X-Git-Url: https://www.flypig.org.uk/git/?a=blobdiff_plain;f=OpenCL%2Fm07600_a0.cl;h=8370297c93c96d8792b274d5bacb40fdf8400823;hb=161a6eb4bc643d8e636e96eda613f5137d30da59;hp=a59d583468b1b64b33fccc023bba29a2391cf7ce;hpb=331188167cc7b4f6f2a336d63daca23fbdfc199c;p=hashcat.git

diff --git a/OpenCL/m07600_a0.cl b/OpenCL/m07600_a0.cl
index a59d583..8370297 100644
--- a/OpenCL/m07600_a0.cl
+++ b/OpenCL/m07600_a0.cl
@@ -1,56 +1,79 @@
 /**
- * Author......: Jens Steube <jens.steube@gmail.com>
+ * Authors.....: Jens Steube <jens.steube@gmail.com>
+ *               magnum <john.magnum@hushmail.com>
+ *
  * License.....: MIT
  */
 
 #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"
-
-#define uint_to_hex_lower8(i) l_bin2asc[(i)]
-
-__kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m07600_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"
+
+#if   VECT_SIZE == 1
+#define uint_to_hex_lower8(i) (u32x) (l_bin2asc[(i)])
+#elif VECT_SIZE == 2
+#define uint_to_hex_lower8(i) (u32x) (l_bin2asc[(i).s0], l_bin2asc[(i).s1])
+#elif VECT_SIZE == 4
+#define uint_to_hex_lower8(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(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(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
+
+__kernel void m07600_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
+   * base
    */
 
+  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);
+  __local u32 l_bin2asc[256];
 
-  u32 pw_buf0[4];
+  for (u32 i = lid; i < 256; i += lsz)
+  {
+    const u32 i0 = (i >> 0) & 15;
+    const u32 i1 = (i >> 4) & 15;
 
-  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];
+    l_bin2asc[i] = ((i0 < 10) ? '0' + i0 : 'a' - 10 + i0) << 8
+                 | ((i1 < 10) ? '0' + i1 : 'a' - 10 + i1) << 0;
+  }
+
+  barrier (CLK_LOCAL_MEM_FENCE);
 
+  if (gid >= gid_max) return;
+
+  /**
+   * base
+   */
+
+  u32 pw_buf0[4];
   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;
 
@@ -59,136 +82,70 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m07600_m04 (__glo
    */
 
   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;
 
-  const u32 total_len = (salt_len + 40) * 8;
-
-  /**
-   * bin2asc table
-   */
-
-  __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) << 8
-                   | ((v401 < 10) ? '0' + v401 : 'a' - 10 + v401) << 0;
-  l_bin2asc[lid41] = ((v410 < 10) ? '0' + v410 : 'a' - 10 + v410) << 8
-                   | ((v411 < 10) ? '0' + v411 : 'a' - 10 + v411) << 0;
-  l_bin2asc[lid42] = ((v420 < 10) ? '0' + v420 : 'a' - 10 + v420) << 8
-                   | ((v421 < 10) ? '0' + v421 : 'a' - 10 + v421) << 0;
-  l_bin2asc[lid43] = ((v430 < 10) ? '0' + v430 : 'a' - 10 + v430) << 8
-                   | ((v431 < 10) ? '0' + v431 : 'a' - 10 + v431) << 0;
-
-  barrier (CLK_LOCAL_MEM_FENCE);
-
-  if (gid >= gid_max) return;
-
   /**
    * 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[4];
-
-    w0[0] = pw_buf0[0];
-    w0[1] = pw_buf0[1];
-    w0[2] = pw_buf0[2];
-    w0[3] = pw_buf0[3];
-
-    u32 w1[4];
+    u32x w0[4] = { 0 };
+    u32x w1[4] = { 0 };
+    u32x w2[4] = { 0 };
+    u32x w3[4] = { 0 };
 
-    w1[0] = pw_buf1[0];
-    w1[1] = pw_buf1[1];
-    w1[2] = pw_buf1[2];
-    w1[3] = pw_buf1[3];
+    const u32x out_len = apply_rules_vect (pw_buf0, pw_buf1, pw_len, rules_buf, il_pos, w0, w1);
 
-    u32 w2[4];
-
-    w2[0] = 0;
-    w2[1] = 0;
-    w2[2] = 0;
-    w2[3] = 0;
-
-    u32 w3[4];
-
-    w3[0] = 0;
-    w3[1] = 0;
-    w3[2] = 0;
-    w3[3] = 0;
-
-    const u32 out_len = apply_rules (rules_buf[il_pos].cmds, w0, w1, pw_len);
-
-    append_0x80_2x4 (w0, w1, out_len);
+    append_0x80_2x4_VV (w0, w1, out_len);
 
     /**
      * sha1
      */
 
-    u32 w0_t = swap32 (w0[0]);
-    u32 w1_t = swap32 (w0[1]);
-    u32 w2_t = swap32 (w0[2]);
-    u32 w3_t = swap32 (w0[3]);
-    u32 w4_t = swap32 (w1[0]);
-    u32 w5_t = swap32 (w1[1]);
-    u32 w6_t = swap32 (w1[2]);
-    u32 w7_t = swap32 (w1[3]);
-    u32 w8_t = swap32 (w2[0]);
-    u32 w9_t = swap32 (w2[1]);
-    u32 wa_t = swap32 (w2[2]);
-    u32 wb_t = swap32 (w2[3]);
-    u32 wc_t = swap32 (w3[0]);
-    u32 wd_t = swap32 (w3[1]);
-    u32 we_t = 0;
-    u32 wf_t = pw_len * 8;
-
-    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 = out_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
@@ -296,35 +253,6 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m07600_m04 (__glo
      * Prepend salt
      */
 
-    u32 w0t[4];
-
-    w0t[0] = uint_to_hex_lower8 ((a >> 24) & 255) <<  0
-           | uint_to_hex_lower8 ((a >> 16) & 255) << 16;
-    w0t[1] = uint_to_hex_lower8 ((a >>  8) & 255) <<  0
-           | uint_to_hex_lower8 ((a >>  0) & 255) << 16;
-    w0t[2] = uint_to_hex_lower8 ((b >> 24) & 255) <<  0
-           | uint_to_hex_lower8 ((b >> 16) & 255) << 16;
-    w0t[3] = uint_to_hex_lower8 ((b >>  8) & 255) <<  0
-           | uint_to_hex_lower8 ((b >>  0) & 255) << 16;
-
-    u32 w1t[4];
-
-    w1t[0] = uint_to_hex_lower8 ((c >> 24) & 255) <<  0
-           | uint_to_hex_lower8 ((c >> 16) & 255) << 16;
-    w1t[1] = uint_to_hex_lower8 ((c >>  8) & 255) <<  0
-           | uint_to_hex_lower8 ((c >>  0) & 255) << 16;
-    w1t[2] = uint_to_hex_lower8 ((d >> 24) & 255) <<  0
-           | uint_to_hex_lower8 ((d >> 16) & 255) << 16;
-    w1t[3] = uint_to_hex_lower8 ((d >>  8) & 255) <<  0
-           | uint_to_hex_lower8 ((d >>  0) & 255) << 16;
-
-    u32 w2t[2];
-
-    w2t[0] = uint_to_hex_lower8 ((e >> 24) & 255) <<  0
-           | uint_to_hex_lower8 ((e >> 16) & 255) << 16;
-    w2t[1] = uint_to_hex_lower8 ((e >>  8) & 255) <<  0
-           | uint_to_hex_lower8 ((e >>  0) & 255) << 16;
-
     w0_t = salt_buf0[0];
     w1_t = salt_buf0[1];
     w2_t = salt_buf0[2];
@@ -333,14 +261,24 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m07600_m04 (__glo
     w5_t = salt_buf1[1];
     w6_t = salt_buf1[2];
     w7_t = salt_buf1[3];
-    w8_t = w0t[0];
-    w9_t = w0t[1];
-    wa_t = w0t[2];
-    wb_t = w0t[3];
-    wc_t = w1t[0];
-    wd_t = w1t[1];
-    we_t = w1t[2];
-    wf_t = w1t[3];
+    w8_t = uint_to_hex_lower8 ((a >> 24) & 255) <<  0
+         | uint_to_hex_lower8 ((a >> 16) & 255) << 16;
+    w9_t = uint_to_hex_lower8 ((a >>  8) & 255) <<  0
+         | uint_to_hex_lower8 ((a >>  0) & 255) << 16;
+    wa_t = uint_to_hex_lower8 ((b >> 24) & 255) <<  0
+         | uint_to_hex_lower8 ((b >> 16) & 255) << 16;
+    wb_t = uint_to_hex_lower8 ((b >>  8) & 255) <<  0
+         | uint_to_hex_lower8 ((b >>  0) & 255) << 16;
+    wc_t = uint_to_hex_lower8 ((c >> 24) & 255) <<  0
+         | uint_to_hex_lower8 ((c >> 16) & 255) << 16;
+    wd_t = uint_to_hex_lower8 ((c >>  8) & 255) <<  0
+         | uint_to_hex_lower8 ((c >>  0) & 255) << 16;
+    we_t = uint_to_hex_lower8 ((d >> 24) & 255) <<  0
+         | uint_to_hex_lower8 ((d >> 16) & 255) << 16;
+    wf_t = uint_to_hex_lower8 ((d >>  8) & 255) <<  0
+         | uint_to_hex_lower8 ((d >>  0) & 255) << 16;
+
+    const u32x e_sav = e;
 
     /**
      * 2nd SHA1
@@ -473,16 +411,18 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m07600_m04 (__glo
     d += SHA1M_D;
     e += SHA1M_E;
 
-    u32 r_a = a;
-    u32 r_b = b;
-    u32 r_c = c;
-    u32 r_d = d;
-    u32 r_e = e;
+    u32x r_a = a;
+    u32x r_b = b;
+    u32x r_c = c;
+    u32x r_d = d;
+    u32x r_e = e;
 
     // 2nd transform
 
-    w0_t = swap32 (w2t[0]);
-    w1_t = swap32 (w2t[1]);
+    w0_t = uint_to_hex_lower8 ((e_sav >> 24) & 255) <<  0
+         | uint_to_hex_lower8 ((e_sav >> 16) & 255) << 16;
+    w1_t = uint_to_hex_lower8 ((e_sav >>  8) & 255) <<  0
+         | uint_to_hex_lower8 ((e_sav >>  0) & 255) << 16;
     w2_t = 0x80000000;
     w3_t = 0;
     w4_t = 0;
@@ -496,7 +436,10 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m07600_m04 (__glo
     wc_t = 0;
     wd_t = 0;
     we_t = 0;
-    wf_t = total_len;
+    wf_t = (salt_len + 40) * 8;
+
+    w0_t = swap32 (w0_t);
+    w1_t = swap32 (w1_t);
 
     #undef K
     #define K SHA1C00
@@ -600,50 +543,62 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m07600_m04 (__glo
     d += r_d;
     e += r_e;
 
-    const u32 r0 = d;
-    const u32 r1 = e;
-    const u32 r2 = c;
-    const u32 r3 = b;
-
-    #include COMPARE_M
+    COMPARE_M_SIMD (d, e, c, b);
   }
 }
 
-__kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m07600_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 rules_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
+__kernel void m07600_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))) m07600_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 rules_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
+__kernel void m07600_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))) m07600_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 m07600_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
+   * base
    */
 
+  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);
+  __local u32 l_bin2asc[256];
 
-  u32 pw_buf0[4];
+  for (u32 i = lid; i < 256; i += lsz)
+  {
+    const u32 i0 = (i >> 0) & 15;
+    const u32 i1 = (i >> 4) & 15;
+
+    l_bin2asc[i] = ((i0 < 10) ? '0' + i0 : 'a' - 10 + i0) << 8
+                 | ((i1 < 10) ? '0' + i1 : 'a' - 10 + i1) << 0;
+  }
+
+  barrier (CLK_LOCAL_MEM_FENCE);
+
+  if (gid >= gid_max) return;
 
-  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];
+  /**
+   * base
+   */
 
+  u32 pw_buf0[4];
   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;
 
@@ -652,72 +607,29 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m07600_s04 (__glo
    */
 
   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;
 
-  const u32 total_len = (salt_len + 40) * 8;
-
-  /**
-   * bin2asc table
-   */
-
-  __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) << 8
-                   | ((v401 < 10) ? '0' + v401 : 'a' - 10 + v401) << 0;
-  l_bin2asc[lid41] = ((v410 < 10) ? '0' + v410 : 'a' - 10 + v410) << 8
-                   | ((v411 < 10) ? '0' + v411 : 'a' - 10 + v411) << 0;
-  l_bin2asc[lid42] = ((v420 < 10) ? '0' + v420 : 'a' - 10 + v420) << 8
-                   | ((v421 < 10) ? '0' + v421 : 'a' - 10 + v421) << 0;
-  l_bin2asc[lid43] = ((v430 < 10) ? '0' + v430 : 'a' - 10 + v430) << 8
-                   | ((v431 < 10) ? '0' + v431 : 'a' - 10 + v431) << 0;
-
-  barrier (CLK_LOCAL_MEM_FENCE);
-
-  if (gid >= gid_max) return;
-
   /**
    * digest
    */
@@ -730,76 +642,47 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m07600_s04 (__glo
     digests_buf[digests_offset].digest_buf[DGST_R3]
   };
 
-  /**
-   * reverse
-   */
-
-  const u32 e_rev = rotl32 (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[4];
-
-    w0[0] = pw_buf0[0];
-    w0[1] = pw_buf0[1];
-    w0[2] = pw_buf0[2];
-    w0[3] = pw_buf0[3];
-
-    u32 w1[4];
+    u32x w0[4] = { 0 };
+    u32x w1[4] = { 0 };
+    u32x w2[4] = { 0 };
+    u32x w3[4] = { 0 };
 
-    w1[0] = pw_buf1[0];
-    w1[1] = pw_buf1[1];
-    w1[2] = pw_buf1[2];
-    w1[3] = pw_buf1[3];
+    const u32x out_len = apply_rules_vect (pw_buf0, pw_buf1, pw_len, rules_buf, il_pos, w0, w1);
 
-    u32 w2[4];
-
-    w2[0] = 0;
-    w2[1] = 0;
-    w2[2] = 0;
-    w2[3] = 0;
-
-    u32 w3[4];
-
-    w3[0] = 0;
-    w3[1] = 0;
-    w3[2] = 0;
-    w3[3] = 0;
-
-    const u32 out_len = apply_rules (rules_buf[il_pos].cmds, w0, w1, pw_len);
-
-    append_0x80_2x4 (w0, w1, out_len);
+    append_0x80_2x4_VV (w0, w1, out_len);
 
     /**
      * sha1
      */
 
-    u32 w0_t = swap32 (w0[0]);
-    u32 w1_t = swap32 (w0[1]);
-    u32 w2_t = swap32 (w0[2]);
-    u32 w3_t = swap32 (w0[3]);
-    u32 w4_t = swap32 (w1[0]);
-    u32 w5_t = swap32 (w1[1]);
-    u32 w6_t = swap32 (w1[2]);
-    u32 w7_t = swap32 (w1[3]);
-    u32 w8_t = swap32 (w2[0]);
-    u32 w9_t = swap32 (w2[1]);
-    u32 wa_t = swap32 (w2[2]);
-    u32 wb_t = swap32 (w2[3]);
-    u32 wc_t = swap32 (w3[0]);
-    u32 wd_t = swap32 (w3[1]);
-    u32 we_t = 0;
-    u32 wf_t = pw_len * 8;
-
-    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 = out_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
@@ -907,35 +790,6 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m07600_s04 (__glo
      * Prepend salt
      */
 
-    u32 w0t[4];
-
-    w0t[0] = uint_to_hex_lower8 ((a >> 24) & 255) <<  0
-           | uint_to_hex_lower8 ((a >> 16) & 255) << 16;
-    w0t[1] = uint_to_hex_lower8 ((a >>  8) & 255) <<  0
-           | uint_to_hex_lower8 ((a >>  0) & 255) << 16;
-    w0t[2] = uint_to_hex_lower8 ((b >> 24) & 255) <<  0
-           | uint_to_hex_lower8 ((b >> 16) & 255) << 16;
-    w0t[3] = uint_to_hex_lower8 ((b >>  8) & 255) <<  0
-           | uint_to_hex_lower8 ((b >>  0) & 255) << 16;
-
-    u32 w1t[4];
-
-    w1t[0] = uint_to_hex_lower8 ((c >> 24) & 255) <<  0
-           | uint_to_hex_lower8 ((c >> 16) & 255) << 16;
-    w1t[1] = uint_to_hex_lower8 ((c >>  8) & 255) <<  0
-           | uint_to_hex_lower8 ((c >>  0) & 255) << 16;
-    w1t[2] = uint_to_hex_lower8 ((d >> 24) & 255) <<  0
-           | uint_to_hex_lower8 ((d >> 16) & 255) << 16;
-    w1t[3] = uint_to_hex_lower8 ((d >>  8) & 255) <<  0
-           | uint_to_hex_lower8 ((d >>  0) & 255) << 16;
-
-    u32 w2t[2];
-
-    w2t[0] = uint_to_hex_lower8 ((e >> 24) & 255) <<  0
-           | uint_to_hex_lower8 ((e >> 16) & 255) << 16;
-    w2t[1] = uint_to_hex_lower8 ((e >>  8) & 255) <<  0
-           | uint_to_hex_lower8 ((e >>  0) & 255) << 16;
-
     w0_t = salt_buf0[0];
     w1_t = salt_buf0[1];
     w2_t = salt_buf0[2];
@@ -944,14 +798,24 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m07600_s04 (__glo
     w5_t = salt_buf1[1];
     w6_t = salt_buf1[2];
     w7_t = salt_buf1[3];
-    w8_t = w0t[0];
-    w9_t = w0t[1];
-    wa_t = w0t[2];
-    wb_t = w0t[3];
-    wc_t = w1t[0];
-    wd_t = w1t[1];
-    we_t = w1t[2];
-    wf_t = w1t[3];
+    w8_t = uint_to_hex_lower8 ((a >> 24) & 255) <<  0
+         | uint_to_hex_lower8 ((a >> 16) & 255) << 16;
+    w9_t = uint_to_hex_lower8 ((a >>  8) & 255) <<  0
+         | uint_to_hex_lower8 ((a >>  0) & 255) << 16;
+    wa_t = uint_to_hex_lower8 ((b >> 24) & 255) <<  0
+         | uint_to_hex_lower8 ((b >> 16) & 255) << 16;
+    wb_t = uint_to_hex_lower8 ((b >>  8) & 255) <<  0
+         | uint_to_hex_lower8 ((b >>  0) & 255) << 16;
+    wc_t = uint_to_hex_lower8 ((c >> 24) & 255) <<  0
+         | uint_to_hex_lower8 ((c >> 16) & 255) << 16;
+    wd_t = uint_to_hex_lower8 ((c >>  8) & 255) <<  0
+         | uint_to_hex_lower8 ((c >>  0) & 255) << 16;
+    we_t = uint_to_hex_lower8 ((d >> 24) & 255) <<  0
+         | uint_to_hex_lower8 ((d >> 16) & 255) << 16;
+    wf_t = uint_to_hex_lower8 ((d >>  8) & 255) <<  0
+         | uint_to_hex_lower8 ((d >>  0) & 255) << 16;
+
+    const u32x e_sav = e;
 
     /**
      * 2nd SHA1
@@ -1084,16 +948,18 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m07600_s04 (__glo
     d += SHA1M_D;
     e += SHA1M_E;
 
-    u32 r_a = a;
-    u32 r_b = b;
-    u32 r_c = c;
-    u32 r_d = d;
-    u32 r_e = e;
+    u32x r_a = a;
+    u32x r_b = b;
+    u32x r_c = c;
+    u32x r_d = d;
+    u32x r_e = e;
 
     // 2nd transform
 
-    w0_t = swap32 (w2t[0]);
-    w1_t = swap32 (w2t[1]);
+    w0_t = uint_to_hex_lower8 ((e_sav >> 24) & 255) <<  0
+         | uint_to_hex_lower8 ((e_sav >> 16) & 255) << 16;
+    w1_t = uint_to_hex_lower8 ((e_sav >>  8) & 255) <<  0
+         | uint_to_hex_lower8 ((e_sav >>  0) & 255) << 16;
     w2_t = 0x80000000;
     w3_t = 0;
     w4_t = 0;
@@ -1107,7 +973,10 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m07600_s04 (__glo
     wc_t = 0;
     wd_t = 0;
     we_t = 0;
-    wf_t = total_len;
+    wf_t = (salt_len + 40) * 8;
+
+    w0_t = swap32 (w0_t);
+    w1_t = swap32 (w1_t);
 
     #undef K
     #define K SHA1C00
@@ -1211,19 +1080,14 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m07600_s04 (__glo
     d += r_d;
     e += r_e;
 
-    const u32 r0 = d;
-    const u32 r1 = e;
-    const u32 r2 = c;
-    const u32 r3 = b;
-
-    #include COMPARE_S
+    COMPARE_S_SIMD (d, e, c, b);
   }
 }
 
-__kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m07600_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 rules_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
+__kernel void m07600_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))) m07600_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 rules_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
+__kernel void m07600_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)
 {
 }