diff --git a/OpenCL/m07600_a0.cl b/OpenCL/m07600_a0.cl index d46c16378..86a001255 100644 --- a/OpenCL/m07600_a0.cl +++ b/OpenCL/m07600_a0.cl @@ -70,18 +70,16 @@ __kernel void m07600_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, */ 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; @@ -90,37 +88,29 @@ __kernel void m07600_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, */ u32 salt_buf0[4]; + u32 salt_buf1[4]; + u32 salt_buf2[4]; + u32 salt_buf3[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]; salt_buf1[3] = salt_bufs[salt_pos].salt_buf[ 7]; - - u32 salt_buf2[4]; - - salt_buf2[0] = 0; - salt_buf2[1] = 0; - salt_buf2[2] = 0; - salt_buf2[3] = 0; - - u32 salt_buf3[4]; - - salt_buf3[0] = 0; - salt_buf3[1] = 0; - salt_buf3[2] = 0; - salt_buf3[3] = 0; + salt_buf2[0] = salt_bufs[salt_pos].salt_buf[ 8]; + salt_buf2[1] = salt_bufs[salt_pos].salt_buf[ 9]; + salt_buf2[2] = salt_bufs[salt_pos].salt_buf[10]; + salt_buf2[3] = salt_bufs[salt_pos].salt_buf[11]; + salt_buf3[0] = salt_bufs[salt_pos].salt_buf[12]; + salt_buf3[1] = salt_bufs[salt_pos].salt_buf[13]; + salt_buf3[2] = salt_bufs[salt_pos].salt_buf[14]; + salt_buf3[3] = salt_bufs[salt_pos].salt_buf[15]; const u32 salt_len = salt_bufs[salt_pos].salt_len; - const u32 total_len = (salt_len + 40) * 8; - /** * loop */ @@ -269,35 +259,6 @@ __kernel void m07600_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, * Prepend salt */ - u32x w0t[4]; - - w0t[0] = uint_to_hex_lower8 ((a >> 24) & 255) << 0 - | uint_to_hex_lower8 ((a >> 16) & 255) << 16; - w0t[1] = uint_to_hex_lower8 ((a >> 8) & 255) << 0 - | uint_to_hex_lower8 ((a >> 0) & 255) << 16; - w0t[2] = uint_to_hex_lower8 ((b >> 24) & 255) << 0 - | uint_to_hex_lower8 ((b >> 16) & 255) << 16; - w0t[3] = uint_to_hex_lower8 ((b >> 8) & 255) << 0 - | uint_to_hex_lower8 ((b >> 0) & 255) << 16; - - u32x w1t[4]; - - w1t[0] = uint_to_hex_lower8 ((c >> 24) & 255) << 0 - | uint_to_hex_lower8 ((c >> 16) & 255) << 16; - w1t[1] = uint_to_hex_lower8 ((c >> 8) & 255) << 0 - | uint_to_hex_lower8 ((c >> 0) & 255) << 16; - w1t[2] = uint_to_hex_lower8 ((d >> 24) & 255) << 0 - | uint_to_hex_lower8 ((d >> 16) & 255) << 16; - w1t[3] = uint_to_hex_lower8 ((d >> 8) & 255) << 0 - | uint_to_hex_lower8 ((d >> 0) & 255) << 16; - - u32x w2t[2]; - - w2t[0] = uint_to_hex_lower8 ((e >> 24) & 255) << 0 - | uint_to_hex_lower8 ((e >> 16) & 255) << 16; - w2t[1] = uint_to_hex_lower8 ((e >> 8) & 255) << 0 - | uint_to_hex_lower8 ((e >> 0) & 255) << 16; - w0_t = salt_buf0[0]; w1_t = salt_buf0[1]; w2_t = salt_buf0[2]; @@ -306,14 +267,24 @@ __kernel void m07600_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, w5_t = salt_buf1[1]; w6_t = salt_buf1[2]; w7_t = salt_buf1[3]; - w8_t = w0t[0]; - w9_t = w0t[1]; - wa_t = w0t[2]; - wb_t = w0t[3]; - wc_t = w1t[0]; - wd_t = w1t[1]; - we_t = w1t[2]; - wf_t = w1t[3]; + w8_t = uint_to_hex_lower8 ((a >> 24) & 255) << 0 + | uint_to_hex_lower8 ((a >> 16) & 255) << 16; + w9_t = uint_to_hex_lower8 ((a >> 8) & 255) << 0 + | uint_to_hex_lower8 ((a >> 0) & 255) << 16; + wa_t = uint_to_hex_lower8 ((b >> 24) & 255) << 0 + | uint_to_hex_lower8 ((b >> 16) & 255) << 16; + wb_t = uint_to_hex_lower8 ((b >> 8) & 255) << 0 + | uint_to_hex_lower8 ((b >> 0) & 255) << 16; + wc_t = uint_to_hex_lower8 ((c >> 24) & 255) << 0 + | uint_to_hex_lower8 ((c >> 16) & 255) << 16; + wd_t = uint_to_hex_lower8 ((c >> 8) & 255) << 0 + | uint_to_hex_lower8 ((c >> 0) & 255) << 16; + we_t = uint_to_hex_lower8 ((d >> 24) & 255) << 0 + | uint_to_hex_lower8 ((d >> 16) & 255) << 16; + wf_t = uint_to_hex_lower8 ((d >> 8) & 255) << 0 + | uint_to_hex_lower8 ((d >> 0) & 255) << 16; + + const u32x e_sav = e; /** * 2nd SHA1 @@ -454,8 +425,10 @@ __kernel void m07600_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, // 2nd transform - w0_t = swap32 (w2t[0]); - w1_t = swap32 (w2t[1]); + w0_t = uint_to_hex_lower8 ((e_sav >> 24) & 255) << 0 + | uint_to_hex_lower8 ((e_sav >> 16) & 255) << 16; + w1_t = uint_to_hex_lower8 ((e_sav >> 8) & 255) << 0 + | uint_to_hex_lower8 ((e_sav >> 0) & 255) << 16; w2_t = 0x80000000; w3_t = 0; w4_t = 0; @@ -469,7 +442,10 @@ __kernel void m07600_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, wc_t = 0; wd_t = 0; we_t = 0; - wf_t = total_len; + wf_t = (salt_len + 40) * 8; + + w0_t = swap32 (w0_t); + w1_t = swap32 (w1_t); #undef K #define K SHA1C00 @@ -619,18 +595,16 @@ __kernel void m07600_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, */ 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; @@ -639,37 +613,29 @@ __kernel void m07600_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, */ u32 salt_buf0[4]; + u32 salt_buf1[4]; + u32 salt_buf2[4]; + u32 salt_buf3[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]; salt_buf1[3] = salt_bufs[salt_pos].salt_buf[ 7]; - - u32 salt_buf2[4]; - - salt_buf2[0] = 0; - salt_buf2[1] = 0; - salt_buf2[2] = 0; - salt_buf2[3] = 0; - - u32 salt_buf3[4]; - - salt_buf3[0] = 0; - salt_buf3[1] = 0; - salt_buf3[2] = 0; - salt_buf3[3] = 0; + salt_buf2[0] = salt_bufs[salt_pos].salt_buf[ 8]; + salt_buf2[1] = salt_bufs[salt_pos].salt_buf[ 9]; + salt_buf2[2] = salt_bufs[salt_pos].salt_buf[10]; + salt_buf2[3] = salt_bufs[salt_pos].salt_buf[11]; + salt_buf3[0] = salt_bufs[salt_pos].salt_buf[12]; + salt_buf3[1] = salt_bufs[salt_pos].salt_buf[13]; + salt_buf3[2] = salt_bufs[salt_pos].salt_buf[14]; + salt_buf3[3] = salt_bufs[salt_pos].salt_buf[15]; const u32 salt_len = salt_bufs[salt_pos].salt_len; - const u32 total_len = (salt_len + 40) * 8; - /** * digest */ @@ -682,12 +648,6 @@ __kernel void m07600_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, digests_buf[digests_offset].digest_buf[DGST_R3] }; - /** - * reverse - */ - - const u32 e_rev = rotl32_S (search[1], 2u); - /** * loop */ @@ -836,35 +796,6 @@ __kernel void m07600_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, * Prepend salt */ - u32x w0t[4]; - - w0t[0] = uint_to_hex_lower8 ((a >> 24) & 255) << 0 - | uint_to_hex_lower8 ((a >> 16) & 255) << 16; - w0t[1] = uint_to_hex_lower8 ((a >> 8) & 255) << 0 - | uint_to_hex_lower8 ((a >> 0) & 255) << 16; - w0t[2] = uint_to_hex_lower8 ((b >> 24) & 255) << 0 - | uint_to_hex_lower8 ((b >> 16) & 255) << 16; - w0t[3] = uint_to_hex_lower8 ((b >> 8) & 255) << 0 - | uint_to_hex_lower8 ((b >> 0) & 255) << 16; - - u32x w1t[4]; - - w1t[0] = uint_to_hex_lower8 ((c >> 24) & 255) << 0 - | uint_to_hex_lower8 ((c >> 16) & 255) << 16; - w1t[1] = uint_to_hex_lower8 ((c >> 8) & 255) << 0 - | uint_to_hex_lower8 ((c >> 0) & 255) << 16; - w1t[2] = uint_to_hex_lower8 ((d >> 24) & 255) << 0 - | uint_to_hex_lower8 ((d >> 16) & 255) << 16; - w1t[3] = uint_to_hex_lower8 ((d >> 8) & 255) << 0 - | uint_to_hex_lower8 ((d >> 0) & 255) << 16; - - u32x w2t[2]; - - w2t[0] = uint_to_hex_lower8 ((e >> 24) & 255) << 0 - | uint_to_hex_lower8 ((e >> 16) & 255) << 16; - w2t[1] = uint_to_hex_lower8 ((e >> 8) & 255) << 0 - | uint_to_hex_lower8 ((e >> 0) & 255) << 16; - w0_t = salt_buf0[0]; w1_t = salt_buf0[1]; w2_t = salt_buf0[2]; @@ -873,14 +804,24 @@ __kernel void m07600_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, w5_t = salt_buf1[1]; w6_t = salt_buf1[2]; w7_t = salt_buf1[3]; - w8_t = w0t[0]; - w9_t = w0t[1]; - wa_t = w0t[2]; - wb_t = w0t[3]; - wc_t = w1t[0]; - wd_t = w1t[1]; - we_t = w1t[2]; - wf_t = w1t[3]; + w8_t = uint_to_hex_lower8 ((a >> 24) & 255) << 0 + | uint_to_hex_lower8 ((a >> 16) & 255) << 16; + w9_t = uint_to_hex_lower8 ((a >> 8) & 255) << 0 + | uint_to_hex_lower8 ((a >> 0) & 255) << 16; + wa_t = uint_to_hex_lower8 ((b >> 24) & 255) << 0 + | uint_to_hex_lower8 ((b >> 16) & 255) << 16; + wb_t = uint_to_hex_lower8 ((b >> 8) & 255) << 0 + | uint_to_hex_lower8 ((b >> 0) & 255) << 16; + wc_t = uint_to_hex_lower8 ((c >> 24) & 255) << 0 + | uint_to_hex_lower8 ((c >> 16) & 255) << 16; + wd_t = uint_to_hex_lower8 ((c >> 8) & 255) << 0 + | uint_to_hex_lower8 ((c >> 0) & 255) << 16; + we_t = uint_to_hex_lower8 ((d >> 24) & 255) << 0 + | uint_to_hex_lower8 ((d >> 16) & 255) << 16; + wf_t = uint_to_hex_lower8 ((d >> 8) & 255) << 0 + | uint_to_hex_lower8 ((d >> 0) & 255) << 16; + + const u32x e_sav = e; /** * 2nd SHA1 @@ -1021,8 +962,10 @@ __kernel void m07600_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, // 2nd transform - w0_t = swap32 (w2t[0]); - w1_t = swap32 (w2t[1]); + w0_t = uint_to_hex_lower8 ((e_sav >> 24) & 255) << 0 + | uint_to_hex_lower8 ((e_sav >> 16) & 255) << 16; + w1_t = uint_to_hex_lower8 ((e_sav >> 8) & 255) << 0 + | uint_to_hex_lower8 ((e_sav >> 0) & 255) << 16; w2_t = 0x80000000; w3_t = 0; w4_t = 0; @@ -1036,7 +979,10 @@ __kernel void m07600_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, wc_t = 0; wd_t = 0; we_t = 0; - wf_t = total_len; + wf_t = (salt_len + 40) * 8; + + w0_t = swap32 (w0_t); + w1_t = swap32 (w1_t); #undef K #define K SHA1C00 diff --git a/OpenCL/m07600_a1.cl b/OpenCL/m07600_a1.cl index 93323d484..531d5f37b 100644 --- a/OpenCL/m07600_a1.cl +++ b/OpenCL/m07600_a1.cl @@ -5,6 +5,8 @@ #define _SHA1_ +#define NEW_SIMD_CODE + #include "include/constants.h" #include "include/kernel_vendor.h" @@ -16,11 +18,19 @@ #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" - -#define uint_to_hex_lower8(i) l_bin2asc[(i)] +#include "OpenCL/simd.c" + +#if VECT_SIZE == 1 +#define uint_to_hex_lower8(i) (u32x) (l_bin2asc[(i)]) +#elif VECT_SIZE == 2 +#define uint_to_hex_lower8(i) (u32x) (l_bin2asc[(i).s0], l_bin2asc[(i).s1]) +#elif VECT_SIZE == 4 +#define uint_to_hex_lower8(i) (u32x) (l_bin2asc[(i).s0], l_bin2asc[(i).s1], l_bin2asc[(i).s2], l_bin2asc[(i).s3]) +#elif VECT_SIZE == 8 +#define uint_to_hex_lower8(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]) +#elif VECT_SIZE == 16 +#define uint_to_hex_lower8(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], l_bin2asc[(i).s8], l_bin2asc[(i).s9], l_bin2asc[(i).sa], l_bin2asc[(i).sb], l_bin2asc[(i).sc], l_bin2asc[(i).sd], l_bin2asc[(i).se], l_bin2asc[(i).sf]) +#endif __kernel void m07600_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 il_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max) { @@ -55,178 +65,147 @@ __kernel void m07600_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, * base */ - 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 pw_buf0[4]; + u32 pw_buf1[4]; - 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; + 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_l_len = pws[gid].pw_len; - if (combs_mode == COMBINATOR_MODE_BASE_RIGHT) - { - append_0x80_2x4 (wordl0, wordl1, pw_l_len); - - switch_buffer_by_offset_le (wordl0, wordl1, wordl2, wordl3, combs_buf[0].pw_len); - } - /** * salt */ u32 salt_buf0[4]; + u32 salt_buf1[4]; + u32 salt_buf2[4]; + u32 salt_buf3[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]; salt_buf1[3] = salt_bufs[salt_pos].salt_buf[ 7]; - - u32 salt_buf2[4]; - - salt_buf2[0] = 0; - salt_buf2[1] = 0; - salt_buf2[2] = 0; - salt_buf2[3] = 0; - - u32 salt_buf3[4]; - - salt_buf3[0] = 0; - salt_buf3[1] = 0; - salt_buf3[2] = 0; - salt_buf3[3] = 0; + salt_buf2[0] = salt_bufs[salt_pos].salt_buf[ 8]; + salt_buf2[1] = salt_bufs[salt_pos].salt_buf[ 9]; + salt_buf2[2] = salt_bufs[salt_pos].salt_buf[10]; + salt_buf2[3] = salt_bufs[salt_pos].salt_buf[11]; + salt_buf3[0] = salt_bufs[salt_pos].salt_buf[12]; + salt_buf3[1] = salt_bufs[salt_pos].salt_buf[13]; + salt_buf3[2] = salt_bufs[salt_pos].salt_buf[14]; + salt_buf3[3] = salt_bufs[salt_pos].salt_buf[15]; const u32 salt_len = salt_bufs[salt_pos].salt_len; - const u32 total_len = (salt_len + 40) * 8; - /** * loop */ - for (u32 il_pos = 0; il_pos < il_cnt; il_pos++) + for (u32 il_pos = 0; il_pos < il_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; + const u32x pw_r_len = pwlenx_create_combt (combs_buf, il_pos); - u32 wordr0[4]; + const u32x pw_len = pw_l_len + pw_r_len; - 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]; - - u32 wordr1[4]; - - 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]; + /** + * concat password candidate + */ - wordr3[0] = 0; - wordr3[1] = 0; - wordr3[2] = 0; - wordr3[3] = 0; + u32x wordl0[4] = { 0 }; + u32x wordl1[4] = { 0 }; + u32x wordl2[4] = { 0 }; + u32x wordl3[4] = { 0 }; + + wordl0[0] = pw_buf0[0]; + wordl0[1] = pw_buf0[1]; + wordl0[2] = pw_buf0[2]; + wordl0[3] = pw_buf0[3]; + wordl1[0] = pw_buf1[0]; + wordl1[1] = pw_buf1[1]; + wordl1[2] = pw_buf1[2]; + wordl1[3] = pw_buf1[3]; + + u32x wordr0[4] = { 0 }; + u32x wordr1[4] = { 0 }; + u32x wordr2[4] = { 0 }; + u32x wordr3[4] = { 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) { - append_0x80_2x4 (wordr0, wordr1, pw_r_len); - - switch_buffer_by_offset_le (wordr0, wordr1, wordr2, wordr3, pw_l_len); + switch_buffer_by_offset_le_VV (wordr0, wordr1, wordr2, wordr3, pw_l_len); + } + else + { + switch_buffer_by_offset_le_VV (wordl0, wordl1, wordl2, wordl3, pw_r_len); } - u32 w0[4]; + u32x w0[4]; + u32x w1[4]; + u32x w2[4]; + u32x w3[4]; w0[0] = wordl0[0] | wordr0[0]; w0[1] = wordl0[1] | wordr0[1]; w0[2] = wordl0[2] | wordr0[2]; w0[3] = wordl0[3] | wordr0[3]; - - u32 w1[4]; - w1[0] = wordl1[0] | wordr1[0]; w1[1] = wordl1[1] | wordr1[1]; w1[2] = wordl1[2] | wordr1[2]; w1[3] = wordl1[3] | wordr1[3]; - - u32 w2[4]; - w2[0] = wordl2[0] | wordr2[0]; w2[1] = wordl2[1] | wordr2[1]; w2[2] = wordl2[2] | wordr2[2]; w2[3] = wordl2[3] | wordr2[3]; - - u32 w3[4]; - w3[0] = wordl3[0] | wordr3[0]; w3[1] = wordl3[1] | wordr3[1]; - w3[2] = 0; - w3[3] = 0; + w3[2] = wordl3[2] | wordr3[2]; + w3[3] = wordl3[3] | wordr3[3]; /** * 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_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_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 @@ -334,35 +313,6 @@ __kernel void m07600_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, * Prepend salt */ - u32 w0t[4]; - - w0t[0] = uint_to_hex_lower8 ((a >> 24) & 255) << 0 - | uint_to_hex_lower8 ((a >> 16) & 255) << 16; - w0t[1] = uint_to_hex_lower8 ((a >> 8) & 255) << 0 - | uint_to_hex_lower8 ((a >> 0) & 255) << 16; - w0t[2] = uint_to_hex_lower8 ((b >> 24) & 255) << 0 - | uint_to_hex_lower8 ((b >> 16) & 255) << 16; - w0t[3] = uint_to_hex_lower8 ((b >> 8) & 255) << 0 - | uint_to_hex_lower8 ((b >> 0) & 255) << 16; - - u32 w1t[4]; - - w1t[0] = uint_to_hex_lower8 ((c >> 24) & 255) << 0 - | uint_to_hex_lower8 ((c >> 16) & 255) << 16; - w1t[1] = uint_to_hex_lower8 ((c >> 8) & 255) << 0 - | uint_to_hex_lower8 ((c >> 0) & 255) << 16; - w1t[2] = uint_to_hex_lower8 ((d >> 24) & 255) << 0 - | uint_to_hex_lower8 ((d >> 16) & 255) << 16; - w1t[3] = uint_to_hex_lower8 ((d >> 8) & 255) << 0 - | uint_to_hex_lower8 ((d >> 0) & 255) << 16; - - u32 w2t[2]; - - w2t[0] = uint_to_hex_lower8 ((e >> 24) & 255) << 0 - | uint_to_hex_lower8 ((e >> 16) & 255) << 16; - w2t[1] = uint_to_hex_lower8 ((e >> 8) & 255) << 0 - | uint_to_hex_lower8 ((e >> 0) & 255) << 16; - w0_t = salt_buf0[0]; w1_t = salt_buf0[1]; w2_t = salt_buf0[2]; @@ -371,14 +321,24 @@ __kernel void m07600_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, w5_t = salt_buf1[1]; w6_t = salt_buf1[2]; w7_t = salt_buf1[3]; - w8_t = w0t[0]; - w9_t = w0t[1]; - wa_t = w0t[2]; - wb_t = w0t[3]; - wc_t = w1t[0]; - wd_t = w1t[1]; - we_t = w1t[2]; - wf_t = w1t[3]; + w8_t = uint_to_hex_lower8 ((a >> 24) & 255) << 0 + | uint_to_hex_lower8 ((a >> 16) & 255) << 16; + w9_t = uint_to_hex_lower8 ((a >> 8) & 255) << 0 + | uint_to_hex_lower8 ((a >> 0) & 255) << 16; + wa_t = uint_to_hex_lower8 ((b >> 24) & 255) << 0 + | uint_to_hex_lower8 ((b >> 16) & 255) << 16; + wb_t = uint_to_hex_lower8 ((b >> 8) & 255) << 0 + | uint_to_hex_lower8 ((b >> 0) & 255) << 16; + wc_t = uint_to_hex_lower8 ((c >> 24) & 255) << 0 + | uint_to_hex_lower8 ((c >> 16) & 255) << 16; + wd_t = uint_to_hex_lower8 ((c >> 8) & 255) << 0 + | uint_to_hex_lower8 ((c >> 0) & 255) << 16; + we_t = uint_to_hex_lower8 ((d >> 24) & 255) << 0 + | uint_to_hex_lower8 ((d >> 16) & 255) << 16; + wf_t = uint_to_hex_lower8 ((d >> 8) & 255) << 0 + | uint_to_hex_lower8 ((d >> 0) & 255) << 16; + + const u32x e_sav = e; /** * 2nd SHA1 @@ -511,16 +471,18 @@ __kernel void m07600_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, d += SHA1M_D; e += SHA1M_E; - u32 r_a = a; - u32 r_b = b; - u32 r_c = c; - u32 r_d = d; - u32 r_e = e; + u32x r_a = a; + u32x r_b = b; + u32x r_c = c; + u32x r_d = d; + u32x r_e = e; // 2nd transform - w0_t = swap32 (w2t[0]); - w1_t = swap32 (w2t[1]); + w0_t = uint_to_hex_lower8 ((e_sav >> 24) & 255) << 0 + | uint_to_hex_lower8 ((e_sav >> 16) & 255) << 16; + w1_t = uint_to_hex_lower8 ((e_sav >> 8) & 255) << 0 + | uint_to_hex_lower8 ((e_sav >> 0) & 255) << 16; w2_t = 0x80000000; w3_t = 0; w4_t = 0; @@ -534,7 +496,10 @@ __kernel void m07600_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, wc_t = 0; wd_t = 0; we_t = 0; - wf_t = total_len; + wf_t = (salt_len + 40) * 8; + + w0_t = swap32 (w0_t); + w1_t = swap32 (w1_t); #undef K #define K SHA1C00 @@ -638,12 +603,7 @@ __kernel void m07600_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, d += r_d; e += r_e; - 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); } } @@ -688,79 +648,48 @@ __kernel void m07600_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, * base */ - 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 pw_buf0[4]; + u32 pw_buf1[4]; - 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 wordl3[4]; - - wordl3[0] = 0; - wordl3[1] = 0; - wordl3[2] = 0; - wordl3[3] = 0; + 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_l_len = pws[gid].pw_len; - if (combs_mode == COMBINATOR_MODE_BASE_RIGHT) - { - append_0x80_2x4 (wordl0, wordl1, pw_l_len); - - switch_buffer_by_offset_le (wordl0, wordl1, wordl2, wordl3, combs_buf[0].pw_len); - } - /** * salt */ u32 salt_buf0[4]; + u32 salt_buf1[4]; + u32 salt_buf2[4]; + u32 salt_buf3[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]; salt_buf1[3] = salt_bufs[salt_pos].salt_buf[ 7]; - - u32 salt_buf2[4]; - - salt_buf2[0] = 0; - salt_buf2[1] = 0; - salt_buf2[2] = 0; - salt_buf2[3] = 0; - - u32 salt_buf3[4]; - - salt_buf3[0] = 0; - salt_buf3[1] = 0; - salt_buf3[2] = 0; - salt_buf3[3] = 0; + salt_buf2[0] = salt_bufs[salt_pos].salt_buf[ 8]; + salt_buf2[1] = salt_bufs[salt_pos].salt_buf[ 9]; + salt_buf2[2] = salt_bufs[salt_pos].salt_buf[10]; + salt_buf2[3] = salt_bufs[salt_pos].salt_buf[11]; + salt_buf3[0] = salt_bufs[salt_pos].salt_buf[12]; + salt_buf3[1] = salt_bufs[salt_pos].salt_buf[13]; + salt_buf3[2] = salt_bufs[salt_pos].salt_buf[14]; + salt_buf3[3] = salt_bufs[salt_pos].salt_buf[15]; const u32 salt_len = salt_bufs[salt_pos].salt_len; - const u32 total_len = (salt_len + 40) * 8; - /** * digest */ @@ -773,111 +702,105 @@ __kernel void m07600_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, digests_buf[digests_offset].digest_buf[DGST_R3] }; - /** - * reverse - */ - - const u32 e_rev = rotl32 (search[1], 2u); - /** * loop */ - for (u32 il_pos = 0; il_pos < il_cnt; il_pos++) + for (u32 il_pos = 0; il_pos < il_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; + const u32x pw_r_len = pwlenx_create_combt (combs_buf, il_pos); - u32 wordr0[4]; + const u32x pw_len = pw_l_len + pw_r_len; - 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]; - - u32 wordr1[4]; - - 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]; + /** + * concat password candidate + */ - wordr3[0] = 0; - wordr3[1] = 0; - wordr3[2] = 0; - wordr3[3] = 0; + u32x wordl0[4] = { 0 }; + u32x wordl1[4] = { 0 }; + u32x wordl2[4] = { 0 }; + u32x wordl3[4] = { 0 }; + + wordl0[0] = pw_buf0[0]; + wordl0[1] = pw_buf0[1]; + wordl0[2] = pw_buf0[2]; + wordl0[3] = pw_buf0[3]; + wordl1[0] = pw_buf1[0]; + wordl1[1] = pw_buf1[1]; + wordl1[2] = pw_buf1[2]; + wordl1[3] = pw_buf1[3]; + + u32x wordr0[4] = { 0 }; + u32x wordr1[4] = { 0 }; + u32x wordr2[4] = { 0 }; + u32x wordr3[4] = { 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) { - append_0x80_2x4 (wordr0, wordr1, pw_r_len); - - switch_buffer_by_offset_le (wordr0, wordr1, wordr2, wordr3, pw_l_len); + switch_buffer_by_offset_le_VV (wordr0, wordr1, wordr2, wordr3, pw_l_len); + } + else + { + switch_buffer_by_offset_le_VV (wordl0, wordl1, wordl2, wordl3, pw_r_len); } - u32 w0[4]; + u32x w0[4]; + u32x w1[4]; + u32x w2[4]; + u32x w3[4]; w0[0] = wordl0[0] | wordr0[0]; w0[1] = wordl0[1] | wordr0[1]; w0[2] = wordl0[2] | wordr0[2]; w0[3] = wordl0[3] | wordr0[3]; - - u32 w1[4]; - w1[0] = wordl1[0] | wordr1[0]; w1[1] = wordl1[1] | wordr1[1]; w1[2] = wordl1[2] | wordr1[2]; w1[3] = wordl1[3] | wordr1[3]; - - u32 w2[4]; - w2[0] = wordl2[0] | wordr2[0]; w2[1] = wordl2[1] | wordr2[1]; w2[2] = wordl2[2] | wordr2[2]; w2[3] = wordl2[3] | wordr2[3]; - - u32 w3[4]; - w3[0] = wordl3[0] | wordr3[0]; w3[1] = wordl3[1] | wordr3[1]; - w3[2] = 0; - w3[3] = 0; + w3[2] = wordl3[2] | wordr3[2]; + w3[3] = wordl3[3] | wordr3[3]; /** * 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_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_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 @@ -985,35 +908,6 @@ __kernel void m07600_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, * Prepend salt */ - u32 w0t[4]; - - w0t[0] = uint_to_hex_lower8 ((a >> 24) & 255) << 0 - | uint_to_hex_lower8 ((a >> 16) & 255) << 16; - w0t[1] = uint_to_hex_lower8 ((a >> 8) & 255) << 0 - | uint_to_hex_lower8 ((a >> 0) & 255) << 16; - w0t[2] = uint_to_hex_lower8 ((b >> 24) & 255) << 0 - | uint_to_hex_lower8 ((b >> 16) & 255) << 16; - w0t[3] = uint_to_hex_lower8 ((b >> 8) & 255) << 0 - | uint_to_hex_lower8 ((b >> 0) & 255) << 16; - - u32 w1t[4]; - - w1t[0] = uint_to_hex_lower8 ((c >> 24) & 255) << 0 - | uint_to_hex_lower8 ((c >> 16) & 255) << 16; - w1t[1] = uint_to_hex_lower8 ((c >> 8) & 255) << 0 - | uint_to_hex_lower8 ((c >> 0) & 255) << 16; - w1t[2] = uint_to_hex_lower8 ((d >> 24) & 255) << 0 - | uint_to_hex_lower8 ((d >> 16) & 255) << 16; - w1t[3] = uint_to_hex_lower8 ((d >> 8) & 255) << 0 - | uint_to_hex_lower8 ((d >> 0) & 255) << 16; - - u32 w2t[2]; - - w2t[0] = uint_to_hex_lower8 ((e >> 24) & 255) << 0 - | uint_to_hex_lower8 ((e >> 16) & 255) << 16; - w2t[1] = uint_to_hex_lower8 ((e >> 8) & 255) << 0 - | uint_to_hex_lower8 ((e >> 0) & 255) << 16; - w0_t = salt_buf0[0]; w1_t = salt_buf0[1]; w2_t = salt_buf0[2]; @@ -1022,14 +916,24 @@ __kernel void m07600_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, w5_t = salt_buf1[1]; w6_t = salt_buf1[2]; w7_t = salt_buf1[3]; - w8_t = w0t[0]; - w9_t = w0t[1]; - wa_t = w0t[2]; - wb_t = w0t[3]; - wc_t = w1t[0]; - wd_t = w1t[1]; - we_t = w1t[2]; - wf_t = w1t[3]; + w8_t = uint_to_hex_lower8 ((a >> 24) & 255) << 0 + | uint_to_hex_lower8 ((a >> 16) & 255) << 16; + w9_t = uint_to_hex_lower8 ((a >> 8) & 255) << 0 + | uint_to_hex_lower8 ((a >> 0) & 255) << 16; + wa_t = uint_to_hex_lower8 ((b >> 24) & 255) << 0 + | uint_to_hex_lower8 ((b >> 16) & 255) << 16; + wb_t = uint_to_hex_lower8 ((b >> 8) & 255) << 0 + | uint_to_hex_lower8 ((b >> 0) & 255) << 16; + wc_t = uint_to_hex_lower8 ((c >> 24) & 255) << 0 + | uint_to_hex_lower8 ((c >> 16) & 255) << 16; + wd_t = uint_to_hex_lower8 ((c >> 8) & 255) << 0 + | uint_to_hex_lower8 ((c >> 0) & 255) << 16; + we_t = uint_to_hex_lower8 ((d >> 24) & 255) << 0 + | uint_to_hex_lower8 ((d >> 16) & 255) << 16; + wf_t = uint_to_hex_lower8 ((d >> 8) & 255) << 0 + | uint_to_hex_lower8 ((d >> 0) & 255) << 16; + + const u32x e_sav = e; /** * 2nd SHA1 @@ -1162,16 +1066,18 @@ __kernel void m07600_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, d += SHA1M_D; e += SHA1M_E; - u32 r_a = a; - u32 r_b = b; - u32 r_c = c; - u32 r_d = d; - u32 r_e = e; + u32x r_a = a; + u32x r_b = b; + u32x r_c = c; + u32x r_d = d; + u32x r_e = e; // 2nd transform - w0_t = swap32 (w2t[0]); - w1_t = swap32 (w2t[1]); + w0_t = uint_to_hex_lower8 ((e_sav >> 24) & 255) << 0 + | uint_to_hex_lower8 ((e_sav >> 16) & 255) << 16; + w1_t = uint_to_hex_lower8 ((e_sav >> 8) & 255) << 0 + | uint_to_hex_lower8 ((e_sav >> 0) & 255) << 16; w2_t = 0x80000000; w3_t = 0; w4_t = 0; @@ -1185,7 +1091,10 @@ __kernel void m07600_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, wc_t = 0; wd_t = 0; we_t = 0; - wf_t = total_len; + wf_t = (salt_len + 40) * 8; + + w0_t = swap32 (w0_t); + w1_t = swap32 (w1_t); #undef K #define K SHA1C00 @@ -1289,12 +1198,7 @@ __kernel void m07600_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, d += r_d; e += r_e; - 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); } } diff --git a/OpenCL/m07600_a3.cl b/OpenCL/m07600_a3.cl index 2b6cabab8..048572621 100644 --- a/OpenCL/m07600_a3.cl +++ b/OpenCL/m07600_a3.cl @@ -49,34 +49,28 @@ static void m07600m (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 pw_le */ u32 salt_buf0[4]; + u32 salt_buf1[4]; + u32 salt_buf2[4]; + u32 salt_buf3[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]; salt_buf1[3] = salt_bufs[salt_pos].salt_buf[ 7]; + salt_buf2[0] = salt_bufs[salt_pos].salt_buf[ 8]; + salt_buf2[1] = salt_bufs[salt_pos].salt_buf[ 9]; + salt_buf2[2] = salt_bufs[salt_pos].salt_buf[10]; + salt_buf2[3] = salt_bufs[salt_pos].salt_buf[11]; + salt_buf3[0] = salt_bufs[salt_pos].salt_buf[12]; + salt_buf3[1] = salt_bufs[salt_pos].salt_buf[13]; + salt_buf3[2] = salt_bufs[salt_pos].salt_buf[14]; + salt_buf3[3] = salt_bufs[salt_pos].salt_buf[15]; - u32 salt_buf2[4]; - - salt_buf2[0] = 0; - salt_buf2[1] = 0; - salt_buf2[2] = 0; - salt_buf2[3] = 0; - - u32 salt_buf3[4]; - - salt_buf3[0] = 0; - salt_buf3[1] = 0; - salt_buf3[2] = 0; - salt_buf3[3] = 0; - - const u32 total_len = (32 + 40) * 8; + const u32 salt_len = salt_bufs[salt_pos].salt_len; /** * loop @@ -223,35 +217,6 @@ static void m07600m (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 pw_le * Prepend salt */ - u32x w0t[4]; - - w0t[0] = uint_to_hex_lower8 ((a >> 24) & 255) << 0 - | uint_to_hex_lower8 ((a >> 16) & 255) << 16; - w0t[1] = uint_to_hex_lower8 ((a >> 8) & 255) << 0 - | uint_to_hex_lower8 ((a >> 0) & 255) << 16; - w0t[2] = uint_to_hex_lower8 ((b >> 24) & 255) << 0 - | uint_to_hex_lower8 ((b >> 16) & 255) << 16; - w0t[3] = uint_to_hex_lower8 ((b >> 8) & 255) << 0 - | uint_to_hex_lower8 ((b >> 0) & 255) << 16; - - u32x w1t[4]; - - w1t[0] = uint_to_hex_lower8 ((c >> 24) & 255) << 0 - | uint_to_hex_lower8 ((c >> 16) & 255) << 16; - w1t[1] = uint_to_hex_lower8 ((c >> 8) & 255) << 0 - | uint_to_hex_lower8 ((c >> 0) & 255) << 16; - w1t[2] = uint_to_hex_lower8 ((d >> 24) & 255) << 0 - | uint_to_hex_lower8 ((d >> 16) & 255) << 16; - w1t[3] = uint_to_hex_lower8 ((d >> 8) & 255) << 0 - | uint_to_hex_lower8 ((d >> 0) & 255) << 16; - - u32x w2t[2]; - - w2t[0] = uint_to_hex_lower8 ((e >> 24) & 255) << 0 - | uint_to_hex_lower8 ((e >> 16) & 255) << 16; - w2t[1] = uint_to_hex_lower8 ((e >> 8) & 255) << 0 - | uint_to_hex_lower8 ((e >> 0) & 255) << 16; - w0_t = salt_buf0[0]; w1_t = salt_buf0[1]; w2_t = salt_buf0[2]; @@ -260,14 +225,24 @@ static void m07600m (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 pw_le w5_t = salt_buf1[1]; w6_t = salt_buf1[2]; w7_t = salt_buf1[3]; - w8_t = w0t[0]; - w9_t = w0t[1]; - wa_t = w0t[2]; - wb_t = w0t[3]; - wc_t = w1t[0]; - wd_t = w1t[1]; - we_t = w1t[2]; - wf_t = w1t[3]; + w8_t = uint_to_hex_lower8 ((a >> 24) & 255) << 0 + | uint_to_hex_lower8 ((a >> 16) & 255) << 16; + w9_t = uint_to_hex_lower8 ((a >> 8) & 255) << 0 + | uint_to_hex_lower8 ((a >> 0) & 255) << 16; + wa_t = uint_to_hex_lower8 ((b >> 24) & 255) << 0 + | uint_to_hex_lower8 ((b >> 16) & 255) << 16; + wb_t = uint_to_hex_lower8 ((b >> 8) & 255) << 0 + | uint_to_hex_lower8 ((b >> 0) & 255) << 16; + wc_t = uint_to_hex_lower8 ((c >> 24) & 255) << 0 + | uint_to_hex_lower8 ((c >> 16) & 255) << 16; + wd_t = uint_to_hex_lower8 ((c >> 8) & 255) << 0 + | uint_to_hex_lower8 ((c >> 0) & 255) << 16; + we_t = uint_to_hex_lower8 ((d >> 24) & 255) << 0 + | uint_to_hex_lower8 ((d >> 16) & 255) << 16; + wf_t = uint_to_hex_lower8 ((d >> 8) & 255) << 0 + | uint_to_hex_lower8 ((d >> 0) & 255) << 16; + + const u32x e_sav = e; /** * 2nd SHA1 @@ -408,8 +383,10 @@ static void m07600m (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 pw_le // 2nd transform - w0_t = swap32 (w2t[0]); - w1_t = swap32 (w2t[1]); + w0_t = uint_to_hex_lower8 ((e_sav >> 24) & 255) << 0 + | uint_to_hex_lower8 ((e_sav >> 16) & 255) << 16; + w1_t = uint_to_hex_lower8 ((e_sav >> 8) & 255) << 0 + | uint_to_hex_lower8 ((e_sav >> 0) & 255) << 16; w2_t = 0x80000000; w3_t = 0; w4_t = 0; @@ -423,7 +400,10 @@ static void m07600m (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 pw_le wc_t = 0; wd_t = 0; we_t = 0; - wf_t = total_len; + wf_t = (salt_len + 40) * 8; + + w0_t = swap32 (w0_t); + w1_t = swap32 (w1_t); #undef K #define K SHA1C00 @@ -545,34 +525,28 @@ static void m07600s (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 pw_le */ u32 salt_buf0[4]; + u32 salt_buf1[4]; + u32 salt_buf2[4]; + u32 salt_buf3[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]; salt_buf1[3] = salt_bufs[salt_pos].salt_buf[ 7]; + salt_buf2[0] = salt_bufs[salt_pos].salt_buf[ 8]; + salt_buf2[1] = salt_bufs[salt_pos].salt_buf[ 9]; + salt_buf2[2] = salt_bufs[salt_pos].salt_buf[10]; + salt_buf2[3] = salt_bufs[salt_pos].salt_buf[11]; + salt_buf3[0] = salt_bufs[salt_pos].salt_buf[12]; + salt_buf3[1] = salt_bufs[salt_pos].salt_buf[13]; + salt_buf3[2] = salt_bufs[salt_pos].salt_buf[14]; + salt_buf3[3] = salt_bufs[salt_pos].salt_buf[15]; - u32 salt_buf2[4]; - - salt_buf2[0] = 0; - salt_buf2[1] = 0; - salt_buf2[2] = 0; - salt_buf2[3] = 0; - - u32 salt_buf3[4]; - - salt_buf3[0] = 0; - salt_buf3[1] = 0; - salt_buf3[2] = 0; - salt_buf3[3] = 0; - - const u32 total_len = (32 + 40) * 8; + const u32 salt_len = salt_bufs[salt_pos].salt_len; /** * digest @@ -731,35 +705,6 @@ static void m07600s (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 pw_le * Prepend salt */ - u32x w0t[4]; - - w0t[0] = uint_to_hex_lower8 ((a >> 24) & 255) << 0 - | uint_to_hex_lower8 ((a >> 16) & 255) << 16; - w0t[1] = uint_to_hex_lower8 ((a >> 8) & 255) << 0 - | uint_to_hex_lower8 ((a >> 0) & 255) << 16; - w0t[2] = uint_to_hex_lower8 ((b >> 24) & 255) << 0 - | uint_to_hex_lower8 ((b >> 16) & 255) << 16; - w0t[3] = uint_to_hex_lower8 ((b >> 8) & 255) << 0 - | uint_to_hex_lower8 ((b >> 0) & 255) << 16; - - u32x w1t[4]; - - w1t[0] = uint_to_hex_lower8 ((c >> 24) & 255) << 0 - | uint_to_hex_lower8 ((c >> 16) & 255) << 16; - w1t[1] = uint_to_hex_lower8 ((c >> 8) & 255) << 0 - | uint_to_hex_lower8 ((c >> 0) & 255) << 16; - w1t[2] = uint_to_hex_lower8 ((d >> 24) & 255) << 0 - | uint_to_hex_lower8 ((d >> 16) & 255) << 16; - w1t[3] = uint_to_hex_lower8 ((d >> 8) & 255) << 0 - | uint_to_hex_lower8 ((d >> 0) & 255) << 16; - - u32x w2t[2]; - - w2t[0] = uint_to_hex_lower8 ((e >> 24) & 255) << 0 - | uint_to_hex_lower8 ((e >> 16) & 255) << 16; - w2t[1] = uint_to_hex_lower8 ((e >> 8) & 255) << 0 - | uint_to_hex_lower8 ((e >> 0) & 255) << 16; - w0_t = salt_buf0[0]; w1_t = salt_buf0[1]; w2_t = salt_buf0[2]; @@ -768,14 +713,24 @@ static void m07600s (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 pw_le w5_t = salt_buf1[1]; w6_t = salt_buf1[2]; w7_t = salt_buf1[3]; - w8_t = w0t[0]; - w9_t = w0t[1]; - wa_t = w0t[2]; - wb_t = w0t[3]; - wc_t = w1t[0]; - wd_t = w1t[1]; - we_t = w1t[2]; - wf_t = w1t[3]; + w8_t = uint_to_hex_lower8 ((a >> 24) & 255) << 0 + | uint_to_hex_lower8 ((a >> 16) & 255) << 16; + w9_t = uint_to_hex_lower8 ((a >> 8) & 255) << 0 + | uint_to_hex_lower8 ((a >> 0) & 255) << 16; + wa_t = uint_to_hex_lower8 ((b >> 24) & 255) << 0 + | uint_to_hex_lower8 ((b >> 16) & 255) << 16; + wb_t = uint_to_hex_lower8 ((b >> 8) & 255) << 0 + | uint_to_hex_lower8 ((b >> 0) & 255) << 16; + wc_t = uint_to_hex_lower8 ((c >> 24) & 255) << 0 + | uint_to_hex_lower8 ((c >> 16) & 255) << 16; + wd_t = uint_to_hex_lower8 ((c >> 8) & 255) << 0 + | uint_to_hex_lower8 ((c >> 0) & 255) << 16; + we_t = uint_to_hex_lower8 ((d >> 24) & 255) << 0 + | uint_to_hex_lower8 ((d >> 16) & 255) << 16; + wf_t = uint_to_hex_lower8 ((d >> 8) & 255) << 0 + | uint_to_hex_lower8 ((d >> 0) & 255) << 16; + + const u32x e_sav = e; /** * 2nd SHA1 @@ -916,8 +871,10 @@ static void m07600s (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 pw_le // 2nd transform - w0_t = swap32 (w2t[0]); - w1_t = swap32 (w2t[1]); + w0_t = uint_to_hex_lower8 ((e_sav >> 24) & 255) << 0 + | uint_to_hex_lower8 ((e_sav >> 16) & 255) << 16; + w1_t = uint_to_hex_lower8 ((e_sav >> 8) & 255) << 0 + | uint_to_hex_lower8 ((e_sav >> 0) & 255) << 16; w2_t = 0x80000000; w3_t = 0; w4_t = 0; @@ -931,7 +888,10 @@ static void m07600s (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const u32 pw_le wc_t = 0; wd_t = 0; we_t = 0; - wf_t = total_len; + wf_t = (salt_len + 40) * 8; + + w0_t = swap32 (w0_t); + w1_t = swap32 (w1_t); #undef K #define K SHA1C00