1 = rotl32 ((we ^ w9 ^ w3 ^ w1), 1u); SHA1_STEP (SHA1_F0o, d, e, a, b, c, w1);
-    w2 = rotl32 ((wf ^ wa ^ w4 ^ w2), 1u); SHA1_STEP (SHA1_F0o, c, d, e, a, b, w2);
-    w3 = rotl32 ((w0 ^ wb ^ w5 ^ w3), 1u); SHA1_STEP (SHA1_F0o, b, c, d, e, a, w3);
+    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);
+    SHA1_STEP (SHA1_F0o, c, d, e, a, b, w3_t);
+    SHA1_STEP (SHA1_F0o, b, c, d, e, a, w4_t);
+    SHA1_STEP (SHA1_F0o, a, b, c, d, e, w5_t);
+    SHA1_STEP (SHA1_F0o, e, a, b, c, d, w6_t);
+    SHA1_STEP (SHA1_F0o, d, e, a, b, c, w7_t);
+    SHA1_STEP (SHA1_F0o, c, d, e, a, b, w8_t);
+    SHA1_STEP (SHA1_F0o, b, c, d, e, a, w9_t);
+    SHA1_STEP (SHA1_F0o, a, b, c, d, e, wa_t);
+    SHA1_STEP (SHA1_F0o, e, a, b, c, d, wb_t);
+    SHA1_STEP (SHA1_F0o, d, e, a, b, c, wc_t);
+    SHA1_STEP (SHA1_F0o, c, d, e, a, b, wd_t);
+    SHA1_STEP (SHA1_F0o, b, c, d, e, a, we_t);
+    SHA1_STEP (SHA1_F0o, a, b, c, d, e, wf_t);
+    w0_t = rotl32 ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP (SHA1_F0o, e, a, b, c, d, w0_t);
+    w1_t = rotl32 ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP (SHA1_F0o, d, e, a, b, c, w1_t);
+    w2_t = rotl32 ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP (SHA1_F0o, c, d, e, a, b, w2_t);
+    w3_t = rotl32 ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP (SHA1_F0o, b, c, d, e, a, w3_t);
 
     #undef K
     #define K SHA1C01
 
-    w4 = rotl32 ((w1 ^ wc ^ w6 ^ w4), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, w4);
-    w5 = rotl32 ((w2 ^ wd ^ w7 ^ w5), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, w5);
-    w6 = rotl32 ((w3 ^ we ^ w8 ^ w6), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, w6);
-    w7 = rotl32 ((w4 ^ wf ^ w9 ^ w7), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, w7);
-    w8 = rotl32 ((w5 ^ w0 ^ wa ^ w8), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, w8);
-    w9 = rotl32 ((w6 ^ w1 ^ wb ^ w9), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, w9);
-    wa = rotl32 ((w7 ^ w2 ^ wc ^ wa), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, wa);
-    wb = rotl32 ((w8 ^ w3 ^ wd ^ wb), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, wb);
-    wc = rotl32 ((w9 ^ w4 ^ we ^ wc), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, wc);
-    wd = rotl32 ((wa ^ w5 ^ wf ^ wd), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, wd);
-    we = rotl32 ((wb ^ w6 ^ w0 ^ we), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, we);
-    wf = rotl32 ((wc ^ w7 ^ w1 ^ wf), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, wf);
-    w0 = rotl32 ((wd ^ w8 ^ w2 ^ w0), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, w0);
-    w1 = rotl32 ((we ^ w9 ^ w3 ^ w1), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, w1);
-    w2 = rotl32 ((wf ^ wa ^ w4 ^ w2), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, w2);
-    w3 = rotl32 ((w0 ^ wb ^ w5 ^ w3), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, w3);
-    w4 = rotl32 ((w1 ^ wc ^ w6 ^ w4), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, w4);
-    w5 = rotl32 ((w2 ^ wd ^ w7 ^ w5), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, w5);
-    w6 = rotl32 ((w3 ^ we ^ w8 ^ w6), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, w6);
-    w7 = rotl32 ((w4 ^ wf ^ w9 ^ w7), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, w7);
+    w4_t = rotl32 ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, w4_t);
+    w5_t = rotl32 ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, w5_t);
+    w6_t = rotl32 ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, w6_t);
+    w7_t = rotl32 ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, w7_t);
+    w8_t = rotl32 ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, w8_t);
+    w9_t = rotl32 ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, w9_t);
+    wa_t = rotl32 ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, wa_t);
+    wb_t = rotl32 ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, wb_t);
+    wc_t = rotl32 ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, wc_t);
+    wd_t = rotl32 ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, wd_t);
+    we_t = rotl32 ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, we_t);
+    wf_t = rotl32 ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, wf_t);
+    w0_t = rotl32 ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, w0_t);
+    w1_t = rotl32 ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, w1_t);
+    w2_t = rotl32 ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, w2_t);
+    w3_t = rotl32 ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, w3_t);
+    w4_t = rotl32 ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, w4_t);
+    w5_t = rotl32 ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, w5_t);
+    w6_t = rotl32 ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, w6_t);
+    w7_t = rotl32 ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, w7_t);
 
     #undef K
     #define K SHA1C02
 
-    w8 = rotl32 ((w5 ^ w0 ^ wa ^ w8), 1u); SHA1_STEP (SHA1_F2o, a, b, c, d, e, w8);
-    w9 = rotl32 ((w6 ^ w1 ^ wb ^ w9), 1u); SHA1_STEP (SHA1_F2o, e, a, b, c, d, w9);
-    wa = rotl32 ((w7 ^ w2 ^ wc ^ wa), 1u); SHA1_STEP (SHA1_F2o, d, e, a, b, c, wa);
-    wb = rotl32 ((w8 ^ w3 ^ wd ^ wb), 1u); SHA1_STEP (SHA1_F2o, c, d, e, a, b, wb);
-    wc = rotl32 ((w9 ^ w4 ^ we ^ wc), 1u); SHA1_STEP (SHA1_F2o, b, c, d, e, a, wc);
-    wd = rotl32 ((wa ^ w5 ^ wf ^ wd), 1u); SHA1_STEP (SHA1_F2o, a, b, c, d, e, wd);
-    we = rotl32 ((wb ^ w6 ^ w0 ^ we), 1u); SHA1_STEP (SHA1_F2o, e, a, b, c, d, we);
-    wf = rotl32 ((wc ^ w7 ^ w1 ^ wf), 1u); SHA1_STEP (SHA1_F2o, d, e, a, b, c, wf);
-    w0 = rotl32 ((wd ^ w8 ^ w2 ^ w0), 1u); SHA1_STEP (SHA1_F2o, c, d, e, a, b, w0);
-    w1 = rotl32 ((we ^ w9 ^ w3 ^ w1), 1u); SHA1_STEP (SHA1_F2o, b, c, d, e, a, w1);
-    w2 = rotl32 ((wf ^ wa ^ w4 ^ w2), 1u); SHA1_STEP (SHA1_F2o, a, b, c, d, e, w2);
-    w3 = rotl32 ((w0 ^ wb ^ w5 ^ w3), 1u); SHA1_STEP (SHA1_F2o, e, a, b, c, d, w3);
-    w4 = rotl32 ((w1 ^ wc ^ w6 ^ w4), 1u); SHA1_STEP (SHA1_F2o, d, e, a, b, c, w4);
-    w5 = rotl32 ((w2 ^ wd ^ w7 ^ w5), 1u); SHA1_STEP (SHA1_F2o, c, d, e, a, b, w5);
-    w6 = rotl32 ((w3 ^ we ^ w8 ^ w6), 1u); SHA1_STEP (SHA1_F2o, b, c, d, e, a, w6);
-    w7 = rotl32 ((w4 ^ wf ^ w9 ^ w7), 1u); SHA1_STEP (SHA1_F2o, a, b, c, d, e, w7);
-    w8 = rotl32 ((w5 ^ w0 ^ wa ^ w8), 1u); SHA1_STEP (SHA1_F2o, e, a, b, c, d, w8);
-    w9 = rotl32 ((w6 ^ w1 ^ wb ^ w9), 1u); SHA1_STEP (SHA1_F2o, d, e, a, b, c, w9);
-    wa = rotl32 ((w7 ^ w2 ^ wc ^ wa), 1u); SHA1_STEP (SHA1_F2o, c, d, e, a, b, wa);
-    wb = rotl32 ((w8 ^ w3 ^ wd ^ wb), 1u); SHA1_STEP (SHA1_F2o, b, c, d, e, a, wb);
+    w8_t = rotl32 ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP (SHA1_F2o, a, b, c, d, e, w8_t);
+    w9_t = rotl32 ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP (SHA1_F2o, e, a, b, c, d, w9_t);
+    wa_t = rotl32 ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP (SHA1_F2o, d, e, a, b, c, wa_t);
+    wb_t = rotl32 ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP (SHA1_F2o, c, d, e, a, b, wb_t);
+    wc_t = rotl32 ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP (SHA1_F2o, b, c, d, e, a, wc_t);
+    wd_t = rotl32 ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP (SHA1_F2o, a, b, c, d, e, wd_t);
+    we_t = rotl32 ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP (SHA1_F2o, e, a, b, c, d, we_t);
+    wf_t = rotl32 ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP (SHA1_F2o, d, e, a, b, c, wf_t);
+    w0_t = rotl32 ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP (SHA1_F2o, c, d, e, a, b, w0_t);
+    w1_t = rotl32 ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP (SHA1_F2o, b, c, d, e, a, w1_t);
+    w2_t = rotl32 ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP (SHA1_F2o, a, b, c, d, e, w2_t);
+    w3_t = rotl32 ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP (SHA1_F2o, e, a, b, c, d, w3_t);
+    w4_t = rotl32 ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP (SHA1_F2o, d, e, a, b, c, w4_t);
+    w5_t = rotl32 ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP (SHA1_F2o, c, d, e, a, b, w5_t);
+    w6_t = rotl32 ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP (SHA1_F2o, b, c, d, e, a, w6_t);
+    w7_t = rotl32 ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP (SHA1_F2o, a, b, c, d, e, w7_t);
+    w8_t = rotl32 ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP (SHA1_F2o, e, a, b, c, d, w8_t);
+    w9_t = rotl32 ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP (SHA1_F2o, d, e, a, b, c, w9_t);
+    wa_t = rotl32 ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP (SHA1_F2o, c, d, e, a, b, wa_t);
+    wb_t = rotl32 ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP (SHA1_F2o, b, c, d, e, a, wb_t);
 
     #undef K
     #define K SHA1C03
 
-    wc = rotl32 ((w9 ^ w4 ^ we ^ wc), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, wc);
-    wd = rotl32 ((wa ^ w5 ^ wf ^ wd), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, wd);
-    we = rotl32 ((wb ^ w6 ^ w0 ^ we), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, we);
-    wf = rotl32 ((wc ^ w7 ^ w1 ^ wf), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, wf);
-    w0 = rotl32 ((wd ^ w8 ^ w2 ^ w0), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, w0);
-    w1 = rotl32 ((we ^ w9 ^ w3 ^ w1), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, w1);
-    w2 = rotl32 ((wf ^ wa ^ w4 ^ w2), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, w2);
-    w3 = rotl32 ((w0 ^ wb ^ w5 ^ w3), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, w3);
-    w4 = rotl32 ((w1 ^ wc ^ w6 ^ w4), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, w4);
-    w5 = rotl32 ((w2 ^ wd ^ w7 ^ w5), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, w5);
-    w6 = rotl32 ((w3 ^ we ^ w8 ^ w6), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, w6);
-    w7 = rotl32 ((w4 ^ wf ^ w9 ^ w7), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, w7);
-    w8 = rotl32 ((w5 ^ w0 ^ wa ^ w8), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, w8);
-    w9 = rotl32 ((w6 ^ w1 ^ wb ^ w9), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, w9);
-    wa = rotl32 ((w7 ^ w2 ^ wc ^ wa), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, wa);
-    wb = rotl32 ((w8 ^ w3 ^ wd ^ wb), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, wb);
-
-    if (allx (e != e_rev)) continue;
-
-    wc = rotl32 ((w9 ^ w4 ^ we ^ wc), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, wc);
-    wd = rotl32 ((wa ^ w5 ^ wf ^ wd), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, wd);
-    we = rotl32 ((wb ^ w6 ^ w0 ^ we), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, we);
-    wf = rotl32 ((wc ^ w7 ^ w1 ^ wf), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, wf);
-
-    const u32 r0 = d;
-    const u32 r1 = e;
-    const u32 r2 = c;
-    const u32 r3 = b;
-
-    #include COMPARE_S
+    wc_t = rotl32 ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, wc_t);
+    wd_t = rotl32 ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, wd_t);
+    we_t = rotl32 ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, we_t);
+    wf_t = rotl32 ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, wf_t);
+    w0_t = rotl32 ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, w0_t);
+    w1_t = rotl32 ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, w1_t);
+    w2_t = rotl32 ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, w2_t);
+    w3_t = rotl32 ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, w3_t);
+    w4_t = rotl32 ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, w4_t);
+    w5_t = rotl32 ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, w5_t);
+    w6_t = rotl32 ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, w6_t);
+    w7_t = rotl32 ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, w7_t);
+    w8_t = rotl32 ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, w8_t);
+    w9_t = rotl32 ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, w9_t);
+    wa_t = rotl32 ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, wa_t);
+    wb_t = rotl32 ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, wb_t);
+
+    if (MATCHES_NONE_VS (e, e_rev)) continue;
+
+    wc_t = rotl32 ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, wc_t);
+    wd_t = rotl32 ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, wd_t);
+    we_t = rotl32 ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, we_t);
+    wf_t = rotl32 ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, wf_t);
+
+    COMPARE_S_SIMD (d, e, c, b);
   }
 }
 
-__kernel void m04900_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_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 m04900_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 m04900_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_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 m04900_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)
 {
 }