2 * Author......: Jens Steube <jens.steube@gmail.com>
8 #include "include/constants.h"
9 #include "include/kernel_vendor.h"
24 #include "include/kernel_functions.c"
26 #include "common_nv.c"
29 #define VECT_COMPARE_S "check_single_vect1_comp4.c"
30 #define VECT_COMPARE_M "check_multi_vect1_comp4.c"
34 #define VECT_COMPARE_S "check_single_vect2_comp4.c"
35 #define VECT_COMPARE_M "check_multi_vect2_comp4.c"
39 #define VECT_COMPARE_S "check_single_vect4_comp4.c"
40 #define VECT_COMPARE_M "check_multi_vect4_comp4.c"
43 __device__ __constant__ comb_t c_combs[1024];
45 __device__ __constant__ u64 k_sha384[80] =
47 SHA384C00, SHA384C01, SHA384C02, SHA384C03,
48 SHA384C04, SHA384C05, SHA384C06, SHA384C07,
49 SHA384C08, SHA384C09, SHA384C0a, SHA384C0b,
50 SHA384C0c, SHA384C0d, SHA384C0e, SHA384C0f,
51 SHA384C10, SHA384C11, SHA384C12, SHA384C13,
52 SHA384C14, SHA384C15, SHA384C16, SHA384C17,
53 SHA384C18, SHA384C19, SHA384C1a, SHA384C1b,
54 SHA384C1c, SHA384C1d, SHA384C1e, SHA384C1f,
55 SHA384C20, SHA384C21, SHA384C22, SHA384C23,
56 SHA384C24, SHA384C25, SHA384C26, SHA384C27,
57 SHA384C28, SHA384C29, SHA384C2a, SHA384C2b,
58 SHA384C2c, SHA384C2d, SHA384C2e, SHA384C2f,
59 SHA384C30, SHA384C31, SHA384C32, SHA384C33,
60 SHA384C34, SHA384C35, SHA384C36, SHA384C37,
61 SHA384C38, SHA384C39, SHA384C3a, SHA384C3b,
62 SHA384C3c, SHA384C3d, SHA384C3e, SHA384C3f,
63 SHA384C40, SHA384C41, SHA384C42, SHA384C43,
64 SHA384C44, SHA384C45, SHA384C46, SHA384C47,
65 SHA384C48, SHA384C49, SHA384C4a, SHA384C4b,
66 SHA384C4c, SHA384C4d, SHA384C4e, SHA384C4f,
69 __device__ static void sha384_transform (const u32x w0[4], const u32x w1[4], const u32x w2[4], const u32x w3[4], u64x digest[8])
71 u64x w0_t = hl32_to_64 (w0[0], w0[1]);
72 u64x w1_t = hl32_to_64 (w0[2], w0[3]);
73 u64x w2_t = hl32_to_64 (w1[0], w1[1]);
74 u64x w3_t = hl32_to_64 (w1[2], w1[3]);
75 u64x w4_t = hl32_to_64 (w2[0], w2[1]);
76 u64x w5_t = hl32_to_64 (w2[2], w2[3]);
77 u64x w6_t = hl32_to_64 (w3[0], w3[1]);
86 u64x wf_t = hl32_to_64 (w3[2], w3[3]);
97 #define ROUND_EXPAND() \
99 w0_t = SHA384_EXPAND (we_t, w9_t, w1_t, w0_t); \
100 w1_t = SHA384_EXPAND (wf_t, wa_t, w2_t, w1_t); \
101 w2_t = SHA384_EXPAND (w0_t, wb_t, w3_t, w2_t); \
102 w3_t = SHA384_EXPAND (w1_t, wc_t, w4_t, w3_t); \
103 w4_t = SHA384_EXPAND (w2_t, wd_t, w5_t, w4_t); \
104 w5_t = SHA384_EXPAND (w3_t, we_t, w6_t, w5_t); \
105 w6_t = SHA384_EXPAND (w4_t, wf_t, w7_t, w6_t); \
106 w7_t = SHA384_EXPAND (w5_t, w0_t, w8_t, w7_t); \
107 w8_t = SHA384_EXPAND (w6_t, w1_t, w9_t, w8_t); \
108 w9_t = SHA384_EXPAND (w7_t, w2_t, wa_t, w9_t); \
109 wa_t = SHA384_EXPAND (w8_t, w3_t, wb_t, wa_t); \
110 wb_t = SHA384_EXPAND (w9_t, w4_t, wc_t, wb_t); \
111 wc_t = SHA384_EXPAND (wa_t, w5_t, wd_t, wc_t); \
112 wd_t = SHA384_EXPAND (wb_t, w6_t, we_t, wd_t); \
113 we_t = SHA384_EXPAND (wc_t, w7_t, wf_t, we_t); \
114 wf_t = SHA384_EXPAND (wd_t, w8_t, w0_t, wf_t); \
117 #define ROUND_STEP(i) \
119 SHA384_STEP (SHA384_F0o, SHA384_F1o, a, b, c, d, e, f, g, h, w0_t, k_sha384[i + 0]); \
120 SHA384_STEP (SHA384_F0o, SHA384_F1o, h, a, b, c, d, e, f, g, w1_t, k_sha384[i + 1]); \
121 SHA384_STEP (SHA384_F0o, SHA384_F1o, g, h, a, b, c, d, e, f, w2_t, k_sha384[i + 2]); \
122 SHA384_STEP (SHA384_F0o, SHA384_F1o, f, g, h, a, b, c, d, e, w3_t, k_sha384[i + 3]); \
123 SHA384_STEP (SHA384_F0o, SHA384_F1o, e, f, g, h, a, b, c, d, w4_t, k_sha384[i + 4]); \
124 SHA384_STEP (SHA384_F0o, SHA384_F1o, d, e, f, g, h, a, b, c, w5_t, k_sha384[i + 5]); \
125 SHA384_STEP (SHA384_F0o, SHA384_F1o, c, d, e, f, g, h, a, b, w6_t, k_sha384[i + 6]); \
126 SHA384_STEP (SHA384_F0o, SHA384_F1o, b, c, d, e, f, g, h, a, w7_t, k_sha384[i + 7]); \
127 SHA384_STEP (SHA384_F0o, SHA384_F1o, a, b, c, d, e, f, g, h, w8_t, k_sha384[i + 8]); \
128 SHA384_STEP (SHA384_F0o, SHA384_F1o, h, a, b, c, d, e, f, g, w9_t, k_sha384[i + 9]); \
129 SHA384_STEP (SHA384_F0o, SHA384_F1o, g, h, a, b, c, d, e, f, wa_t, k_sha384[i + 10]); \
130 SHA384_STEP (SHA384_F0o, SHA384_F1o, f, g, h, a, b, c, d, e, wb_t, k_sha384[i + 11]); \
131 SHA384_STEP (SHA384_F0o, SHA384_F1o, e, f, g, h, a, b, c, d, wc_t, k_sha384[i + 12]); \
132 SHA384_STEP (SHA384_F0o, SHA384_F1o, d, e, f, g, h, a, b, c, wd_t, k_sha384[i + 13]); \
133 SHA384_STEP (SHA384_F0o, SHA384_F1o, c, d, e, f, g, h, a, b, we_t, k_sha384[i + 14]); \
134 SHA384_STEP (SHA384_F0o, SHA384_F1o, b, c, d, e, f, g, h, a, wf_t, k_sha384[i + 15]); \
139 for (int i = 16; i < 80; i += 16)
141 ROUND_EXPAND (); ROUND_STEP (i);
165 extern "C" __global__ void __launch_bounds__ (256, 1) m10800_m04 (const pw_t *pws, const gpu_rule_t *rules_buf, const comb_t *combs_buf, const bf_t *bfs_buf, const void *tmps, void *hooks, const u32 *bitmaps_buf_s1_a, const u32 *bitmaps_buf_s1_b, const u32 *bitmaps_buf_s1_c, const u32 *bitmaps_buf_s1_d, const u32 *bitmaps_buf_s2_a, const u32 *bitmaps_buf_s2_b, const u32 *bitmaps_buf_s2_c, const u32 *bitmaps_buf_s2_d, plain_t *plains_buf, const digest_t *digests_buf, u32 *hashes_shown, const salt_t *salt_bufs, const void *esalt_bufs, u32 *d_return_buf, 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)
171 const u32 lid = threadIdx.x;
177 const u32 gid = (blockIdx.x * blockDim.x) + threadIdx.x;
179 if (gid >= gid_max) return;
183 wordl0[0] = pws[gid].i[ 0];
184 wordl0[1] = pws[gid].i[ 1];
185 wordl0[2] = pws[gid].i[ 2];
186 wordl0[3] = pws[gid].i[ 3];
190 wordl1[0] = pws[gid].i[ 4];
191 wordl1[1] = pws[gid].i[ 5];
192 wordl1[2] = pws[gid].i[ 6];
193 wordl1[3] = pws[gid].i[ 7];
209 const u32 pw_l_len = pws[gid].pw_len;
211 if (combs_mode == COMBINATOR_MODE_BASE_RIGHT)
213 append_0x80_2 (wordl0, wordl1, pw_l_len);
215 switch_buffer_by_offset (wordl0, wordl1, wordl2, wordl3, c_combs[0].pw_len);
222 for (u32 il_pos = 0; il_pos < combs_cnt; il_pos++)
224 const u32 pw_r_len = c_combs[il_pos].pw_len;
226 const u32 pw_len = pw_l_len + pw_r_len;
230 wordr0[0] = c_combs[il_pos].i[0];
231 wordr0[1] = c_combs[il_pos].i[1];
232 wordr0[2] = c_combs[il_pos].i[2];
233 wordr0[3] = c_combs[il_pos].i[3];
237 wordr1[0] = c_combs[il_pos].i[4];
238 wordr1[1] = c_combs[il_pos].i[5];
239 wordr1[2] = c_combs[il_pos].i[6];
240 wordr1[3] = c_combs[il_pos].i[7];
256 if (combs_mode == COMBINATOR_MODE_BASE_LEFT)
258 append_0x80_2 (wordr0, wordr1, pw_r_len);
260 switch_buffer_by_offset (wordr0, wordr1, wordr2, wordr3, pw_l_len);
268 w0[0] = wordl0[0] | wordr0[0];
269 w0[1] = wordl0[1] | wordr0[1];
270 w0[2] = wordl0[2] | wordr0[2];
271 w0[3] = wordl0[3] | wordr0[3];
272 w1[0] = wordl1[0] | wordr1[0];
273 w1[1] = wordl1[1] | wordr1[1];
274 w1[2] = wordl1[2] | wordr1[2];
275 w1[3] = wordl1[3] | wordr1[3];
276 w2[0] = wordl2[0] | wordr2[0];
277 w2[1] = wordl2[1] | wordr2[1];
278 w2[2] = wordl2[2] | wordr2[2];
279 w2[3] = wordl2[3] | wordr2[3];
280 w3[0] = wordl3[0] | wordr3[0];
281 w3[1] = wordl3[1] | wordr3[1];
282 w3[2] = wordl3[2] | wordr3[2];
283 w3[3] = wordl3[3] | wordr3[3];
294 w0_t[0] = swap_workaround (w0[0]);
295 w0_t[1] = swap_workaround (w0[1]);
296 w0_t[2] = swap_workaround (w0[2]);
297 w0_t[3] = swap_workaround (w0[3]);
298 w1_t[0] = swap_workaround (w1[0]);
299 w1_t[1] = swap_workaround (w1[1]);
300 w1_t[2] = swap_workaround (w1[2]);
301 w1_t[3] = swap_workaround (w1[3]);
302 w2_t[0] = swap_workaround (w2[0]);
303 w2_t[1] = swap_workaround (w2[1]);
304 w2_t[2] = swap_workaround (w2[2]);
305 w2_t[3] = swap_workaround (w2[3]);
306 w3_t[0] = swap_workaround (w3[0]);
307 w3_t[1] = swap_workaround (w3[1]);
309 w3_t[3] = pw_len * 8;
313 digest[0] = SHA384M_A;
314 digest[1] = SHA384M_B;
315 digest[2] = SHA384M_C;
316 digest[3] = SHA384M_D;
317 digest[4] = SHA384M_E;
318 digest[5] = SHA384M_F;
319 digest[6] = SHA384M_G;
320 digest[7] = SHA384M_H;
322 sha384_transform (w0_t, w1_t, w2_t, w3_t, digest);
324 const u32x r0 = l32_from_64 (digest[3]);
325 const u32x r1 = h32_from_64 (digest[3]);
326 const u32x r2 = l32_from_64 (digest[2]);
327 const u32x r3 = h32_from_64 (digest[2]);
329 #include VECT_COMPARE_M
333 extern "C" __global__ void __launch_bounds__ (256, 1) m10800_m08 (const pw_t *pws, const gpu_rule_t *rules_buf, const comb_t *combs_buf, const bf_t *bfs_buf, const void *tmps, void *hooks, const u32 *bitmaps_buf_s1_a, const u32 *bitmaps_buf_s1_b, const u32 *bitmaps_buf_s1_c, const u32 *bitmaps_buf_s1_d, const u32 *bitmaps_buf_s2_a, const u32 *bitmaps_buf_s2_b, const u32 *bitmaps_buf_s2_c, const u32 *bitmaps_buf_s2_d, plain_t *plains_buf, const digest_t *digests_buf, u32 *hashes_shown, const salt_t *salt_bufs, const void *esalt_bufs, u32 *d_return_buf, 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)
337 extern "C" __global__ void __launch_bounds__ (256, 1) m10800_m16 (const pw_t *pws, const gpu_rule_t *rules_buf, const comb_t *combs_buf, const bf_t *bfs_buf, const void *tmps, void *hooks, const u32 *bitmaps_buf_s1_a, const u32 *bitmaps_buf_s1_b, const u32 *bitmaps_buf_s1_c, const u32 *bitmaps_buf_s1_d, const u32 *bitmaps_buf_s2_a, const u32 *bitmaps_buf_s2_b, const u32 *bitmaps_buf_s2_c, const u32 *bitmaps_buf_s2_d, plain_t *plains_buf, const digest_t *digests_buf, u32 *hashes_shown, const salt_t *salt_bufs, const void *esalt_bufs, u32 *d_return_buf, 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)
341 extern "C" __global__ void __launch_bounds__ (256, 1) m10800_s04 (const pw_t *pws, const gpu_rule_t *rules_buf, const comb_t *combs_buf, const bf_t *bfs_buf, const void *tmps, void *hooks, const u32 *bitmaps_buf_s1_a, const u32 *bitmaps_buf_s1_b, const u32 *bitmaps_buf_s1_c, const u32 *bitmaps_buf_s1_d, const u32 *bitmaps_buf_s2_a, const u32 *bitmaps_buf_s2_b, const u32 *bitmaps_buf_s2_c, const u32 *bitmaps_buf_s2_d, plain_t *plains_buf, const digest_t *digests_buf, u32 *hashes_shown, const salt_t *salt_bufs, const void *esalt_bufs, u32 *d_return_buf, 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)
347 const u32 lid = threadIdx.x;
353 const u32 gid = (blockIdx.x * blockDim.x) + threadIdx.x;
355 if (gid >= gid_max) return;
359 wordl0[0] = pws[gid].i[ 0];
360 wordl0[1] = pws[gid].i[ 1];
361 wordl0[2] = pws[gid].i[ 2];
362 wordl0[3] = pws[gid].i[ 3];
366 wordl1[0] = pws[gid].i[ 4];
367 wordl1[1] = pws[gid].i[ 5];
368 wordl1[2] = pws[gid].i[ 6];
369 wordl1[3] = pws[gid].i[ 7];
385 const u32 pw_l_len = pws[gid].pw_len;
387 if (combs_mode == COMBINATOR_MODE_BASE_RIGHT)
389 append_0x80_2 (wordl0, wordl1, pw_l_len);
391 switch_buffer_by_offset (wordl0, wordl1, wordl2, wordl3, c_combs[0].pw_len);
398 const u32 search[4] =
400 digests_buf[digests_offset].digest_buf[DGST_R0],
401 digests_buf[digests_offset].digest_buf[DGST_R1],
402 digests_buf[digests_offset].digest_buf[DGST_R2],
403 digests_buf[digests_offset].digest_buf[DGST_R3]
410 for (u32 il_pos = 0; il_pos < combs_cnt; il_pos++)
412 const u32 pw_r_len = c_combs[il_pos].pw_len;
414 const u32 pw_len = pw_l_len + pw_r_len;
418 wordr0[0] = c_combs[il_pos].i[0];
419 wordr0[1] = c_combs[il_pos].i[1];
420 wordr0[2] = c_combs[il_pos].i[2];
421 wordr0[3] = c_combs[il_pos].i[3];
425 wordr1[0] = c_combs[il_pos].i[4];
426 wordr1[1] = c_combs[il_pos].i[5];
427 wordr1[2] = c_combs[il_pos].i[6];
428 wordr1[3] = c_combs[il_pos].i[7];
444 if (combs_mode == COMBINATOR_MODE_BASE_LEFT)
446 append_0x80_2 (wordr0, wordr1, pw_r_len);
448 switch_buffer_by_offset (wordr0, wordr1, wordr2, wordr3, pw_l_len);
456 w0[0] = wordl0[0] | wordr0[0];
457 w0[1] = wordl0[1] | wordr0[1];
458 w0[2] = wordl0[2] | wordr0[2];
459 w0[3] = wordl0[3] | wordr0[3];
460 w1[0] = wordl1[0] | wordr1[0];
461 w1[1] = wordl1[1] | wordr1[1];
462 w1[2] = wordl1[2] | wordr1[2];
463 w1[3] = wordl1[3] | wordr1[3];
464 w2[0] = wordl2[0] | wordr2[0];
465 w2[1] = wordl2[1] | wordr2[1];
466 w2[2] = wordl2[2] | wordr2[2];
467 w2[3] = wordl2[3] | wordr2[3];
468 w3[0] = wordl3[0] | wordr3[0];
469 w3[1] = wordl3[1] | wordr3[1];
470 w3[2] = wordl3[2] | wordr3[2];
471 w3[3] = wordl3[3] | wordr3[3];
482 w0_t[0] = swap_workaround (w0[0]);
483 w0_t[1] = swap_workaround (w0[1]);
484 w0_t[2] = swap_workaround (w0[2]);
485 w0_t[3] = swap_workaround (w0[3]);
486 w1_t[0] = swap_workaround (w1[0]);
487 w1_t[1] = swap_workaround (w1[1]);
488 w1_t[2] = swap_workaround (w1[2]);
489 w1_t[3] = swap_workaround (w1[3]);
490 w2_t[0] = swap_workaround (w2[0]);
491 w2_t[1] = swap_workaround (w2[1]);
492 w2_t[2] = swap_workaround (w2[2]);
493 w2_t[3] = swap_workaround (w2[3]);
494 w3_t[0] = swap_workaround (w3[0]);
495 w3_t[1] = swap_workaround (w3[1]);
497 w3_t[3] = pw_len * 8;
501 digest[0] = SHA384M_A;
502 digest[1] = SHA384M_B;
503 digest[2] = SHA384M_C;
504 digest[3] = SHA384M_D;
505 digest[4] = SHA384M_E;
506 digest[5] = SHA384M_F;
507 digest[6] = SHA384M_G;
508 digest[7] = SHA384M_H;
510 sha384_transform (w0_t, w1_t, w2_t, w3_t, digest);
512 const u32x r0 = l32_from_64 (digest[3]);
513 const u32x r1 = h32_from_64 (digest[3]);
514 const u32x r2 = l32_from_64 (digest[2]);
515 const u32x r3 = h32_from_64 (digest[2]);
517 #include VECT_COMPARE_S
521 extern "C" __global__ void __launch_bounds__ (256, 1) m10800_s08 (const pw_t *pws, const gpu_rule_t *rules_buf, const comb_t *combs_buf, const bf_t *bfs_buf, const void *tmps, void *hooks, const u32 *bitmaps_buf_s1_a, const u32 *bitmaps_buf_s1_b, const u32 *bitmaps_buf_s1_c, const u32 *bitmaps_buf_s1_d, const u32 *bitmaps_buf_s2_a, const u32 *bitmaps_buf_s2_b, const u32 *bitmaps_buf_s2_c, const u32 *bitmaps_buf_s2_d, plain_t *plains_buf, const digest_t *digests_buf, u32 *hashes_shown, const salt_t *salt_bufs, const void *esalt_bufs, u32 *d_return_buf, 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)
525 extern "C" __global__ void __launch_bounds__ (256, 1) m10800_s16 (const pw_t *pws, const gpu_rule_t *rules_buf, const comb_t *combs_buf, const bf_t *bfs_buf, const void *tmps, void *hooks, const u32 *bitmaps_buf_s1_a, const u32 *bitmaps_buf_s1_b, const u32 *bitmaps_buf_s1_c, const u32 *bitmaps_buf_s1_d, const u32 *bitmaps_buf_s2_a, const u32 *bitmaps_buf_s2_b, const u32 *bitmaps_buf_s2_c, const u32 *bitmaps_buf_s2_d, plain_t *plains_buf, const digest_t *digests_buf, u32 *hashes_shown, const salt_t *salt_bufs, const void *esalt_bufs, u32 *d_return_buf, 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)