X-Git-Url: https://www.flypig.org.uk/git/?a=blobdiff_plain;f=OpenCL%2Fm10420_a0.cl;h=af512d137c93464107cf90bb8d91ba373bfedee5;hb=8702d0e3e155e6a61ca1b0dcdc1ec924d92b8436;hp=65f17c5e92cd1a2e2dd1e664a982e27b941fd820;hpb=0bf4e3c34a6a799ccc34f403bed70119574ca9c8;p=hashcat.git

diff --git a/OpenCL/m10420_a0.cl b/OpenCL/m10420_a0.cl
index 65f17c5..af512d1 100644
--- a/OpenCL/m10420_a0.cl
+++ b/OpenCL/m10420_a0.cl
@@ -5,22 +5,16 @@
 
 #define _MD5_
 
-#include "include/constants.h"
-#include "include/kernel_vendor.h"
+#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 "types_ocl.c"
-#include "common.c"
-#include "include/rp_gpu.h"
-#include "rp.c"
-
-#define COMPARE_S "check_single_comp4.c"
-#define COMPARE_M "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] =
 {
@@ -34,29 +28,29 @@ __constant u32 padding[8] =
   0x7a695364
 };
 
-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 u32x w0[4], const u32x w1[4], const u32x w2[4], const u32x w3[4], u32x digest[4])
 {
-  u32 a = digest[0];
-  u32 b = digest[1];
-  u32 c = digest[2];
-  u32 d = digest[3];
-
-  u32 w0_t = w0[0];
-  u32 w1_t = w0[1];
-  u32 w2_t = w0[2];
-  u32 w3_t = w0[3];
-  u32 w4_t = w1[0];
-  u32 w5_t = w1[1];
-  u32 w6_t = w1[2];
-  u32 w7_t = w1[3];
-  u32 w8_t = w2[0];
-  u32 w9_t = w2[1];
-  u32 wa_t = w2[2];
-  u32 wb_t = w2[3];
-  u32 wc_t = w3[0];
-  u32 wd_t = w3[1];
-  u32 we_t = w3[2];
-  u32 wf_t = w3[3];
+  u32x a = digest[0];
+  u32x b = digest[1];
+  u32x c = digest[2];
+  u32x d = digest[3];
+
+  u32x w0_t = w0[0];
+  u32x w1_t = w0[1];
+  u32x w2_t = w0[2];
+  u32x w3_t = w0[3];
+  u32x w4_t = w1[0];
+  u32x w5_t = w1[1];
+  u32x w6_t = w1[2];
+  u32x w7_t = w1[3];
+  u32x w8_t = w2[0];
+  u32x w9_t = w2[1];
+  u32x wa_t = w2[2];
+  u32x wb_t = w2[3];
+  u32x wc_t = w3[0];
+  u32x wd_t = w3[1];
+  u32x we_t = w3[2];
+  u32x wf_t = w3[3];
 
   MD5_STEP (MD5_Fo, a, b, c, d, w0_t, MD5C00, MD5S00);
   MD5_STEP (MD5_Fo, d, a, b, c, w1_t, MD5C01, MD5S01);
@@ -132,7 +126,7 @@ 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))) m10420_m04 (__global pw_t *pws, __global gpu_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global void *tmps, __global void *hooks, __global u32 *bitmaps_buf_s1_a, __global u32 *bitmaps_buf_s1_b, __global u32 *bitmaps_buf_s1_c, __global u32 *bitmaps_buf_s1_d, __global u32 *bitmaps_buf_s2_a, __global u32 *bitmaps_buf_s2_b, __global u32 *bitmaps_buf_s2_c, __global u32 *bitmaps_buf_s2_d, __global plain_t *plains_buf, __global digest_t *digests_buf, __global u32 *hashes_shown, __global salt_t *salt_bufs, __global 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 m10420_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
@@ -149,14 +143,12 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m10420_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];
@@ -192,134 +184,110 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m10420_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];
+    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[0] = pw_buf0[0];
-    w0[1] = pw_buf0[1];
-    w0[2] = pw_buf0[2];
-    w0[3] = pw_buf0[3];
+    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 w1[4];
+    u32 digest[4];
 
-    w1[0] = pw_buf1[0];
-    w1[1] = pw_buf1[1];
-    w1[2] = pw_buf1[2];
-    w1[3] = pw_buf1[3];
+    digest[0] = MD5M_A;
+    digest[1] = MD5M_B;
+    digest[2] = MD5M_C;
+    digest[3] = MD5M_D;
 
-    u32 w2[4];
+    md5_transform (w0, w1, w2, w3, digest);
 
+    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;
-
-    u32 w3[4];
-
     w3[0] = 0;
     w3[1] = 0;
-    w3[2] = 0;
+    w3[2] = 84 * 8;
     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);
+    md5_transform (w0, w1, w2, w3, digest);
 
-    // add password
-    // truncate at 32 is wanted, not a bug!
-    // add o_buf
+    u32x a = digest[0];
+    u32x b = digest[1] & 0xff;
+    u32x c = 0;
+    u32x d = 0;
 
-    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];
-
-    u32 digest[4];
-
-    digest[0] = MD5M_A;
-    digest[1] = MD5M_B;
-    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);
-
-    u32 a = digest[0];
-    u32 b = digest[1] & 0xff;
-
-    const u32 r0 = a;
-    const u32 r1 = b;
-    const u32 r2 = 0;
-    const u32 r3 = 0;
-
-    #include COMPARE_M
+    COMPARE_M_SIMD (a, b, c, d);
   }
 }
 
-__kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m10420_m08 (__global pw_t *pws, __global gpu_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global void *tmps, __global void *hooks, __global u32 *bitmaps_buf_s1_a, __global u32 *bitmaps_buf_s1_b, __global u32 *bitmaps_buf_s1_c, __global u32 *bitmaps_buf_s1_d, __global u32 *bitmaps_buf_s2_a, __global u32 *bitmaps_buf_s2_b, __global u32 *bitmaps_buf_s2_c, __global u32 *bitmaps_buf_s2_d, __global plain_t *plains_buf, __global digest_t *digests_buf, __global u32 *hashes_shown, __global salt_t *salt_bufs, __global 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 m10420_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 rules__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))) m10420_m16 (__global pw_t *pws, __global gpu_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global void *tmps, __global void *hooks, __global u32 *bitmaps_buf_s1_a, __global u32 *bitmaps_buf_s1_b, __global u32 *bitmaps_buf_s1_c, __global u32 *bitmaps_buf_s1_d, __global u32 *bitmaps_buf_s2_a, __global u32 *bitmaps_buf_s2_b, __global u32 *bitmaps_buf_s2_c, __global u32 *bitmaps_buf_s2_d, __global plain_t *plains_buf, __global digest_t *digests_buf, __global u32 *hashes_shown, __global salt_t *salt_bufs, __global 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 m10420_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))) m10420_s04 (__global pw_t *pws, __global gpu_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global void *tmps, __global void *hooks, __global u32 *bitmaps_buf_s1_a, __global u32 *bitmaps_buf_s1_b, __global u32 *bitmaps_buf_s1_c, __global u32 *bitmaps_buf_s1_d, __global u32 *bitmaps_buf_s2_a, __global u32 *bitmaps_buf_s2_b, __global u32 *bitmaps_buf_s2_c, __global u32 *bitmaps_buf_s2_d, __global plain_t *plains_buf, __global digest_t *digests_buf, __global u32 *hashes_shown, __global salt_t *salt_bufs, __global 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 m10420_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
@@ -336,14 +304,12 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m10420_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];
@@ -351,18 +317,6 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m10420_s04 (__glo
 
   const u32 pw_len = pws[gid].pw_len;
 
-  /**
-   * 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
    */
@@ -387,133 +341,121 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m10420_s04 (__glo
   id_buf[2] = pdf_bufs[salt_pos].id_buf[2];
   id_buf[3] = pdf_bufs[salt_pos].id_buf[3];
 
+  /**
+   * 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]
+  };
+
   /**
    * 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];
+    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);
 
-    w0[0] = pw_buf0[0];
-    w0[1] = pw_buf0[1];
-    w0[2] = pw_buf0[2];
-    w0[3] = pw_buf0[3];
+    // add password
+    // truncate at 32 is wanted, not a bug!
+    // add o_buf
 
-    u32 w1[4];
+    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];
 
-    w1[0] = pw_buf1[0];
-    w1[1] = pw_buf1[1];
-    w1[2] = pw_buf1[2];
-    w1[3] = pw_buf1[3];
+    u32 digest[4];
+
+    digest[0] = MD5M_A;
+    digest[1] = MD5M_B;
+    digest[2] = MD5M_C;
+    digest[3] = MD5M_D;
 
-    u32 w2[4];
+    md5_transform (w0, w1, w2, w3, digest);
 
+    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;
-
-    u32 w3[4];
-
     w3[0] = 0;
     w3[1] = 0;
-    w3[2] = 0;
+    w3[2] = 84 * 8;
     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);
+    md5_transform (w0, w1, w2, w3, digest);
 
-    // 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];
-
-    u32 digest[4];
-
-    digest[0] = MD5M_A;
-    digest[1] = MD5M_B;
-    digest[2] = MD5M_C;
-    digest[3] = MD5M_D;
+    u32x a = digest[0];
+    u32x b = digest[1] & 0xff;
+    u32x c = 0;
+    u32x d = 0;
 
-    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);
-
-    u32 a = digest[0];
-    u32 b = digest[1] & 0xff;
-
-    const u32 r0 = a;
-    const u32 r1 = b;
-    const u32 r2 = 0;
-    const u32 r3 = 0;
-
-    #include COMPARE_S
+    COMPARE_S_SIMD (a, b, c, d);
   }
 }
 
-__kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m10420_s08 (__global pw_t *pws, __global gpu_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global void *tmps, __global void *hooks, __global u32 *bitmaps_buf_s1_a, __global u32 *bitmaps_buf_s1_b, __global u32 *bitmaps_buf_s1_c, __global u32 *bitmaps_buf_s1_d, __global u32 *bitmaps_buf_s2_a, __global u32 *bitmaps_buf_s2_b, __global u32 *bitmaps_buf_s2_c, __global u32 *bitmaps_buf_s2_d, __global plain_t *plains_buf, __global digest_t *digests_buf, __global u32 *hashes_shown, __global salt_t *salt_bufs, __global 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 m10420_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))) m10420_s16 (__global pw_t *pws, __global gpu_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global void *tmps, __global void *hooks, __global u32 *bitmaps_buf_s1_a, __global u32 *bitmaps_buf_s1_b, __global u32 *bitmaps_buf_s1_c, __global u32 *bitmaps_buf_s1_d, __global u32 *bitmaps_buf_s2_a, __global u32 *bitmaps_buf_s2_b, __global u32 *bitmaps_buf_s2_c, __global u32 *bitmaps_buf_s2_d, __global plain_t *plains_buf, __global digest_t *digests_buf, __global u32 *hashes_shown, __global salt_t *salt_bufs, __global 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 m10420_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)
 {
 }