arno/hashcat
1
0
Fork 0
mirror of https://github.com/hashcat/hashcat.git synced 2024-11-22 16:18:09 +00:00
Code Issues Releases Wiki Activity

Converted to new SIMD: -m 10800 -a 0

Browse Source
This commit is contained in:
Jens Steube 2016-02-04 22:52:01 +01:00
parent 1bede41eac
commit 09dfc98797
1 changed files with 60 additions and 106 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

166
OpenCL/m10800_a0.cl
View File

@ -5,6 +5,8 @@
#define _SHA384_
#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"
__constant u64 k_sha384[80] =
{
@ -46,33 +46,33 @@ __constant u64 k_sha384[80] =
SHA384C4c, SHA384C4d, SHA384C4e, SHA384C4f,
};
static void sha384_transform (const u32 w0[4], const u32 w1[4], const u32 w2[4], const u32 w3[4], u64 digest[8])
static void sha384_transform (const u32x w0[4], const u32x w1[4], const u32x w2[4], const u32x w3[4], u64x digest[8])
{
u64 w0_t = hl32_to_64 (w0[0], w0[1]);
u64 w1_t = hl32_to_64 (w0[2], w0[3]);
u64 w2_t = hl32_to_64 (w1[0], w1[1]);
u64 w3_t = hl32_to_64 (w1[2], w1[3]);
u64 w4_t = hl32_to_64 (w2[0], w2[1]);
u64 w5_t = hl32_to_64 (w2[2], w2[3]);
u64 w6_t = hl32_to_64 (w3[0], w3[1]);
u64 w7_t = 0;
u64 w8_t = 0;
u64 w9_t = 0;
u64 wa_t = 0;
u64 wb_t = 0;
u64 wc_t = 0;
u64 wd_t = 0;
u64 we_t = 0;
u64 wf_t = hl32_to_64 (w3[2], w3[3]);
u64x w0_t = hl32_to_64 (w0[0], w0[1]);
u64x w1_t = hl32_to_64 (w0[2], w0[3]);
u64x w2_t = hl32_to_64 (w1[0], w1[1]);
u64x w3_t = hl32_to_64 (w1[2], w1[3]);
u64x w4_t = hl32_to_64 (w2[0], w2[1]);
u64x w5_t = hl32_to_64 (w2[2], w2[3]);
u64x w6_t = hl32_to_64 (w3[0], w3[1]);
u64x w7_t = 0;
u64x w8_t = 0;
u64x w9_t = 0;
u64x wa_t = 0;
u64x wb_t = 0;
u64x wc_t = 0;
u64x wd_t = 0;
u64x we_t = 0;
u64x wf_t = hl32_to_64 (w3[2], w3[3]);
u64 a = digest[0];
u64 b = digest[1];
u64 c = digest[2];
u64 d = digest[3];
u64 e = digest[4];
u64 f = digest[5];
u64 g = digest[6];
u64 h = digest[7];
u64x a = digest[0];
u64x b = digest[1];
u64x c = digest[2];
u64x d = digest[3];
u64x e = digest[4];
u64x f = digest[5];
u64x g = digest[6];
u64x h = digest[7];
#define ROUND_EXPAND() \
{ \
@ -178,37 +178,14 @@ __kernel void m10800_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);
@ -216,10 +193,10 @@ __kernel void m10800_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
* SHA384
*/
u32 w0_t[4];
u32 w1_t[4];
u32 w2_t[4];
u32 w3_t[4];
u32x w0_t[4];
u32x w1_t[4];
u32x w2_t[4];
u32x w3_t[4];
w0_t[0] = swap32 (w0[0]);
w0_t[1] = swap32 (w0[1]);
@ -238,7 +215,7 @@ __kernel void m10800_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
w3_t[2] = 0;
w3_t[3] = out_len * 8;
u64 digest[8];
u64x digest[8];
digest[0] = SHA384M_A;
digest[1] = SHA384M_B;
@ -251,12 +228,12 @@ __kernel void m10800_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
sha384_transform (w0_t, w1_t, w2_t, w3_t, digest);
const u32 r0 = l32_from_64 (digest[3]);
const u32 r1 = h32_from_64 (digest[3]);
const u32 r2 = l32_from_64 (digest[2]);
const u32 r3 = h32_from_64 (digest[2]);
const u32x r0 = l32_from_64 (digest[3]);
const u32x r1 = h32_from_64 (digest[3]);
const u32x r2 = l32_from_64 (digest[2]);
const u32x r3 = h32_from_64 (digest[2]);
#include COMPARE_M
COMPARE_M_SIMD (r0, r1, r2, r3);
}
}
@ -316,37 +293,14 @@ __kernel void m10800_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);
@ -354,10 +308,10 @@ __kernel void m10800_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
* SHA384
*/
u32 w0_t[4];
u32 w1_t[4];
u32 w2_t[4];
u32 w3_t[4];
u32x w0_t[4];
u32x w1_t[4];
u32x w2_t[4];
u32x w3_t[4];
w0_t[0] = swap32 (w0[0]);
w0_t[1] = swap32 (w0[1]);
@ -376,7 +330,7 @@ __kernel void m10800_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
w3_t[2] = 0;
w3_t[3] = out_len * 8;
u64 digest[8];
u64x digest[8];
digest[0] = SHA384M_A;
digest[1] = SHA384M_B;
@ -389,12 +343,12 @@ __kernel void m10800_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf,
sha384_transform (w0_t, w1_t, w2_t, w3_t, digest);
const u32 r0 = l32_from_64 (digest[3]);
const u32 r1 = h32_from_64 (digest[3]);
const u32 r2 = l32_from_64 (digest[2]);
const u32 r3 = h32_from_64 (digest[2]);
const u32x r0 = l32_from_64 (digest[3]);
const u32x r1 = h32_from_64 (digest[3]);
const u32x r2 = l32_from_64 (digest[2]);
const u32x r3 = h32_from_64 (digest[2]);
#include COMPARE_S
COMPARE_S_SIMD (r0, r1, r2, r3);
}
}