diff --git a/OpenCL/m01440_a0.cl b/OpenCL/m01440_a0.cl index c490b3c06..183f6e1e6 100644 --- a/OpenCL/m01440_a0.cl +++ b/OpenCL/m01440_a0.cl @@ -5,6 +5,8 @@ #define _SHA256_ +#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" __kernel void m01440_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 rules_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max) { @@ -78,31 +78,14 @@ __kernel void m01440_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]; - u32 w1[4]; - u32 w2[4]; - u32 w3[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]; - w1[0] = pw_buf1[0]; - w1[1] = pw_buf1[1]; - w1[2] = pw_buf1[2]; - w1[3] = pw_buf1[3]; - w2[0] = 0; - w2[1] = 0; - w2[2] = 0; - w2[3] = 0; - 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); /** * prepend salt @@ -110,10 +93,10 @@ __kernel void m01440_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, const u32 out_salt_len = (out_len * 2) + salt_len; - u32 w0_t2[4]; - u32 w1_t2[4]; - u32 w2_t2[4]; - u32 w3_t2[4]; + u32x w0_t2[4]; + u32x w1_t2[4]; + u32x w2_t2[4]; + u32x w3_t2[4]; make_unicode (w0, w0_t2, w1_t2); make_unicode (w1, w2_t2, w3_t2); @@ -135,31 +118,31 @@ __kernel void m01440_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, * sha256 */ - u32 w0_t = swap32 (w0_t2[0]); - u32 w1_t = swap32 (w0_t2[1]); - u32 w2_t = swap32 (w0_t2[2]); - u32 w3_t = swap32 (w0_t2[3]); - u32 w4_t = swap32 (w1_t2[0]); - u32 w5_t = swap32 (w1_t2[1]); - u32 w6_t = swap32 (w1_t2[2]); - u32 w7_t = swap32 (w1_t2[3]); - u32 w8_t = swap32 (w2_t2[0]); - u32 w9_t = swap32 (w2_t2[1]); - u32 wa_t = swap32 (w2_t2[2]); - u32 wb_t = swap32 (w2_t2[3]); - u32 wc_t = swap32 (w3_t2[0]); - u32 wd_t = swap32 (w3_t2[1]); - u32 we_t = 0; - u32 wf_t = out_salt_len * 8; + u32x w0_t = swap32 (w0_t2[0]); + u32x w1_t = swap32 (w0_t2[1]); + u32x w2_t = swap32 (w0_t2[2]); + u32x w3_t = swap32 (w0_t2[3]); + u32x w4_t = swap32 (w1_t2[0]); + u32x w5_t = swap32 (w1_t2[1]); + u32x w6_t = swap32 (w1_t2[2]); + u32x w7_t = swap32 (w1_t2[3]); + u32x w8_t = swap32 (w2_t2[0]); + u32x w9_t = swap32 (w2_t2[1]); + u32x wa_t = swap32 (w2_t2[2]); + u32x wb_t = swap32 (w2_t2[3]); + u32x wc_t = swap32 (w3_t2[0]); + u32x wd_t = swap32 (w3_t2[1]); + u32x we_t = 0; + u32x wf_t = out_salt_len * 8; - u32 a = SHA256M_A; - u32 b = SHA256M_B; - u32 c = SHA256M_C; - u32 d = SHA256M_D; - u32 e = SHA256M_E; - u32 f = SHA256M_F; - u32 g = SHA256M_G; - u32 h = SHA256M_H; + u32x a = SHA256M_A; + u32x b = SHA256M_B; + u32x c = SHA256M_C; + u32x d = SHA256M_D; + u32x e = SHA256M_E; + u32x f = SHA256M_F; + u32x g = SHA256M_G; + u32x h = SHA256M_H; SHA256_STEP (SHA256_F0o, SHA256_F1o, a, b, c, d, e, f, g, h, w0_t, SHA256C00); SHA256_STEP (SHA256_F0o, SHA256_F1o, h, a, b, c, d, e, f, g, w1_t, SHA256C01); @@ -229,13 +212,7 @@ __kernel void m01440_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, we_t = SHA256_EXPAND (wc_t, w7_t, wf_t, we_t); SHA256_STEP (SHA256_F0o, SHA256_F1o, c, d, e, f, g, h, a, b, we_t, SHA256C3e); wf_t = SHA256_EXPAND (wd_t, w8_t, w0_t, wf_t); SHA256_STEP (SHA256_F0o, SHA256_F1o, b, c, d, e, f, g, h, a, wf_t, SHA256C3f); - - const u32 r0 = d; - const u32 r1 = h; - const u32 r2 = c; - const u32 r3 = g; - - #include COMPARE_M + COMPARE_M_SIMD (d, h, c, g); } } @@ -315,31 +292,14 @@ __kernel void m01440_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]; - u32 w1[4]; - u32 w2[4]; - u32 w3[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]; - w1[0] = pw_buf1[0]; - w1[1] = pw_buf1[1]; - w1[2] = pw_buf1[2]; - w1[3] = pw_buf1[3]; - w2[0] = 0; - w2[1] = 0; - w2[2] = 0; - w2[3] = 0; - 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); /** * prepend salt @@ -347,10 +307,10 @@ __kernel void m01440_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, const u32 out_salt_len = (out_len * 2) + salt_len; - u32 w0_t2[4]; - u32 w1_t2[4]; - u32 w2_t2[4]; - u32 w3_t2[4]; + u32x w0_t2[4]; + u32x w1_t2[4]; + u32x w2_t2[4]; + u32x w3_t2[4]; make_unicode (w0, w0_t2, w1_t2); make_unicode (w1, w2_t2, w3_t2); @@ -372,31 +332,31 @@ __kernel void m01440_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, * sha256 */ - u32 w0_t = swap32 (w0_t2[0]); - u32 w1_t = swap32 (w0_t2[1]); - u32 w2_t = swap32 (w0_t2[2]); - u32 w3_t = swap32 (w0_t2[3]); - u32 w4_t = swap32 (w1_t2[0]); - u32 w5_t = swap32 (w1_t2[1]); - u32 w6_t = swap32 (w1_t2[2]); - u32 w7_t = swap32 (w1_t2[3]); - u32 w8_t = swap32 (w2_t2[0]); - u32 w9_t = swap32 (w2_t2[1]); - u32 wa_t = swap32 (w2_t2[2]); - u32 wb_t = swap32 (w2_t2[3]); - u32 wc_t = swap32 (w3_t2[0]); - u32 wd_t = swap32 (w3_t2[1]); - u32 we_t = 0; - u32 wf_t = out_salt_len * 8; + u32x w0_t = swap32 (w0_t2[0]); + u32x w1_t = swap32 (w0_t2[1]); + u32x w2_t = swap32 (w0_t2[2]); + u32x w3_t = swap32 (w0_t2[3]); + u32x w4_t = swap32 (w1_t2[0]); + u32x w5_t = swap32 (w1_t2[1]); + u32x w6_t = swap32 (w1_t2[2]); + u32x w7_t = swap32 (w1_t2[3]); + u32x w8_t = swap32 (w2_t2[0]); + u32x w9_t = swap32 (w2_t2[1]); + u32x wa_t = swap32 (w2_t2[2]); + u32x wb_t = swap32 (w2_t2[3]); + u32x wc_t = swap32 (w3_t2[0]); + u32x wd_t = swap32 (w3_t2[1]); + u32x we_t = 0; + u32x wf_t = out_salt_len * 8; - u32 a = SHA256M_A; - u32 b = SHA256M_B; - u32 c = SHA256M_C; - u32 d = SHA256M_D; - u32 e = SHA256M_E; - u32 f = SHA256M_F; - u32 g = SHA256M_G; - u32 h = SHA256M_H; + u32x a = SHA256M_A; + u32x b = SHA256M_B; + u32x c = SHA256M_C; + u32x d = SHA256M_D; + u32x e = SHA256M_E; + u32x f = SHA256M_F; + u32x g = SHA256M_G; + u32x h = SHA256M_H; SHA256_STEP (SHA256_F0o, SHA256_F1o, a, b, c, d, e, f, g, h, w0_t, SHA256C00); SHA256_STEP (SHA256_F0o, SHA256_F1o, h, a, b, c, d, e, f, g, w1_t, SHA256C01); @@ -466,13 +426,7 @@ __kernel void m01440_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, we_t = SHA256_EXPAND (wc_t, w7_t, wf_t, we_t); SHA256_STEP (SHA256_F0o, SHA256_F1o, c, d, e, f, g, h, a, b, we_t, SHA256C3e); wf_t = SHA256_EXPAND (wd_t, w8_t, w0_t, wf_t); SHA256_STEP (SHA256_F0o, SHA256_F1o, b, c, d, e, f, g, h, a, wf_t, SHA256C3f); - - const u32 r0 = d; - const u32 r1 = h; - const u32 r2 = c; - const u32 r3 = g; - - #include COMPARE_S + COMPARE_S_SIMD (d, h, c, g); } }