github/hashcat
1
Fork 0
mirror of https://github.com/hashcat/hashcat synced 2024-11-24 14:27:14 +01:00
Code Releases Activity

Converted to new SIMD: -m 4500 -a 0

Browse Source
This commit is contained in:
Jens Steube 2016-02-06 10:07:54 +01:00
parent 6369c0b899
commit e6393454f0
1 changed files with 61 additions and 118 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

179
OpenCL/m04500_a0.cl
View File

@ -5,6 +5,8 @@
#define _SHA1_ #define _SHA1_
#define NEW_SIMD_CODE
#include "include/constants.h" #include "include/constants.h"
#include "include/kernel_vendor.h" #include "include/kernel_vendor.h"
@ -18,9 +20,7 @@
#include "OpenCL/common.c" #include "OpenCL/common.c"
#include "include/rp_kernel.h" #include "include/rp_kernel.h"
#include "OpenCL/rp.c" #include "OpenCL/rp.c"
#include "OpenCL/simd.c"
#define COMPARE_S "OpenCL/check_single_comp4.c"
#define COMPARE_M "OpenCL/check_multi_comp4.c"
#if VECT_SIZE == 1 #if VECT_SIZE == 1
#define uint_to_hex_lower8_le(i) (u32x) (l_bin2asc[(i)]) #define uint_to_hex_lower8_le(i) (u32x) (l_bin2asc[(i)])
@ -85,37 +85,14 @@ __kernel void m04500_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
* loop * 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]; u32x w0[4] = { 0 };
u32x w1[4] = { 0 };
u32x w2[4] = { 0 };
u32x w3[4] = { 0 };
w0[0] = pw_buf0[0]; const u32 out_len = apply_rules_vect (pw_buf0, pw_buf1, pw_len, rules_buf, il_pos, w0, w1);
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);
append_0x80_2x4 (w0, w1, out_len); append_0x80_2x4 (w0, w1, out_len);
@ -123,28 +100,28 @@ __kernel void m04500_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
* sha1 * sha1
*/ */
u32 w0_t = swap32 (w0[0]); u32x w0_t = swap32 (w0[0]);
u32 w1_t = swap32 (w0[1]); u32x w1_t = swap32 (w0[1]);
u32 w2_t = swap32 (w0[2]); u32x w2_t = swap32 (w0[2]);
u32 w3_t = swap32 (w0[3]); u32x w3_t = swap32 (w0[3]);
u32 w4_t = swap32 (w1[0]); u32x w4_t = swap32 (w1[0]);
u32 w5_t = swap32 (w1[1]); u32x w5_t = swap32 (w1[1]);
u32 w6_t = swap32 (w1[2]); u32x w6_t = swap32 (w1[2]);
u32 w7_t = swap32 (w1[3]); u32x w7_t = swap32 (w1[3]);
u32 w8_t = swap32 (w2[0]); u32x w8_t = swap32 (w2[0]);
u32 w9_t = swap32 (w2[1]); u32x w9_t = swap32 (w2[1]);
u32 wa_t = swap32 (w2[2]); u32x wa_t = swap32 (w2[2]);
u32 wb_t = swap32 (w2[3]); u32x wb_t = swap32 (w2[3]);
u32 wc_t = swap32 (w3[0]); u32x wc_t = swap32 (w3[0]);
u32 wd_t = swap32 (w3[1]); u32x wd_t = swap32 (w3[1]);
u32 we_t = 0; u32x we_t = 0;
u32 wf_t = out_len * 8; u32x wf_t = out_len * 8;
u32 a = SHA1M_A; u32x a = SHA1M_A;
u32 b = SHA1M_B; u32x b = SHA1M_B;
u32 c = SHA1M_C; u32x c = SHA1M_C;
u32 d = SHA1M_D; u32x d = SHA1M_D;
u32 e = SHA1M_E; u32x e = SHA1M_E;
#undef K #undef K
#define K SHA1C00 #define K SHA1C00
@ -377,18 +354,12 @@ __kernel void m04500_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
w9_t = rotl32 ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, w9_t); w9_t = rotl32 ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, w9_t);
wa_t = rotl32 ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, wa_t); 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); wb_t = rotl32 ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, wb_t);
wc_t = rotl32 ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, wc_t); 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); 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); 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); 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; COMPARE_M_SIMD (d, e, c, b);
const u32 r1 = e;
const u32 r2 = c;
const u32 r3 = b;
#include COMPARE_M
} }
} }
@ -465,43 +436,20 @@ __kernel void m04500_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
* reverse * reverse
*/ */
const u32 e_rev = rotl32 (search[1], 2u); const u32 e_rev = rotl32_S (search[1], 2u);
/** /**
* loop * 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]; u32x w0[4] = { 0 };
u32x w1[4] = { 0 };
u32x w2[4] = { 0 };
u32x w3[4] = { 0 };
w0[0] = pw_buf0[0]; const u32 out_len = apply_rules_vect (pw_buf0, pw_buf1, pw_len, rules_buf, il_pos, w0, w1);
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);
append_0x80_2x4 (w0, w1, out_len); append_0x80_2x4 (w0, w1, out_len);
@ -509,28 +457,28 @@ __kernel void m04500_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
* sha1 * sha1
*/ */
u32 w0_t = swap32 (w0[0]); u32x w0_t = swap32 (w0[0]);
u32 w1_t = swap32 (w0[1]); u32x w1_t = swap32 (w0[1]);
u32 w2_t = swap32 (w0[2]); u32x w2_t = swap32 (w0[2]);
u32 w3_t = swap32 (w0[3]); u32x w3_t = swap32 (w0[3]);
u32 w4_t = swap32 (w1[0]); u32x w4_t = swap32 (w1[0]);
u32 w5_t = swap32 (w1[1]); u32x w5_t = swap32 (w1[1]);
u32 w6_t = swap32 (w1[2]); u32x w6_t = swap32 (w1[2]);
u32 w7_t = swap32 (w1[3]); u32x w7_t = swap32 (w1[3]);
u32 w8_t = swap32 (w2[0]); u32x w8_t = swap32 (w2[0]);
u32 w9_t = swap32 (w2[1]); u32x w9_t = swap32 (w2[1]);
u32 wa_t = swap32 (w2[2]); u32x wa_t = swap32 (w2[2]);
u32 wb_t = swap32 (w2[3]); u32x wb_t = swap32 (w2[3]);
u32 wc_t = swap32 (w3[0]); u32x wc_t = swap32 (w3[0]);
u32 wd_t = swap32 (w3[1]); u32x wd_t = swap32 (w3[1]);
u32 we_t = 0; u32x we_t = 0;
u32 wf_t = out_len * 8; u32x wf_t = out_len * 8;
u32 a = SHA1M_A; u32x a = SHA1M_A;
u32 b = SHA1M_B; u32x b = SHA1M_B;
u32 c = SHA1M_C; u32x c = SHA1M_C;
u32 d = SHA1M_D; u32x d = SHA1M_D;
u32 e = SHA1M_E; u32x e = SHA1M_E;
#undef K #undef K
#define K SHA1C00 #define K SHA1C00
@ -764,19 +712,14 @@ __kernel void m04500_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
wa_t = rotl32 ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, wa_t); 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); 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); 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); 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); 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); 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; COMPARE_S_SIMD (d, e, c, b);
const u32 r1 = e;
const u32 r2 = c;
const u32 r3 = b;
#include COMPARE_S
} }
} }