X-Git-Url: https://www.flypig.org.uk/git/?a=blobdiff_plain;f=OpenCL%2Fm10400_a0.cl;h=9f148efb7b5637d3c2084d234394b25e0338b52d;hb=161a6eb4bc643d8e636e96eda613f5137d30da59;hp=e5ac03d1945a7f4dbb0f558279f0ac899dcef0c0;hpb=331188167cc7b4f6f2a336d63daca23fbdfc199c;p=hashcat.git

diff --git a/OpenCL/m10400_a0.cl b/OpenCL/m10400_a0.cl
index e5ac03d..9f148ef 100644
--- a/OpenCL/m10400_a0.cl
+++ b/OpenCL/m10400_a0.cl
@@ -1,26 +1,23 @@
 /**
- * Author......: Jens Steube <jens.steube@gmail.com>
+ * Authors.....: Jens Steube <jens.steube@gmail.com>
+ *               Gabriele Gristina <matrix@hashcat.net>
+ *
  * License.....: MIT
  */
 
 #define _MD5_
 
-#include "include/constants.h"
-#include "include/kernel_vendor.h"
+//too much register pressure
+//#define NEW_SIMD_CODE
 
-#define DGST_R0 0
-#define DGST_R1 1
-#define DGST_R2 2
-#define DGST_R3 3
-
-#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"
+#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"
 
 __constant u32 padding[8] =
 {
@@ -42,7 +39,7 @@ typedef struct
 
 } RC4_KEY;
 
-static void swap (__local RC4_KEY *rc4_key, const u8 i, const u8 j)
+void swap (__local RC4_KEY *rc4_key, const u8 i, const u8 j)
 {
   u8 tmp;
 
@@ -51,14 +48,16 @@ static void swap (__local RC4_KEY *rc4_key, const u8 i, const u8 j)
   rc4_key->S[j] = tmp;
 }
 
-static void rc4_init_16 (__local RC4_KEY *rc4_key, const u32 data[4])
+void rc4_init_16 (__local RC4_KEY *rc4_key, const u32 data[4])
 {
   u32 v = 0x03020100;
   u32 a = 0x04040404;
 
   __local u32 *ptr = (__local u32 *) rc4_key->S;
 
+  #ifdef _unroll
   #pragma unroll
+  #endif
   for (u32 i = 0; i < 64; i++)
   {
     ptr[i] = v; v += a;
@@ -72,7 +71,9 @@ static void rc4_init_16 (__local RC4_KEY *rc4_key, const u32 data[4])
 
   u32 j = 0;
 
+  #ifdef _unroll
   #pragma unroll
+  #endif
   for (u32 i = 0; i < 255; i += 5)
   {
     j += rc4_key->S[i + 0] + d0; swap (rc4_key, i + 0, j);
@@ -85,9 +86,11 @@ static void rc4_init_16 (__local RC4_KEY *rc4_key, const u32 data[4])
   j += rc4_key->S[255] + d0; swap (rc4_key, 255, j);
 }
 
-static u8 rc4_next_16 (__local RC4_KEY *rc4_key, u8 i, u8 j, __constant u32 in[4], u32 out[4])
+u8 rc4_next_16 (__local RC4_KEY *rc4_key, u8 i, u8 j, __constant u32 *in, u32 out[4])
 {
-  #pragma unroll 4
+  #ifdef _unroll
+  #pragma unroll
+  #endif
   for (u32 k = 0; k < 4; k++)
   {
     u32 xor4 = 0;
@@ -136,7 +139,7 @@ static u8 rc4_next_16 (__local RC4_KEY *rc4_key, u8 i, u8 j, __constant u32 in[4
   return j;
 }
 
-static void md5_transform (const u32 w0[4], const u32 w1[4], const u32 w2[4], const u32 w3[4], u32 digest[4])
+void md5_transform (const u32 w0[4], const u32 w1[4], const u32 w2[4], const u32 w3[4], u32 digest[4])
 {
   u32 a = digest[0];
   u32 b = digest[1];
@@ -234,7 +237,7 @@ static void md5_transform (const u32 w0[4], const u32 w1[4], const u32 w2[4], co
   digest[3] += d;
 }
 
-__kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m10400_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 pdf_t *pdf_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 m10400_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 pdf_t *pdf_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
@@ -251,14 +254,12 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m10400_m04 (__glo
   if (gid >= gid_max) return;
 
   u32 pw_buf0[4];
+  u32 pw_buf1[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];
@@ -267,7 +268,7 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m10400_m04 (__glo
   const u32 pw_len = pws[gid].pw_len;
 
   /**
-   * key
+   * shared
    */
 
   __local RC4_KEY rc4_keys[64];
@@ -302,84 +303,63 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m10400_m04 (__glo
    * 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];
-
-    w1[0] = pw_buf1[0];
-    w1[1] = pw_buf1[1];
-    w1[2] = pw_buf1[2];
-    w1[3] = pw_buf1[3];
-
-    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);
-
-    u32 w0_t[4];
-    u32 w1_t[4];
-    u32 w2_t[4];
-    u32 w3_t[4];
-
-    // max length supported by pdf11 is 32
-
-    w0_t[0] = padding[0];
-    w0_t[1] = padding[1];
-    w0_t[2] = padding[2];
-    w0_t[3] = padding[3];
-    w1_t[0] = padding[4];
-    w1_t[1] = padding[5];
-    w1_t[2] = padding[6];
-    w1_t[3] = padding[7];
-    w2_t[0] = 0;
-    w2_t[1] = 0;
-    w2_t[2] = 0;
-    w2_t[3] = 0;
-    w3_t[0] = 0;
-    w3_t[1] = 0;
-    w3_t[2] = 0;
-    w3_t[3] = 0;
-
-    switch_buffer_by_offset (w0_t, w1_t, w2_t, w3_t, pw_len);
+    u32x w0[4] = { 0 };
+    u32x w1[4] = { 0 };
+    u32x w2[4] = { 0 };
+    u32x w3[4] = { 0 };
+
+    const u32x out_len = apply_rules_vect (pw_buf0, pw_buf1, pw_len, rules_buf, il_pos, w0, w1);
+
+    /**
+     * pdf
+     */
+
+    u32 p0[4];
+    u32 p1[4];
+    u32 p2[4];
+    u32 p3[4];
+
+    p0[0] = padding[0];
+    p0[1] = padding[1];
+    p0[2] = padding[2];
+    p0[3] = padding[3];
+    p1[0] = padding[4];
+    p1[1] = padding[5];
+    p1[2] = padding[6];
+    p1[3] = padding[7];
+    p2[0] = 0;
+    p2[1] = 0;
+    p2[2] = 0;
+    p2[3] = 0;
+    p3[0] = 0;
+    p3[1] = 0;
+    p3[2] = 0;
+    p3[3] = 0;
+
+    switch_buffer_by_offset_le (p0, p1, p2, p3, out_len);
 
     // add password
     // truncate at 32 is wanted, not a bug!
     // add o_buf
 
-    w0_t[0] |= w0[0];
-    w0_t[1] |= w0[1];
-    w0_t[2] |= w0[2];
-    w0_t[3] |= w0[3];
-    w1_t[0] |= w1[0];
-    w1_t[1] |= w1[1];
-    w1_t[2] |= w1[2];
-    w1_t[3] |= w1[3];
-    w2_t[0]  = o_buf[0];
-    w2_t[1]  = o_buf[1];
-    w2_t[2]  = o_buf[2];
-    w2_t[3]  = o_buf[3];
-    w3_t[0]  = o_buf[4];
-    w3_t[1]  = o_buf[5];
-    w3_t[2]  = o_buf[6];
-    w3_t[3]  = o_buf[7];
+    w0[0] |= p0[0];
+    w0[1] |= p0[1];
+    w0[2] |= p0[2];
+    w0[3] |= p0[3];
+    w1[0] |= p1[0];
+    w1[1] |= p1[1];
+    w1[2] |= p1[2];
+    w1[3] |= p1[3];
+    w2[0]  = o_buf[0];
+    w2[1]  = o_buf[1];
+    w2[2]  = o_buf[2];
+    w2[3]  = o_buf[3];
+    w3[0]  = o_buf[4];
+    w3[1]  = o_buf[5];
+    w3[2]  = o_buf[6];
+    w3[3]  = o_buf[7];
 
     u32 digest[4];
 
@@ -388,60 +368,52 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m10400_m04 (__glo
     digest[2] = MD5M_C;
     digest[3] = MD5M_D;
 
-    md5_transform (w0_t, w1_t, w2_t, w3_t, digest);
-
-    w0_t[0] = P;
-    w0_t[1] = id_buf[0];
-    w0_t[2] = id_buf[1];
-    w0_t[3] = id_buf[2];
-    w1_t[0] = id_buf[3];
-    w1_t[1] = 0x80;
-    w1_t[2] = 0;
-    w1_t[3] = 0;
-    w2_t[0] = 0;
-    w2_t[1] = 0;
-    w2_t[2] = 0;
-    w2_t[3] = 0;
-    w3_t[0] = 0;
-    w3_t[1] = 0;
-    w3_t[2] = 84 * 8;
-    w3_t[3] = 0;
-
-    md5_transform (w0_t, w1_t, w2_t, w3_t, digest);
+    md5_transform (w0, w1, w2, w3, digest);
 
-    // now the RC4 part
+    w0[0] = P;
+    w0[1] = id_buf[0];
+    w0[2] = id_buf[1];
+    w0[3] = id_buf[2];
+    w1[0] = id_buf[3];
+    w1[1] = 0x80;
+    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] = 84 * 8;
+    w3[3] = 0;
 
-    u32 key[4];
+    md5_transform (w0, w1, w2, w3, digest);
 
-    key[0] = digest[0];
-    key[1] = digest[1] & 0xff;
-    key[2] = 0;
-    key[3] = 0;
+    // now the RC4 part
 
-    rc4_init_16 (rc4_key, key);
+    digest[1] = digest[1] & 0xff;
+    digest[2] = 0;
+    digest[3] = 0;
+
+    rc4_init_16 (rc4_key, digest);
 
     u32 out[4];
 
     rc4_next_16 (rc4_key, 0, 0, padding, out);
 
-    const u32 r0 = out[0];
-    const u32 r1 = out[1];
-    const u32 r2 = out[2];
-    const u32 r3 = out[3];
-
-    #include COMPARE_M
+    COMPARE_M_SIMD (out[0], out[1], out[2], out[3]);
   }
 }
 
-__kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m10400_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 pdf_t *pdf_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 m10400_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 pdf_t *pdf_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))) m10400_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 pdf_t *pdf_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 m10400_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 pdf_t *pdf_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))) m10400_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 pdf_t *pdf_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 m10400_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 pdf_t *pdf_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
@@ -458,14 +430,12 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m10400_s04 (__glo
   if (gid >= gid_max) return;
 
   u32 pw_buf0[4];
+  u32 pw_buf1[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];
@@ -474,25 +444,13 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m10400_s04 (__glo
   const u32 pw_len = pws[gid].pw_len;
 
   /**
-   * key
+   * shared
    */
 
   __local RC4_KEY rc4_keys[64];
 
   __local RC4_KEY *rc4_key = &rc4_keys[lid];
 
-  /**
-   * digest
-   */
-
-  const u32 search[4] =
-  {
-    digests_buf[digests_offset].digest_buf[DGST_R0],
-    digests_buf[digests_offset].digest_buf[DGST_R1],
-    digests_buf[digests_offset].digest_buf[DGST_R2],
-    digests_buf[digests_offset].digest_buf[DGST_R3]
-  };
-
   /**
    * U_buf
    */
@@ -518,87 +476,78 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m10400_s04 (__glo
   id_buf[3] = pdf_bufs[salt_pos].id_buf[3];
 
   /**
-   * loop
+   * digest
    */
 
-  for (u32 il_pos = 0; il_pos < rules_cnt; il_pos++)
+  const u32 search[4] =
   {
-    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];
-
-    w1[0] = pw_buf1[0];
-    w1[1] = pw_buf1[1];
-    w1[2] = pw_buf1[2];
-    w1[3] = pw_buf1[3];
-
-    u32 w2[4];
-
-    w2[0] = 0;
-    w2[1] = 0;
-    w2[2] = 0;
-    w2[3] = 0;
-
-    u32 w3[4];
+    digests_buf[digests_offset].digest_buf[DGST_R0],
+    digests_buf[digests_offset].digest_buf[DGST_R1],
+    digests_buf[digests_offset].digest_buf[DGST_R2],
+    digests_buf[digests_offset].digest_buf[DGST_R3]
+  };
 
-    w3[0] = 0;
-    w3[1] = 0;
-    w3[2] = 0;
-    w3[3] = 0;
+  /**
+   * loop
+   */
 
-    const u32 out_len = apply_rules (rules_buf[il_pos].cmds, w0, w1, pw_len);
-
-    u32 w0_t[4];
-    u32 w1_t[4];
-    u32 w2_t[4];
-    u32 w3_t[4];
-
-    // max length supported by pdf11 is 32
-
-    w0_t[0] = padding[0];
-    w0_t[1] = padding[1];
-    w0_t[2] = padding[2];
-    w0_t[3] = padding[3];
-    w1_t[0] = padding[4];
-    w1_t[1] = padding[5];
-    w1_t[2] = padding[6];
-    w1_t[3] = padding[7];
-    w2_t[0] = 0;
-    w2_t[1] = 0;
-    w2_t[2] = 0;
-    w2_t[3] = 0;
-    w3_t[0] = 0;
-    w3_t[1] = 0;
-    w3_t[2] = 0;
-    w3_t[3] = 0;
-
-    switch_buffer_by_offset (w0_t, w1_t, w2_t, w3_t, pw_len);
+  for (u32 il_pos = 0; il_pos < il_cnt; il_pos += VECT_SIZE)
+  {
+    u32x w0[4] = { 0 };
+    u32x w1[4] = { 0 };
+    u32x w2[4] = { 0 };
+    u32x w3[4] = { 0 };
+
+    const u32x out_len = apply_rules_vect (pw_buf0, pw_buf1, pw_len, rules_buf, il_pos, w0, w1);
+
+    /**
+     * pdf
+     */
+
+    u32 p0[4];
+    u32 p1[4];
+    u32 p2[4];
+    u32 p3[4];
+
+    p0[0] = padding[0];
+    p0[1] = padding[1];
+    p0[2] = padding[2];
+    p0[3] = padding[3];
+    p1[0] = padding[4];
+    p1[1] = padding[5];
+    p1[2] = padding[6];
+    p1[3] = padding[7];
+    p2[0] = 0;
+    p2[1] = 0;
+    p2[2] = 0;
+    p2[3] = 0;
+    p3[0] = 0;
+    p3[1] = 0;
+    p3[2] = 0;
+    p3[3] = 0;
+
+    switch_buffer_by_offset_le (p0, p1, p2, p3, out_len);
 
     // add password
     // truncate at 32 is wanted, not a bug!
     // add o_buf
 
-    w0_t[0] |= w0[0];
-    w0_t[1] |= w0[1];
-    w0_t[2] |= w0[2];
-    w0_t[3] |= w0[3];
-    w1_t[0] |= w1[0];
-    w1_t[1] |= w1[1];
-    w1_t[2] |= w1[2];
-    w1_t[3] |= w1[3];
-    w2_t[0]  = o_buf[0];
-    w2_t[1]  = o_buf[1];
-    w2_t[2]  = o_buf[2];
-    w2_t[3]  = o_buf[3];
-    w3_t[0]  = o_buf[4];
-    w3_t[1]  = o_buf[5];
-    w3_t[2]  = o_buf[6];
-    w3_t[3]  = o_buf[7];
+    w0[0] |= p0[0];
+    w0[1] |= p0[1];
+    w0[2] |= p0[2];
+    w0[3] |= p0[3];
+    w1[0] |= p1[0];
+    w1[1] |= p1[1];
+    w1[2] |= p1[2];
+    w1[3] |= p1[3];
+    w2[0]  = o_buf[0];
+    w2[1]  = o_buf[1];
+    w2[2]  = o_buf[2];
+    w2[3]  = o_buf[3];
+    w3[0]  = o_buf[4];
+    w3[1]  = o_buf[5];
+    w3[2]  = o_buf[6];
+    w3[3]  = o_buf[7];
 
     u32 digest[4];
 
@@ -607,55 +556,47 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m10400_s04 (__glo
     digest[2] = MD5M_C;
     digest[3] = MD5M_D;
 
-    md5_transform (w0_t, w1_t, w2_t, w3_t, digest);
-
-    w0_t[0] = P;
-    w0_t[1] = id_buf[0];
-    w0_t[2] = id_buf[1];
-    w0_t[3] = id_buf[2];
-    w1_t[0] = id_buf[3];
-    w1_t[1] = 0x80;
-    w1_t[2] = 0;
-    w1_t[3] = 0;
-    w2_t[0] = 0;
-    w2_t[1] = 0;
-    w2_t[2] = 0;
-    w2_t[3] = 0;
-    w3_t[0] = 0;
-    w3_t[1] = 0;
-    w3_t[2] = 84 * 8;
-    w3_t[3] = 0;
-
-    md5_transform (w0_t, w1_t, w2_t, w3_t, digest);
+    md5_transform (w0, w1, w2, w3, digest);
 
-    // now the RC4 part
+    w0[0] = P;
+    w0[1] = id_buf[0];
+    w0[2] = id_buf[1];
+    w0[3] = id_buf[2];
+    w1[0] = id_buf[3];
+    w1[1] = 0x80;
+    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] = 84 * 8;
+    w3[3] = 0;
 
-    u32 key[4];
+    md5_transform (w0, w1, w2, w3, digest);
 
-    key[0] = digest[0];
-    key[1] = digest[1] & 0xff;
-    key[2] = 0;
-    key[3] = 0;
+    // now the RC4 part
 
-    rc4_init_16 (rc4_key, key);
+    digest[1] = digest[1] & 0xff;
+    digest[2] = 0;
+    digest[3] = 0;
+
+    rc4_init_16 (rc4_key, digest);
 
     u32 out[4];
 
     rc4_next_16 (rc4_key, 0, 0, padding, out);
 
-    const u32 r0 = out[0];
-    const u32 r1 = out[1];
-    const u32 r2 = out[2];
-    const u32 r3 = out[3];
-
-    #include COMPARE_S
+    COMPARE_S_SIMD (out[0], out[1], out[2], out[3]);
   }
 }
 
-__kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m10400_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 pdf_t *pdf_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 m10400_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 pdf_t *pdf_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))) m10400_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 pdf_t *pdf_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 m10400_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 pdf_t *pdf_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)
 {
 }