Move macros DGST_R0 - DGST_R3 to host, define dgst_size for opencl kernel from host...

[hashcat.git] / OpenCL / m10800_a0.cl

diff --git a/OpenCL/m10800_a0.cl b/OpenCL/m10800_a0.cl
index 370bd15..f32f934 100644 (file)
--- a/OpenCL/m10800_a0.cl
+++ b/OpenCL/m10800_a0.cl
@@ -7,20 +7,14 @@
 
 #define NEW_SIMD_CODE
 
-#include "include/constants.h"
-#include "include/kernel_vendor.h"
-
-#define DGST_R0 6
-#define DGST_R1 7
-#define DGST_R2 4
-#define DGST_R3 5
-
-#include "include/kernel_functions.c"
-#include "OpenCL/types_ocl.c"
-#include "OpenCL/common.c"
-#include "include/rp_kernel.h"
-#include "OpenCL/rp.c"
-#include "OpenCL/simd.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 u64 k_sha384[80] =
 {
@@ -46,7 +40,7 @@ __constant u64 k_sha384[80] =
   SHA384C4c, SHA384C4d, SHA384C4e, SHA384C4f,
 };
 
-static void sha384_transform (const u32x w0[4], const u32x w1[4], const u32x w2[4], const u32x w3[4], u64x digest[8])
+void sha384_transform (const u32x w0[4], const u32x w1[4], const u32x w2[4], const u32x w3[4], u64x digest[8])
 {
   u64x w0_t = hl32_to_64 (w0[0], w0[1]);
   u64x w1_t = hl32_to_64 (w0[2], w0[3]);
@@ -116,10 +110,13 @@ static void sha384_transform (const u32x w0[4], const u32x w1[4], const u32x w2[
 
   ROUND_STEP (0);
 
-  ROUND_EXPAND (); ROUND_STEP (16);
-  ROUND_EXPAND (); ROUND_STEP (32);
-  ROUND_EXPAND (); ROUND_STEP (48);
-  ROUND_EXPAND (); ROUND_STEP (64);
+  #ifdef _unroll
+  #pragma unroll
+  #endif
+  for (int i = 16; i < 80; i += 16)
+  {
+    ROUND_EXPAND (); ROUND_STEP (i);
+  }
 
   /* rev
   digest[0] += a;
@@ -159,18 +156,16 @@ __kernel void m10800_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
   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;
 
@@ -190,7 +185,7 @@ __kernel void m10800_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
     append_0x80_2x4_VV (w0, w1, out_len);
 
     /**
-     * SHA384
+     * sha512
      */
 
     u32x w0_t[4];
@@ -262,18 +257,16 @@ __kernel void m10800_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
   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;
 
@@ -305,7 +298,7 @@ __kernel void m10800_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
     append_0x80_2x4_VV (w0, w1, out_len);
 
     /**
-     * SHA384
+     * sha512
      */
 
     u32x w0_t[4];