From: Jens Steube <jens.steube@gmail.com>
Date: Sat, 12 Mar 2016 20:32:41 +0000 (+0100)
Subject: Cleanup -m 2410 kernels to latest standard
X-Git-Tag: v3.00-beta~178
X-Git-Url: https://www.flypig.org.uk/git/?a=commitdiff_plain;h=57f1a73f0d2dd7458cc56e50bf9da264f9faef35;p=hashcat.git

Cleanup -m 2410 kernels to latest standard
---

diff --git a/OpenCL/m02410_a0.cl b/OpenCL/m02410_a0.cl
index 41b40d4..bb1ce35 100644
--- a/OpenCL/m02410_a0.cl
+++ b/OpenCL/m02410_a0.cl
@@ -39,18 +39,16 @@ __kernel void m02410_m04 (__global pw_t *pws, __global kernel_rule_t *  rules_bu
   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;
 
@@ -59,13 +57,26 @@ __kernel void m02410_m04 (__global pw_t *pws, __global kernel_rule_t *  rules_bu
    */
 
   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[0] = salt_bufs[salt_pos].salt_buf[ 0];
   salt_buf0[1] = 0;
   salt_buf0[2] = 0;
   salt_buf0[3] = 0;
-
-  const u32 salt_len = (salt_bufs[salt_pos].salt_len < 4) ? salt_bufs[salt_pos].salt_len : 4;
+  salt_buf1[0] = 0;
+  salt_buf1[1] = 0;
+  salt_buf1[2] = 0;
+  salt_buf1[3] = 0;
+  salt_buf2[0] = 0;
+  salt_buf2[1] = 0;
+  salt_buf2[2] = 0;
+  salt_buf2[3] = 0;
+  salt_buf3[0] = 0;
+  salt_buf3[1] = 0;
+  salt_buf3[2] = 0;
+  salt_buf3[3] = 0;
 
   /**
    * loop
@@ -80,45 +91,31 @@ __kernel void m02410_m04 (__global pw_t *pws, __global kernel_rule_t *  rules_bu
 
     const u32x out_len = apply_rules_vect (pw_buf0, pw_buf1, pw_len, rules_buf, il_pos, w0, w1);
 
-    //not required?
-    //truncate_block_VV (w0, out_len);
-
-    w1[0] = 0;
-    w1[1] = 0;
-    w1[2] = 0;
-    w1[3] = 0;
-
     /**
      * append salt
      */
 
     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];
-
-    u32x s1[4];
-
-    s1[0] = 0;
-    s1[1] = 0;
-    s1[2] = 0;
-    s1[3] = 0;
-
-    u32x s2[4];
-
-    s2[0] = 0;
-    s2[1] = 0;
-    s2[2] = 0;
-    s2[3] = 0;
-
-    u32x s3[4];
-
-    s3[0] = 0;
-    s3[1] = 0;
-    s3[2] = 0;
-    s3[3] = 0;
+    s1[0] = salt_buf1[0];
+    s1[1] = salt_buf1[1];
+    s1[2] = salt_buf1[2];
+    s1[3] = salt_buf1[3];
+    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_len);
 
@@ -126,14 +123,35 @@ __kernel void m02410_m04 (__global pw_t *pws, __global kernel_rule_t *  rules_bu
     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 pw_salt_len = out_len + salt_len;
-
-    //not required?
-    //truncate_block_VV (w0, pw_salt_len);
+    /**
+     * md5
+     */
 
     w1[0] = 0x80;
+    w1[1] = 0;
+    w1[2] = 0;
+    w1[3] = 0;
+    w2[0] = 0;
+    w2[1] = 0;
+    w2[2] = 0;
+    w2[3] = 0;
+    w3[0] = 0;
+    w3[1] = 0;
     w3[2] = 16 * 8;
+    w3[3] = 0;
 
     u32x a = MD5M_A;
     u32x b = MD5M_B;
@@ -242,18 +260,16 @@ __kernel void m02410_s04 (__global pw_t *pws, __global kernel_rule_t *  rules_bu
   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;
 
@@ -262,13 +278,26 @@ __kernel void m02410_s04 (__global pw_t *pws, __global kernel_rule_t *  rules_bu
    */
 
   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[0] = salt_bufs[salt_pos].salt_buf[ 0];
   salt_buf0[1] = 0;
   salt_buf0[2] = 0;
   salt_buf0[3] = 0;
-
-  const u32 salt_len = (salt_bufs[salt_pos].salt_len < 4) ? salt_bufs[salt_pos].salt_len : 4;
+  salt_buf1[0] = 0;
+  salt_buf1[1] = 0;
+  salt_buf1[2] = 0;
+  salt_buf1[3] = 0;
+  salt_buf2[0] = 0;
+  salt_buf2[1] = 0;
+  salt_buf2[2] = 0;
+  salt_buf2[3] = 0;
+  salt_buf3[0] = 0;
+  salt_buf3[1] = 0;
+  salt_buf3[2] = 0;
+  salt_buf3[3] = 0;
 
   /**
    * digest
@@ -295,45 +324,31 @@ __kernel void m02410_s04 (__global pw_t *pws, __global kernel_rule_t *  rules_bu
 
     const u32x out_len = apply_rules_vect (pw_buf0, pw_buf1, pw_len, rules_buf, il_pos, w0, w1);
 
-    //not required?
-    //truncate_block_VV (w0, out_len);
-
-    w1[0] = 0;
-    w1[1] = 0;
-    w1[2] = 0;
-    w1[3] = 0;
-
     /**
      * append salt
      */
 
     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];
-
-    u32x s1[4];
-
-    s1[0] = 0;
-    s1[1] = 0;
-    s1[2] = 0;
-    s1[3] = 0;
-
-    u32x s2[4];
-
-    s2[0] = 0;
-    s2[1] = 0;
-    s2[2] = 0;
-    s2[3] = 0;
-
-    u32x s3[4];
-
-    s3[0] = 0;
-    s3[1] = 0;
-    s3[2] = 0;
-    s3[3] = 0;
+    s1[0] = salt_buf1[0];
+    s1[1] = salt_buf1[1];
+    s1[2] = salt_buf1[2];
+    s1[3] = salt_buf1[3];
+    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_len);
 
@@ -341,14 +356,35 @@ __kernel void m02410_s04 (__global pw_t *pws, __global kernel_rule_t *  rules_bu
     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 pw_salt_len = out_len + salt_len;
-
-    //not required?
-    //truncate_block_VV (w0, pw_salt_len);
+    /**
+     * md5
+     */
 
     w1[0] = 0x80;
+    w1[1] = 0;
+    w1[2] = 0;
+    w1[3] = 0;
+    w2[0] = 0;
+    w2[1] = 0;
+    w2[2] = 0;
+    w2[3] = 0;
+    w3[0] = 0;
+    w3[1] = 0;
     w3[2] = 16 * 8;
+    w3[3] = 0;
 
     u32x a = MD5M_A;
     u32x b = MD5M_B;
diff --git a/OpenCL/m02410_a1.cl b/OpenCL/m02410_a1.cl
index 7e7b137..a6b0a59 100644
--- a/OpenCL/m02410_a1.cl
+++ b/OpenCL/m02410_a1.cl
@@ -5,6 +5,8 @@
 
 #define _MD5_
 
+#define NEW_SIMD_CODE
+
 #include "include/constants.h"
 #include "include/kernel_vendor.h"
 
@@ -16,9 +18,7 @@
 #include "include/kernel_functions.c"
 #include "OpenCL/types_ocl.c"
 #include "OpenCL/common.c"
-
-#define COMPARE_S "OpenCL/check_single_comp4.c"
-#define COMPARE_M "OpenCL/check_multi_comp4.c"
+#include "OpenCL/simd.c"
 
 __kernel void m02410_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 il_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
 {
@@ -36,170 +36,185 @@ __kernel void m02410_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
 
   if (gid >= gid_max) return;
 
-  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];
+  u32 pw_buf0[4];
+  u32 pw_buf1[4];
 
-  wordl3[0] = 0;
-  wordl3[1] = 0;
-  wordl3[2] = 0;
-  wordl3[3] = 0;
+  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_l_len = pws[gid].pw_len;
 
-  if (combs_mode == COMBINATOR_MODE_BASE_RIGHT)
-  {
-    switch_buffer_by_offset_le (wordl0, wordl1, wordl2, wordl3, combs_buf[0].pw_len);
-  }
-
   /**
    * salt
    */
 
   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[0] = salt_bufs[salt_pos].salt_buf[ 0];
   salt_buf0[1] = 0;
   salt_buf0[2] = 0;
   salt_buf0[3] = 0;
-
-  const u32 salt_len = (salt_bufs[salt_pos].salt_len < 4) ? salt_bufs[salt_pos].salt_len : 4;
+  salt_buf1[0] = 0;
+  salt_buf1[1] = 0;
+  salt_buf1[2] = 0;
+  salt_buf1[3] = 0;
+  salt_buf2[0] = 0;
+  salt_buf2[1] = 0;
+  salt_buf2[2] = 0;
+  salt_buf2[3] = 0;
+  salt_buf3[0] = 0;
+  salt_buf3[1] = 0;
+  salt_buf3[2] = 0;
+  salt_buf3[3] = 0;
 
   /**
    * loop
    */
 
-  for (u32 il_pos = 0; il_pos < il_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];
+    const u32x pw_r_len = pwlenx_create_combt (combs_buf, il_pos);
 
-    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];
+    const u32x pw_len = pw_l_len + pw_r_len;
 
-    u32 wordr2[4];
-
-    wordr2[0] = 0;
-    wordr2[1] = 0;
-    wordr2[2] = 0;
-    wordr2[3] = 0;
-
-    u32 wordr3[4];
+    /**
+     * concat password candidate
+     */
 
-    wordr3[0] = 0;
-    wordr3[1] = 0;
-    wordr3[2] = 0;
-    wordr3[3] = 0;
+    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)
     {
-      switch_buffer_by_offset_le (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);
     }
 
-    /**
-     * append salt
-     */
-
-    u32 s0[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] = 0;
-    s1[1] = 0;
-    s1[2] = 0;
-    s1[3] = 0;
-
-    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, pw_len);
-
-    const u32 pw_salt_len = pw_len + salt_len;
-
-    u32 w0[4];
-
-    w0[0] = wordl0[0] | wordr0[0] | s0[0];
-    w0[1] = wordl0[1] | wordr0[1] | s0[1];
-    w0[2] = wordl0[2] | wordr0[2] | s0[2];
-    w0[3] = wordl0[3] | wordr0[3] | s0[3];
-
-    u32 w1[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];
     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] = 0;
-    w3[3] = 0;
+    w3[2] = wordl3[2] | wordr3[2];
+    w3[3] = wordl3[3] | wordr3[3];
+
+    /**
+     * append salt
+     */
 
-    truncate_block (w0, pw_salt_len);
+    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];
+    s1[0] = salt_buf1[0];
+    s1[1] = salt_buf1[1];
+    s1[2] = salt_buf1[2];
+    s1[3] = salt_buf1[3];
+    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, pw_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];
+
+    /**
+     * md5
+     */
 
     w1[0] = 0x80;
+    w1[1] = 0;
+    w1[2] = 0;
+    w1[3] = 0;
+    w2[0] = 0;
+    w2[1] = 0;
+    w2[2] = 0;
+    w2[3] = 0;
+    w3[0] = 0;
+    w3[1] = 0;
     w3[2] = 16 * 8;
+    w3[3] = 0;
 
-    u32 a = MD5M_A;
-    u32 b = MD5M_B;
-    u32 c = MD5M_C;
-    u32 d = MD5M_D;
+    u32x a = MD5M_A;
+    u32x b = MD5M_B;
+    u32x c = MD5M_C;
+    u32x d = MD5M_D;
 
     MD5_STEP (MD5_Fo, a, b, c, d, w0[0], MD5C00, MD5S00);
     MD5_STEP (MD5_Fo, d, a, b, c, w0[1], MD5C01, MD5S01);
@@ -274,12 +289,7 @@ __kernel void m02410_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
     c &= 0x00ffffff;
     b &= 0x00ffffff;
 
-    const u32 r0 = a;
-    const u32 r1 = d;
-    const u32 r2 = c;
-    const u32 r3 = b;
-
-    #include COMPARE_M
+    COMPARE_M_SIMD (a, d, c, b);
   }
 }
 
@@ -307,53 +317,45 @@ __kernel void m02410_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
 
   if (gid >= gid_max) return;
 
-  u32 wordl0[4];
+  u32 pw_buf0[4];
+  u32 pw_buf1[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;
+  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_l_len = pws[gid].pw_len;
 
-  if (combs_mode == COMBINATOR_MODE_BASE_RIGHT)
-  {
-    switch_buffer_by_offset_le (wordl0, wordl1, wordl2, wordl3, combs_buf[0].pw_len);
-  }
-
   /**
    * salt
    */
 
   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[0] = salt_bufs[salt_pos].salt_buf[ 0];
   salt_buf0[1] = 0;
   salt_buf0[2] = 0;
   salt_buf0[3] = 0;
-
-  const u32 salt_len = (salt_bufs[salt_pos].salt_len < 4) ? salt_bufs[salt_pos].salt_len : 4;
+  salt_buf1[0] = 0;
+  salt_buf1[1] = 0;
+  salt_buf1[2] = 0;
+  salt_buf1[3] = 0;
+  salt_buf2[0] = 0;
+  salt_buf2[1] = 0;
+  salt_buf2[2] = 0;
+  salt_buf2[3] = 0;
+  salt_buf3[0] = 0;
+  salt_buf3[1] = 0;
+  salt_buf3[2] = 0;
+  salt_buf3[3] = 0;
 
   /**
    * digest
@@ -371,118 +373,141 @@ __kernel void m02410_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
    * loop
    */
 
-  for (u32 il_pos = 0; il_pos < il_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];
+    const u32x pw_r_len = pwlenx_create_combt (combs_buf, il_pos);
 
-    u32 wordr2[4];
+    const u32x pw_len = pw_l_len + pw_r_len;
 
-    wordr2[0] = 0;
-    wordr2[1] = 0;
-    wordr2[2] = 0;
-    wordr2[3] = 0;
-
-    u32 wordr3[4];
+    /**
+     * concat password candidate
+     */
 
-    wordr3[0] = 0;
-    wordr3[1] = 0;
-    wordr3[2] = 0;
-    wordr3[3] = 0;
+    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)
     {
-      switch_buffer_by_offset_le (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);
     }
 
-    /**
-     * append salt
-     */
-
-    u32 s0[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] = 0;
-    s1[1] = 0;
-    s1[2] = 0;
-    s1[3] = 0;
-
-    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, pw_len);
-
-    const u32 pw_salt_len = pw_len + salt_len;
-
-    u32 w0[4];
-
-    w0[0] = wordl0[0] | wordr0[0] | s0[0];
-    w0[1] = wordl0[1] | wordr0[1] | s0[1];
-    w0[2] = wordl0[2] | wordr0[2] | s0[2];
-    w0[3] = wordl0[3] | wordr0[3] | s0[3];
-
-    u32 w1[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];
     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] = 0;
-    w3[3] = 0;
+    w3[2] = wordl3[2] | wordr3[2];
+    w3[3] = wordl3[3] | wordr3[3];
 
-    truncate_block (w0, pw_salt_len);
+    /**
+     * append salt
+     */
+
+    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];
+    s1[0] = salt_buf1[0];
+    s1[1] = salt_buf1[1];
+    s1[2] = salt_buf1[2];
+    s1[3] = salt_buf1[3];
+    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, pw_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];
+
+    /**
+     * md5
+     */
 
     w1[0] = 0x80;
+    w1[1] = 0;
+    w1[2] = 0;
+    w1[3] = 0;
+    w2[0] = 0;
+    w2[1] = 0;
+    w2[2] = 0;
+    w2[3] = 0;
+    w3[0] = 0;
+    w3[1] = 0;
     w3[2] = 16 * 8;
+    w3[3] = 0;
 
-    u32 a = MD5M_A;
-    u32 b = MD5M_B;
-    u32 c = MD5M_C;
-    u32 d = MD5M_D;
+    u32x a = MD5M_A;
+    u32x b = MD5M_B;
+    u32x c = MD5M_C;
+    u32x d = MD5M_D;
 
     MD5_STEP (MD5_Fo, a, b, c, d, w0[0], MD5C00, MD5S00);
     MD5_STEP (MD5_Fo, d, a, b, c, w0[1], MD5C01, MD5S01);
@@ -549,9 +574,7 @@ __kernel void m02410_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
     MD5_STEP (MD5_I , b, c, d, a, w3[1], MD5C3b, MD5S33);
     MD5_STEP (MD5_I , a, b, c, d, w1[0], MD5C3c, MD5S30);
 
-    bool q_cond = allx ((a & 0x00ffffff) != search[0]);
-
-    if (q_cond) continue;
+    if (MATCHES_NONE_VS ((a & 0x00ffffff), search[0])) continue;
 
     MD5_STEP (MD5_I , d, a, b, c, w2[3], MD5C3d, MD5S31);
     MD5_STEP (MD5_I , c, d, a, b, w0[2], MD5C3e, MD5S32);
@@ -562,12 +585,7 @@ __kernel void m02410_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
     c &= 0x00ffffff;
     b &= 0x00ffffff;
 
-    const u32 r0 = a;
-    const u32 r1 = d;
-    const u32 r2 = c;
-    const u32 r3 = b;
-
-    #include COMPARE_S
+    COMPARE_S_SIMD (a, d, c, b);
   }
 }
 
diff --git a/OpenCL/m02410_a3.cl b/OpenCL/m02410_a3.cl
index 8736dc3..c26c17a 100644
--- a/OpenCL/m02410_a3.cl
+++ b/OpenCL/m02410_a3.cl
@@ -34,28 +34,22 @@ static void m02410m (u32 w[16], const u32 pw_len, __global pw_t *pws, __global k
    */
 
   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] = 0;
   salt_buf0[2] = 0;
   salt_buf0[3] = 0;
-
-  u32 salt_buf1[4];
-
   salt_buf1[0] = 0;
   salt_buf1[1] = 0;
   salt_buf1[2] = 0;
   salt_buf1[3] = 0;
-
-  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;
@@ -63,23 +57,39 @@ static void m02410m (u32 w[16], const u32 pw_len, __global pw_t *pws, __global k
 
   switch_buffer_by_offset_le_S (salt_buf0, salt_buf1, salt_buf2, salt_buf3, pw_len);
 
-  w[0] |= salt_buf0[0];
-  w[1] |= salt_buf0[1];
-  w[2] |= salt_buf0[2];
-  w[3] |= salt_buf0[3];
-
-  const u32 salt_len = (salt_bufs[salt_pos].salt_len < 4) ? salt_bufs[salt_pos].salt_len : 4;
-
-  const u32 pw_salt_len = pw_len + salt_len;
-
-  truncate_block_S (w, pw_salt_len);
+  w[ 0] |= salt_buf0[0];
+  w[ 1] |= salt_buf0[1];
+  w[ 2] |= salt_buf0[2];
+  w[ 3] |= salt_buf0[3];
+  w[ 4] |= salt_buf1[0];
+  w[ 5] |= salt_buf1[1];
+  w[ 6] |= salt_buf1[2];
+  w[ 7] |= salt_buf1[3];
+  w[ 8] |= salt_buf2[0];
+  w[ 9] |= salt_buf2[1];
+  w[10] |= salt_buf2[2];
+  w[11] |= salt_buf2[3];
+  w[12] |= salt_buf3[0];
+  w[13] |= salt_buf3[1];
+  w[14] |= salt_buf3[2];
+  w[15] |= salt_buf3[3];
 
   /**
    * algorithm specific
    */
 
   w[ 4] = 0x80;
+  w[ 5] = 0;
+  w[ 6] = 0;
+  w[ 7] = 0;
+  w[ 8] = 0;
+  w[ 9] = 0;
+  w[10] = 0;
+  w[11] = 0;
+  w[12] = 0;
+  w[13] = 0;
   w[14] = 16 * 8;
+  w[15] = 0;
 
   /**
    * base
@@ -261,28 +271,22 @@ static void m02410s (u32 w[16], const u32 pw_len, __global pw_t *pws, __global k
    */
 
   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] = 0;
   salt_buf0[2] = 0;
   salt_buf0[3] = 0;
-
-  u32 salt_buf1[4];
-
   salt_buf1[0] = 0;
   salt_buf1[1] = 0;
   salt_buf1[2] = 0;
   salt_buf1[3] = 0;
-
-  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;
@@ -290,23 +294,39 @@ static void m02410s (u32 w[16], const u32 pw_len, __global pw_t *pws, __global k
 
   switch_buffer_by_offset_le_S (salt_buf0, salt_buf1, salt_buf2, salt_buf3, pw_len);
 
-  w[0] |= salt_buf0[0];
-  w[1] |= salt_buf0[1];
-  w[2] |= salt_buf0[2];
-  w[3] |= salt_buf0[3];
-
-  const u32 salt_len = (salt_bufs[salt_pos].salt_len < 4) ? salt_bufs[salt_pos].salt_len : 4;
-
-  const u32 pw_salt_len = pw_len + salt_len;
-
-  truncate_block_S (w, pw_salt_len);
+  w[ 0] |= salt_buf0[0];
+  w[ 1] |= salt_buf0[1];
+  w[ 2] |= salt_buf0[2];
+  w[ 3] |= salt_buf0[3];
+  w[ 4] |= salt_buf1[0];
+  w[ 5] |= salt_buf1[1];
+  w[ 6] |= salt_buf1[2];
+  w[ 7] |= salt_buf1[3];
+  w[ 8] |= salt_buf2[0];
+  w[ 9] |= salt_buf2[1];
+  w[10] |= salt_buf2[2];
+  w[11] |= salt_buf2[3];
+  w[12] |= salt_buf3[0];
+  w[13] |= salt_buf3[1];
+  w[14] |= salt_buf3[2];
+  w[15] |= salt_buf3[3];
 
   /**
    * algorithm specific
    */
 
   w[ 4] = 0x80;
+  w[ 5] = 0;
+  w[ 6] = 0;
+  w[ 7] = 0;
+  w[ 8] = 0;
+  w[ 9] = 0;
+  w[10] = 0;
+  w[11] = 0;
+  w[12] = 0;
+  w[13] = 0;
   w[14] = 16 * 8;
+  w[15] = 0;
 
   /**
    * base