#define _SHA1_
+#define NEW_SIMD_CODE
+
#include "include/constants.h"
#include "include/kernel_vendor.h"
#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 m00110_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 combs_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
{
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];
+ u32 pws0[4] = { 0 };
+ u32 pws1[4] = { 0 };
- 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;
+ pws0[0] = pws[gid].i[0];
+ pws0[1] = pws[gid].i[1];
+ pws0[2] = pws[gid].i[2];
+ pws0[3] = pws[gid].i[3];
+ pws1[0] = pws[gid].i[4];
+ pws1[1] = pws[gid].i[5];
+ pws1[2] = pws[gid].i[6];
+ pws1[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];
salt_buf0[0] = salt_bufs[salt_pos].salt_buf[0];
salt_buf0[1] = salt_bufs[salt_pos].salt_buf[1];
salt_buf0[2] = salt_bufs[salt_pos].salt_buf[2];
salt_buf0[3] = salt_bufs[salt_pos].salt_buf[3];
-
- u32 salt_buf1[4];
-
salt_buf1[0] = salt_bufs[salt_pos].salt_buf[4];
salt_buf1[1] = salt_bufs[salt_pos].salt_buf[5];
salt_buf1[2] = salt_bufs[salt_pos].salt_buf[6];
* loop
*/
- for (u32 il_pos = 0; il_pos < combs_cnt; il_pos++)
+ for (u32 il_pos = 0; il_pos < combs_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];
+ const u32x pw_r_len = pwlenx_create_combt (combs_buf, il_pos);
- 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];
+ const u32x pw_len = pw_l_len + pw_r_len;
- u32 wordr1[4];
+ u32x wordr0[4] = { 0 };
+ u32x wordr1[4] = { 0 };
+ u32x wordr2[4] = { 0 };
+ u32x wordr3[4] = { 0 };
- 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];
-
- u32 wordr2[4];
-
- wordr2[0] = 0;
- wordr2[1] = 0;
- wordr2[2] = 0;
- wordr2[3] = 0;
-
- u32 wordr3[4];
-
- wordr3[0] = 0;
- wordr3[1] = 0;
- wordr3[2] = 0;
- wordr3[3] = 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)
{
s3[2] = 0;
s3[3] = 0;
- switch_buffer_by_offset_le (s0, s1, s2, s3, pw_len);
+ switch_buffer_by_offset_le_S (s0, s1, s2, s3, pw_len);
- const u32 pw_salt_len = pw_len + salt_len;
+ const u32x pw_salt_len = pw_len + salt_len;
- u32 w0[4];
+ u32x 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 w1[4];
w1[0] = wordl1[0] | wordr1[0] | s1[0];
w1[1] = wordl1[1] | wordr1[1] | s1[1];
w1[2] = wordl1[2] | wordr1[2] | s1[2];
w1[3] = wordl1[3] | wordr1[3] | s1[3];
- u32 w2[4];
+ u32x w2[4];
w2[0] = wordl2[0] | wordr2[0] | s2[0];
w2[1] = wordl2[1] | wordr2[1] | s2[1];
w2[2] = wordl2[2] | wordr2[2] | s2[2];
w2[3] = wordl2[3] | wordr2[3] | s2[3];
- u32 w3[4];
+ u32x w3[4];
w3[0] = wordl3[0] | wordr3[0] | s3[0];
w3[1] = wordl3[1] | wordr3[1] | s3[1];
* sha1
*/
- u32 w0_t = swap32 (w0[0]);
- u32 w1_t = swap32 (w0[1]);
- u32 w2_t = swap32 (w0[2]);
- u32 w3_t = swap32 (w0[3]);
- u32 w4_t = swap32 (w1[0]);
- u32 w5_t = swap32 (w1[1]);
- u32 w6_t = swap32 (w1[2]);
- u32 w7_t = swap32 (w1[3]);
- u32 w8_t = swap32 (w2[0]);
- u32 w9_t = swap32 (w2[1]);
- u32 wa_t = swap32 (w2[2]);
- u32 wb_t = swap32 (w2[3]);
- u32 wc_t = swap32 (w3[0]);
- u32 wd_t = swap32 (w3[1]);
- u32 we_t = 0;
- u32 wf_t = pw_salt_len * 8;
-
- u32 a = SHA1M_A;
- u32 b = SHA1M_B;
- u32 c = SHA1M_C;
- u32 d = SHA1M_D;
- u32 e = SHA1M_E;
+ u32x w0_t = swap32 (w0[0]);
+ u32x w1_t = swap32 (w0[1]);
+ u32x w2_t = swap32 (w0[2]);
+ u32x w3_t = swap32 (w0[3]);
+ u32x w4_t = swap32 (w1[0]);
+ u32x w5_t = swap32 (w1[1]);
+ u32x w6_t = swap32 (w1[2]);
+ u32x w7_t = swap32 (w1[3]);
+ u32x w8_t = swap32 (w2[0]);
+ u32x w9_t = swap32 (w2[1]);
+ u32x wa_t = swap32 (w2[2]);
+ u32x wb_t = swap32 (w2[3]);
+ u32x wc_t = swap32 (w3[0]);
+ u32x wd_t = swap32 (w3[1]);
+ u32x we_t = 0;
+ u32x wf_t = pw_salt_len * 8;
+
+ u32x a = SHA1M_A;
+ u32x b = SHA1M_B;
+ u32x c = SHA1M_C;
+ u32x d = SHA1M_D;
+ u32x e = SHA1M_E;
#undef K
#define K SHA1C00
we_t = rotl32 ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, we_t);
wf_t = rotl32 ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, wf_t);
-
- const u32 r0 = d;
- const u32 r1 = e;
- const u32 r2 = c;
- const u32 r3 = b;
-
- #include COMPARE_M
+ COMPARE_M_SIMD (d, e, c, b);
}
}
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 pws0[4] = { 0 };
+ u32 pws1[4] = { 0 };
- u32 wordl3[4];
-
- wordl3[0] = 0;
- wordl3[1] = 0;
- wordl3[2] = 0;
- wordl3[3] = 0;
+ pws0[0] = pws[gid].i[0];
+ pws0[1] = pws[gid].i[1];
+ pws0[2] = pws[gid].i[2];
+ pws0[3] = pws[gid].i[3];
+ pws1[0] = pws[gid].i[4];
+ pws1[1] = pws[gid].i[5];
+ pws1[2] = pws[gid].i[6];
+ pws1[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];
salt_buf0[0] = salt_bufs[salt_pos].salt_buf[0];
salt_buf0[1] = salt_bufs[salt_pos].salt_buf[1];
salt_buf0[2] = salt_bufs[salt_pos].salt_buf[2];
salt_buf0[3] = salt_bufs[salt_pos].salt_buf[3];
-
- u32 salt_buf1[4];
-
salt_buf1[0] = salt_bufs[salt_pos].salt_buf[4];
salt_buf1[1] = salt_bufs[salt_pos].salt_buf[5];
salt_buf1[2] = salt_bufs[salt_pos].salt_buf[6];
* reverse
*/
- const u32 e_rev = rotl32 (search[1], 2u);
+ const u32 e_rev = rotl32_S (search[1], 2u);
/**
* loop
*/
- for (u32 il_pos = 0; il_pos < combs_cnt; il_pos++)
+ for (u32 il_pos = 0; il_pos < combs_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];
+ const u32x pw_r_len = pwlenx_create_combt (combs_buf, il_pos);
- u32 wordr1[4];
+ const u32x pw_len = pw_l_len + pw_r_len;
- 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];
+ u32x wordr0[4] = { 0 };
+ u32x wordr1[4] = { 0 };
+ u32x wordr2[4] = { 0 };
+ u32x wordr3[4] = { 0 };
- u32 wordr2[4];
-
- wordr2[0] = 0;
- wordr2[1] = 0;
- wordr2[2] = 0;
- wordr2[3] = 0;
-
- u32 wordr3[4];
-
- wordr3[0] = 0;
- wordr3[1] = 0;
- wordr3[2] = 0;
- wordr3[3] = 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)
{
s3[2] = 0;
s3[3] = 0;
- switch_buffer_by_offset_le (s0, s1, s2, s3, pw_len);
+ switch_buffer_by_offset_le_S (s0, s1, s2, s3, pw_len);
- const u32 pw_salt_len = pw_len + salt_len;
+ const u32x pw_salt_len = pw_len + salt_len;
- u32 w0[4];
+ u32x 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 w1[4];
w1[0] = wordl1[0] | wordr1[0] | s1[0];
w1[1] = wordl1[1] | wordr1[1] | s1[1];
w1[2] = wordl1[2] | wordr1[2] | s1[2];
w1[3] = wordl1[3] | wordr1[3] | s1[3];
- u32 w2[4];
+ u32x w2[4];
w2[0] = wordl2[0] | wordr2[0] | s2[0];
w2[1] = wordl2[1] | wordr2[1] | s2[1];
w2[2] = wordl2[2] | wordr2[2] | s2[2];
w2[3] = wordl2[3] | wordr2[3] | s2[3];
- u32 w3[4];
+ u32x w3[4];
w3[0] = wordl3[0] | wordr3[0] | s3[0];
w3[1] = wordl3[1] | wordr3[1] | s3[1];
* sha1
*/
- u32 w0_t = swap32 (w0[0]);
- u32 w1_t = swap32 (w0[1]);
- u32 w2_t = swap32 (w0[2]);
- u32 w3_t = swap32 (w0[3]);
- u32 w4_t = swap32 (w1[0]);
- u32 w5_t = swap32 (w1[1]);
- u32 w6_t = swap32 (w1[2]);
- u32 w7_t = swap32 (w1[3]);
- u32 w8_t = swap32 (w2[0]);
- u32 w9_t = swap32 (w2[1]);
- u32 wa_t = swap32 (w2[2]);
- u32 wb_t = swap32 (w2[3]);
- u32 wc_t = swap32 (w3[0]);
- u32 wd_t = swap32 (w3[1]);
- u32 we_t = 0;
- u32 wf_t = pw_salt_len * 8;
-
- u32 a = SHA1M_A;
- u32 b = SHA1M_B;
- u32 c = SHA1M_C;
- u32 d = SHA1M_D;
- u32 e = SHA1M_E;
+ u32x w0_t = swap32 (w0[0]);
+ u32x w1_t = swap32 (w0[1]);
+ u32x w2_t = swap32 (w0[2]);
+ u32x w3_t = swap32 (w0[3]);
+ u32x w4_t = swap32 (w1[0]);
+ u32x w5_t = swap32 (w1[1]);
+ u32x w6_t = swap32 (w1[2]);
+ u32x w7_t = swap32 (w1[3]);
+ u32x w8_t = swap32 (w2[0]);
+ u32x w9_t = swap32 (w2[1]);
+ u32x wa_t = swap32 (w2[2]);
+ u32x wb_t = swap32 (w2[3]);
+ u32x wc_t = swap32 (w3[0]);
+ u32x wd_t = swap32 (w3[1]);
+ u32x we_t = 0;
+ u32x wf_t = pw_salt_len * 8;
+
+ u32x a = SHA1M_A;
+ u32x b = SHA1M_B;
+ u32x c = SHA1M_C;
+ u32x d = SHA1M_D;
+ u32x e = SHA1M_E;
#undef K
#define K SHA1C00
wa_t = rotl32 ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, wa_t);
wb_t = rotl32 ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, wb_t);
- if (allx (e != e_rev)) continue;
+ if (MATCHES_NONE_VS (e, e_rev)) continue;
wc_t = rotl32 ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, wc_t);
wd_t = rotl32 ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, wd_t);
we_t = rotl32 ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, we_t);
wf_t = rotl32 ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, wf_t);
-
- const u32 r0 = d;
- const u32 r1 = e;
- const u32 r2 = c;
- const u32 r3 = b;
-
- #include COMPARE_S
+ COMPARE_S_SIMD (d, e, c, b);
}
}