Converted to new SIMD: -m 5100 -a 0

diff --git a/OpenCL/m05100_a0.cl b/OpenCL/m05100_a0.cl
index 124c291..41d9f2b 100644 (file)
--- a/OpenCL/m05100_a0.cl
+++ b/OpenCL/m05100_a0.cl
@@ -5,6 +5,8 @@
 
 #define _MD5H_
 
+#define NEW_SIMD_CODE
+
 #include "include/constants.h"
 #include "include/kernel_vendor.h"
 
@@ -18,9 +20,7 @@
 #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 "OpenCL/simd.c"
 
 __kernel void m05100_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)
 {
@@ -58,46 +58,23 @@ __kernel void m05100_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
    * loop
    */
 
-  for (u32 il_pos = 0; il_pos < rules_cnt; il_pos++)
+  for (u32 il_pos = 0; il_pos < rules_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];
+    u32x w0[4] = { 0 };
+    u32x w1[4] = { 0 };
+    u32x w2[4] = { 0 };
+    u32x w3[4] = { 0 };
 
-    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);
+    const u32 out_len = apply_rules_vect (pw_buf0, pw_buf1, pw_len, rules_buf, il_pos, w0, w1);
 
     append_0x80_2x4 (w0, w1, out_len);
 
     w3[2] = out_len * 8;
 
-    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);
@@ -172,32 +149,14 @@ __kernel void m05100_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
     c += MD5M_C;
     d += MD5M_D;
 
-    {
-      const u32 r0 = a;
-      const u32 r1 = b;
-      const u32 r2 = 0;
-      const u32 r3 = 0;
-
-      #include COMPARE_M
-    }
+    u32x e = 0;
+    u32x f = 0;
 
-    {
-      const u32 r0 = b;
-      const u32 r1 = c;
-      const u32 r2 = 0;
-      const u32 r3 = 0;
+    COMPARE_M_SIMD (a, b, e, f);
 
-      #include COMPARE_M
-    }
+    COMPARE_M_SIMD (b, c, e, f);
 
-    {
-      const u32 r0 = c;
-      const u32 r1 = d;
-      const u32 r2 = 0;
-      const u32 r3 = 0;
-
-      #include COMPARE_M
-    }
+    COMPARE_M_SIMD (c, d, e, f);
   }
 }
 
@@ -257,46 +216,23 @@ __kernel void m05100_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
    * loop
    */
 
-  for (u32 il_pos = 0; il_pos < rules_cnt; il_pos++)
+  for (u32 il_pos = 0; il_pos < rules_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];
+    u32x w0[4] = { 0 };
+    u32x w1[4] = { 0 };
+    u32x w2[4] = { 0 };
+    u32x w3[4] = { 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[3] = 0;
-
-    const u32 out_len = apply_rules (rules_buf[il_pos].cmds, w0, w1, pw_len);
+    const u32 out_len = apply_rules_vect (pw_buf0, pw_buf1, pw_len, rules_buf, il_pos, w0, w1);
 
     append_0x80_2x4 (w0, w1, out_len);
 
     w3[2] = out_len * 8;
 
-    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);
@@ -371,32 +307,14 @@ __kernel void m05100_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
     c += MD5M_C;
     d += MD5M_D;
 
-    {
-      const u32 r0 = a;
-      const u32 r1 = b;
-      const u32 r2 = 0;
-      const u32 r3 = 0;
-
-      #include COMPARE_S
-    }
-
-    {
-      const u32 r0 = b;
-      const u32 r1 = c;
-      const u32 r2 = 0;
-      const u32 r3 = 0;
+    u32x e = 0;
+    u32x f = 0;
 
-      #include COMPARE_S
-    }
+    COMPARE_S_SIMD (a, b, e, f);
 
-    {
-      const u32 r0 = c;
-      const u32 r1 = d;
-      const u32 r2 = 0;
-      const u32 r3 = 0;
+    COMPARE_S_SIMD (b, c, e, f);
 
-      #include COMPARE_S
-    }
+    COMPARE_S_SIMD (c, d, e, f);
   }
 }