1
0
mirror of https://github.com/hashcat/hashcat.git synced 2024-11-23 00:28:11 +00:00

Converted to new SIMD: -m 8400 -a 0

This commit is contained in:
Jens Steube 2016-02-06 11:01:22 +01:00
parent 1f14182585
commit 60d43ed064

View File

@ -5,6 +5,8 @@
#define _SHA1_
#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"
#if VECT_SIZE == 1
#define uint_to_hex_lower8_le(i) (u32x) (l_bin2asc[(i)])
@ -32,30 +32,30 @@
#define uint_to_hex_lower8_le(i) (u32x) (l_bin2asc[(i).s0], l_bin2asc[(i).s1], l_bin2asc[(i).s2], l_bin2asc[(i).s3], l_bin2asc[(i).s4], l_bin2asc[(i).s5], l_bin2asc[(i).s6], l_bin2asc[(i).s7])
#endif
static void sha1_transform (const u32 w0[4], const u32 w1[4], const u32 w2[4], const u32 w3[4], u32 digest[5])
static void sha1_transform (const u32x w0[4], const u32x w1[4], const u32x w2[4], const u32x w3[4], u32x digest[5])
{
u32 A = digest[0];
u32 B = digest[1];
u32 C = digest[2];
u32 D = digest[3];
u32 E = digest[4];
u32x A = digest[0];
u32x B = digest[1];
u32x C = digest[2];
u32x D = digest[3];
u32x E = digest[4];
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 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];
#undef K
#define K SHA1C00
@ -215,22 +215,22 @@ __kernel void m08400_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
u32 salt_buf0[4];
salt_buf0[0] = swap32 (salt_bufs[salt_pos].salt_buf[ 0]);
salt_buf0[1] = swap32 (salt_bufs[salt_pos].salt_buf[ 1]);
salt_buf0[2] = swap32 (salt_bufs[salt_pos].salt_buf[ 2]);
salt_buf0[3] = swap32 (salt_bufs[salt_pos].salt_buf[ 3]);
salt_buf0[0] = swap32_S (salt_bufs[salt_pos].salt_buf[ 0]);
salt_buf0[1] = swap32_S (salt_bufs[salt_pos].salt_buf[ 1]);
salt_buf0[2] = swap32_S (salt_bufs[salt_pos].salt_buf[ 2]);
salt_buf0[3] = swap32_S (salt_bufs[salt_pos].salt_buf[ 3]);
u32 salt_buf1[4];
salt_buf1[0] = swap32 (salt_bufs[salt_pos].salt_buf[ 4]);
salt_buf1[1] = swap32 (salt_bufs[salt_pos].salt_buf[ 5]);
salt_buf1[2] = swap32 (salt_bufs[salt_pos].salt_buf[ 6]);
salt_buf1[3] = swap32 (salt_bufs[salt_pos].salt_buf[ 7]);
salt_buf1[0] = swap32_S (salt_bufs[salt_pos].salt_buf[ 4]);
salt_buf1[1] = swap32_S (salt_bufs[salt_pos].salt_buf[ 5]);
salt_buf1[2] = swap32_S (salt_bufs[salt_pos].salt_buf[ 6]);
salt_buf1[3] = swap32_S (salt_bufs[salt_pos].salt_buf[ 7]);
u32 salt_buf2[4];
salt_buf2[0] = swap32 (salt_bufs[salt_pos].salt_buf[ 8]);
salt_buf2[1] = swap32 (salt_bufs[salt_pos].salt_buf[ 9]);
salt_buf2[0] = swap32_S (salt_bufs[salt_pos].salt_buf[ 8]);
salt_buf2[1] = swap32_S (salt_bufs[salt_pos].salt_buf[ 9]);
salt_buf2[2] = 0;
salt_buf2[3] = 0;
@ -240,69 +240,46 @@ __kernel void m08400_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];
u32x w0[4] = { 0 };
u32x w1[4] = { 0 };
u32x w2[4] = { 0 };
u32x w3[4] = { 0 };
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];
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);
u32 w0_t[4];
u32x w0_t[4];
w0_t[0] = swap32 (w0[0]);
w0_t[1] = swap32 (w0[1]);
w0_t[2] = swap32 (w0[2]);
w0_t[3] = swap32 (w0[3]);
u32 w1_t[4];
u32x w1_t[4];
w1_t[0] = swap32 (w1[0]);
w1_t[1] = swap32 (w1[1]);
w1_t[2] = swap32 (w1[2]);
w1_t[3] = swap32 (w1[3]);
u32 w2_t[4];
u32x w2_t[4];
w2_t[0] = swap32 (w2[0]);
w2_t[1] = swap32 (w2[1]);
w2_t[2] = swap32 (w2[2]);
w2_t[3] = swap32 (w2[3]);
u32 w3_t[4];
u32x w3_t[4];
w3_t[0] = swap32 (w3[0]);
w3_t[1] = swap32 (w3[1]);
w3_t[2] = 0;
w3_t[3] = out_len * 8;
u32 digest[5];
u32x digest[5];
digest[0] = SHA1M_A;
digest[1] = SHA1M_B;
@ -312,11 +289,11 @@ __kernel void m08400_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
sha1_transform (w0_t, w1_t, w2_t, w3_t, digest);
u32 a;
u32 b;
u32 c;
u32 d;
u32 e;
u32x a;
u32x b;
u32x c;
u32x d;
u32x e;
a = digest[0];
b = digest[1];
@ -438,12 +415,7 @@ __kernel void m08400_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
sha1_transform (w0_t, w1_t, w2_t, w3_t, digest);
const u32 r0 = digest[3];
const u32 r1 = digest[4];
const u32 r2 = digest[2];
const u32 r3 = digest[1];
#include COMPARE_M
COMPARE_M_SIMD (digest[3], digest[4], digest[2], digest[1]);
}
}
@ -510,22 +482,22 @@ __kernel void m08400_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
u32 salt_buf0[4];
salt_buf0[0] = swap32 (salt_bufs[salt_pos].salt_buf[ 0]);
salt_buf0[1] = swap32 (salt_bufs[salt_pos].salt_buf[ 1]);
salt_buf0[2] = swap32 (salt_bufs[salt_pos].salt_buf[ 2]);
salt_buf0[3] = swap32 (salt_bufs[salt_pos].salt_buf[ 3]);
salt_buf0[0] = swap32_S (salt_bufs[salt_pos].salt_buf[ 0]);
salt_buf0[1] = swap32_S (salt_bufs[salt_pos].salt_buf[ 1]);
salt_buf0[2] = swap32_S (salt_bufs[salt_pos].salt_buf[ 2]);
salt_buf0[3] = swap32_S (salt_bufs[salt_pos].salt_buf[ 3]);
u32 salt_buf1[4];
salt_buf1[0] = swap32 (salt_bufs[salt_pos].salt_buf[ 4]);
salt_buf1[1] = swap32 (salt_bufs[salt_pos].salt_buf[ 5]);
salt_buf1[2] = swap32 (salt_bufs[salt_pos].salt_buf[ 6]);
salt_buf1[3] = swap32 (salt_bufs[salt_pos].salt_buf[ 7]);
salt_buf1[0] = swap32_S (salt_bufs[salt_pos].salt_buf[ 4]);
salt_buf1[1] = swap32_S (salt_bufs[salt_pos].salt_buf[ 5]);
salt_buf1[2] = swap32_S (salt_bufs[salt_pos].salt_buf[ 6]);
salt_buf1[3] = swap32_S (salt_bufs[salt_pos].salt_buf[ 7]);
u32 salt_buf2[4];
salt_buf2[0] = swap32 (salt_bufs[salt_pos].salt_buf[ 8]);
salt_buf2[1] = swap32 (salt_bufs[salt_pos].salt_buf[ 9]);
salt_buf2[0] = swap32_S (salt_bufs[salt_pos].salt_buf[ 8]);
salt_buf2[1] = swap32_S (salt_bufs[salt_pos].salt_buf[ 9]);
salt_buf2[2] = 0;
salt_buf2[3] = 0;
@ -547,69 +519,46 @@ __kernel void m08400_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];
u32x w0[4] = { 0 };
u32x w1[4] = { 0 };
u32x w2[4] = { 0 };
u32x w3[4] = { 0 };
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];
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);
u32 w0_t[4];
u32x w0_t[4];
w0_t[0] = swap32 (w0[0]);
w0_t[1] = swap32 (w0[1]);
w0_t[2] = swap32 (w0[2]);
w0_t[3] = swap32 (w0[3]);
u32 w1_t[4];
u32x w1_t[4];
w1_t[0] = swap32 (w1[0]);
w1_t[1] = swap32 (w1[1]);
w1_t[2] = swap32 (w1[2]);
w1_t[3] = swap32 (w1[3]);
u32 w2_t[4];
u32x w2_t[4];
w2_t[0] = swap32 (w2[0]);
w2_t[1] = swap32 (w2[1]);
w2_t[2] = swap32 (w2[2]);
w2_t[3] = swap32 (w2[3]);
u32 w3_t[4];
u32x w3_t[4];
w3_t[0] = swap32 (w3[0]);
w3_t[1] = swap32 (w3[1]);
w3_t[2] = 0;
w3_t[3] = out_len * 8;
u32 digest[5];
u32x digest[5];
digest[0] = SHA1M_A;
digest[1] = SHA1M_B;
@ -619,11 +568,11 @@ __kernel void m08400_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
sha1_transform (w0_t, w1_t, w2_t, w3_t, digest);
u32 a;
u32 b;
u32 c;
u32 d;
u32 e;
u32x a;
u32x b;
u32x c;
u32x d;
u32x e;
a = digest[0];
b = digest[1];
@ -745,12 +694,7 @@ __kernel void m08400_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
sha1_transform (w0_t, w1_t, w2_t, w3_t, digest);
const u32 r0 = digest[3];
const u32 r1 = digest[4];
const u32 r2 = digest[2];
const u32 r3 = digest[1];
#include COMPARE_S
COMPARE_S_SIMD (digest[3], digest[4], digest[2], digest[1]);
}
}