arno/hashcat
1
0
Fork 0
mirror of https://github.com/hashcat/hashcat.git synced 2024-12-16 11:48:15 +00:00
Code Issues Releases Wiki Activity
1ccb643fba
hashcat/OpenCL/m06000_a1.cl

527 lines
26 KiB
Common Lisp
Raw Normal View History

Initial commit
2015-12-04 14:47:52 +00:00
/**
* Author......: Jens Steube <jens.steube@gmail.com>
* License.....: MIT
*/
#define _RIPEMD160_
Cleanup -m 60xx kernels to latest standard
2016-04-13 10:58:22 +00:00
#define NEW_SIMD_CODE
Initial commit
2015-12-04 14:47:52 +00:00
#include "include/constants.h"
#include "include/kernel_vendor.h"
#define DGST_R0 0
#define DGST_R1 1
#define DGST_R2 2
#define DGST_R3 3
#include "include/kernel_functions.c"
Fix path to includes
2016-01-03 00:48:05 +00:00
#include "OpenCL/types_ocl.c"
#include "OpenCL/common.c"
Cleanup -m 60xx kernels to latest standard
2016-04-13 10:58:22 +00:00
#include "OpenCL/simd.c"
Revert "Zero pws_buf before reuse" This reverts commit b409e5e9e162927b25af88e672326b246f17ec5f.
2016-02-22 20:32:38 +00:00
Cleanup -m 60xx kernels to latest standard
2016-04-13 10:58:22 +00:00
static void ripemd160_transform (const u32x w[16], u32x dgst[5])
Initial commit
2015-12-04 14:47:52 +00:00
{
Cleanup -m 60xx kernels to latest standard
2016-04-13 10:58:22 +00:00
u32x a1 = dgst[0];
u32x b1 = dgst[1];
u32x c1 = dgst[2];
u32x d1 = dgst[3];
u32x e1 = dgst[4];
Initial commit
2015-12-04 14:47:52 +00:00
RIPEMD160_STEP (RIPEMD160_F , a1, b1, c1, d1, e1, w[ 0], RIPEMD160C00, RIPEMD160S00);
RIPEMD160_STEP (RIPEMD160_F , e1, a1, b1, c1, d1, w[ 1], RIPEMD160C00, RIPEMD160S01);
RIPEMD160_STEP (RIPEMD160_F , d1, e1, a1, b1, c1, w[ 2], RIPEMD160C00, RIPEMD160S02);
RIPEMD160_STEP (RIPEMD160_F , c1, d1, e1, a1, b1, w[ 3], RIPEMD160C00, RIPEMD160S03);
RIPEMD160_STEP (RIPEMD160_F , b1, c1, d1, e1, a1, w[ 4], RIPEMD160C00, RIPEMD160S04);
RIPEMD160_STEP (RIPEMD160_F , a1, b1, c1, d1, e1, w[ 5], RIPEMD160C00, RIPEMD160S05);
RIPEMD160_STEP (RIPEMD160_F , e1, a1, b1, c1, d1, w[ 6], RIPEMD160C00, RIPEMD160S06);
RIPEMD160_STEP (RIPEMD160_F , d1, e1, a1, b1, c1, w[ 7], RIPEMD160C00, RIPEMD160S07);
RIPEMD160_STEP (RIPEMD160_F , c1, d1, e1, a1, b1, w[ 8], RIPEMD160C00, RIPEMD160S08);
RIPEMD160_STEP (RIPEMD160_F , b1, c1, d1, e1, a1, w[ 9], RIPEMD160C00, RIPEMD160S09);
RIPEMD160_STEP (RIPEMD160_F , a1, b1, c1, d1, e1, w[10], RIPEMD160C00, RIPEMD160S0A);
RIPEMD160_STEP (RIPEMD160_F , e1, a1, b1, c1, d1, w[11], RIPEMD160C00, RIPEMD160S0B);
RIPEMD160_STEP (RIPEMD160_F , d1, e1, a1, b1, c1, w[12], RIPEMD160C00, RIPEMD160S0C);
RIPEMD160_STEP (RIPEMD160_F , c1, d1, e1, a1, b1, w[13], RIPEMD160C00, RIPEMD160S0D);
RIPEMD160_STEP (RIPEMD160_F , b1, c1, d1, e1, a1, w[14], RIPEMD160C00, RIPEMD160S0E);
RIPEMD160_STEP (RIPEMD160_F , a1, b1, c1, d1, e1, w[15], RIPEMD160C00, RIPEMD160S0F);
RIPEMD160_STEP (RIPEMD160_Go, e1, a1, b1, c1, d1, w[ 7], RIPEMD160C10, RIPEMD160S10);
RIPEMD160_STEP (RIPEMD160_Go, d1, e1, a1, b1, c1, w[ 4], RIPEMD160C10, RIPEMD160S11);
RIPEMD160_STEP (RIPEMD160_Go, c1, d1, e1, a1, b1, w[13], RIPEMD160C10, RIPEMD160S12);
RIPEMD160_STEP (RIPEMD160_Go, b1, c1, d1, e1, a1, w[ 1], RIPEMD160C10, RIPEMD160S13);
RIPEMD160_STEP (RIPEMD160_Go, a1, b1, c1, d1, e1, w[10], RIPEMD160C10, RIPEMD160S14);
RIPEMD160_STEP (RIPEMD160_Go, e1, a1, b1, c1, d1, w[ 6], RIPEMD160C10, RIPEMD160S15);
RIPEMD160_STEP (RIPEMD160_Go, d1, e1, a1, b1, c1, w[15], RIPEMD160C10, RIPEMD160S16);
RIPEMD160_STEP (RIPEMD160_Go, c1, d1, e1, a1, b1, w[ 3], RIPEMD160C10, RIPEMD160S17);
RIPEMD160_STEP (RIPEMD160_Go, b1, c1, d1, e1, a1, w[12], RIPEMD160C10, RIPEMD160S18);
RIPEMD160_STEP (RIPEMD160_Go, a1, b1, c1, d1, e1, w[ 0], RIPEMD160C10, RIPEMD160S19);
RIPEMD160_STEP (RIPEMD160_Go, e1, a1, b1, c1, d1, w[ 9], RIPEMD160C10, RIPEMD160S1A);
RIPEMD160_STEP (RIPEMD160_Go, d1, e1, a1, b1, c1, w[ 5], RIPEMD160C10, RIPEMD160S1B);
RIPEMD160_STEP (RIPEMD160_Go, c1, d1, e1, a1, b1, w[ 2], RIPEMD160C10, RIPEMD160S1C);
RIPEMD160_STEP (RIPEMD160_Go, b1, c1, d1, e1, a1, w[14], RIPEMD160C10, RIPEMD160S1D);
RIPEMD160_STEP (RIPEMD160_Go, a1, b1, c1, d1, e1, w[11], RIPEMD160C10, RIPEMD160S1E);
RIPEMD160_STEP (RIPEMD160_Go, e1, a1, b1, c1, d1, w[ 8], RIPEMD160C10, RIPEMD160S1F);
RIPEMD160_STEP (RIPEMD160_H , d1, e1, a1, b1, c1, w[ 3], RIPEMD160C20, RIPEMD160S20);
RIPEMD160_STEP (RIPEMD160_H , c1, d1, e1, a1, b1, w[10], RIPEMD160C20, RIPEMD160S21);
RIPEMD160_STEP (RIPEMD160_H , b1, c1, d1, e1, a1, w[14], RIPEMD160C20, RIPEMD160S22);
RIPEMD160_STEP (RIPEMD160_H , a1, b1, c1, d1, e1, w[ 4], RIPEMD160C20, RIPEMD160S23);
RIPEMD160_STEP (RIPEMD160_H , e1, a1, b1, c1, d1, w[ 9], RIPEMD160C20, RIPEMD160S24);
RIPEMD160_STEP (RIPEMD160_H , d1, e1, a1, b1, c1, w[15], RIPEMD160C20, RIPEMD160S25);
RIPEMD160_STEP (RIPEMD160_H , c1, d1, e1, a1, b1, w[ 8], RIPEMD160C20, RIPEMD160S26);
RIPEMD160_STEP (RIPEMD160_H , b1, c1, d1, e1, a1, w[ 1], RIPEMD160C20, RIPEMD160S27);
RIPEMD160_STEP (RIPEMD160_H , a1, b1, c1, d1, e1, w[ 2], RIPEMD160C20, RIPEMD160S28);
RIPEMD160_STEP (RIPEMD160_H , e1, a1, b1, c1, d1, w[ 7], RIPEMD160C20, RIPEMD160S29);
RIPEMD160_STEP (RIPEMD160_H , d1, e1, a1, b1, c1, w[ 0], RIPEMD160C20, RIPEMD160S2A);
RIPEMD160_STEP (RIPEMD160_H , c1, d1, e1, a1, b1, w[ 6], RIPEMD160C20, RIPEMD160S2B);
RIPEMD160_STEP (RIPEMD160_H , b1, c1, d1, e1, a1, w[13], RIPEMD160C20, RIPEMD160S2C);
RIPEMD160_STEP (RIPEMD160_H , a1, b1, c1, d1, e1, w[11], RIPEMD160C20, RIPEMD160S2D);
RIPEMD160_STEP (RIPEMD160_H , e1, a1, b1, c1, d1, w[ 5], RIPEMD160C20, RIPEMD160S2E);
RIPEMD160_STEP (RIPEMD160_H , d1, e1, a1, b1, c1, w[12], RIPEMD160C20, RIPEMD160S2F);
RIPEMD160_STEP (RIPEMD160_Io, c1, d1, e1, a1, b1, w[ 1], RIPEMD160C30, RIPEMD160S30);
RIPEMD160_STEP (RIPEMD160_Io, b1, c1, d1, e1, a1, w[ 9], RIPEMD160C30, RIPEMD160S31);
RIPEMD160_STEP (RIPEMD160_Io, a1, b1, c1, d1, e1, w[11], RIPEMD160C30, RIPEMD160S32);
RIPEMD160_STEP (RIPEMD160_Io, e1, a1, b1, c1, d1, w[10], RIPEMD160C30, RIPEMD160S33);
RIPEMD160_STEP (RIPEMD160_Io, d1, e1, a1, b1, c1, w[ 0], RIPEMD160C30, RIPEMD160S34);
RIPEMD160_STEP (RIPEMD160_Io, c1, d1, e1, a1, b1, w[ 8], RIPEMD160C30, RIPEMD160S35);
RIPEMD160_STEP (RIPEMD160_Io, b1, c1, d1, e1, a1, w[12], RIPEMD160C30, RIPEMD160S36);
RIPEMD160_STEP (RIPEMD160_Io, a1, b1, c1, d1, e1, w[ 4], RIPEMD160C30, RIPEMD160S37);
RIPEMD160_STEP (RIPEMD160_Io, e1, a1, b1, c1, d1, w[13], RIPEMD160C30, RIPEMD160S38);
RIPEMD160_STEP (RIPEMD160_Io, d1, e1, a1, b1, c1, w[ 3], RIPEMD160C30, RIPEMD160S39);
RIPEMD160_STEP (RIPEMD160_Io, c1, d1, e1, a1, b1, w[ 7], RIPEMD160C30, RIPEMD160S3A);
RIPEMD160_STEP (RIPEMD160_Io, b1, c1, d1, e1, a1, w[15], RIPEMD160C30, RIPEMD160S3B);
RIPEMD160_STEP (RIPEMD160_Io, a1, b1, c1, d1, e1, w[14], RIPEMD160C30, RIPEMD160S3C);
RIPEMD160_STEP (RIPEMD160_Io, e1, a1, b1, c1, d1, w[ 5], RIPEMD160C30, RIPEMD160S3D);
RIPEMD160_STEP (RIPEMD160_Io, d1, e1, a1, b1, c1, w[ 6], RIPEMD160C30, RIPEMD160S3E);
RIPEMD160_STEP (RIPEMD160_Io, c1, d1, e1, a1, b1, w[ 2], RIPEMD160C30, RIPEMD160S3F);
RIPEMD160_STEP (RIPEMD160_J , b1, c1, d1, e1, a1, w[ 4], RIPEMD160C40, RIPEMD160S40);
RIPEMD160_STEP (RIPEMD160_J , a1, b1, c1, d1, e1, w[ 0], RIPEMD160C40, RIPEMD160S41);
RIPEMD160_STEP (RIPEMD160_J , e1, a1, b1, c1, d1, w[ 5], RIPEMD160C40, RIPEMD160S42);
RIPEMD160_STEP (RIPEMD160_J , d1, e1, a1, b1, c1, w[ 9], RIPEMD160C40, RIPEMD160S43);
RIPEMD160_STEP (RIPEMD160_J , c1, d1, e1, a1, b1, w[ 7], RIPEMD160C40, RIPEMD160S44);
RIPEMD160_STEP (RIPEMD160_J , b1, c1, d1, e1, a1, w[12], RIPEMD160C40, RIPEMD160S45);
RIPEMD160_STEP (RIPEMD160_J , a1, b1, c1, d1, e1, w[ 2], RIPEMD160C40, RIPEMD160S46);
RIPEMD160_STEP (RIPEMD160_J , e1, a1, b1, c1, d1, w[10], RIPEMD160C40, RIPEMD160S47);
RIPEMD160_STEP (RIPEMD160_J , d1, e1, a1, b1, c1, w[14], RIPEMD160C40, RIPEMD160S48);
RIPEMD160_STEP (RIPEMD160_J , c1, d1, e1, a1, b1, w[ 1], RIPEMD160C40, RIPEMD160S49);
RIPEMD160_STEP (RIPEMD160_J , b1, c1, d1, e1, a1, w[ 3], RIPEMD160C40, RIPEMD160S4A);
RIPEMD160_STEP (RIPEMD160_J , a1, b1, c1, d1, e1, w[ 8], RIPEMD160C40, RIPEMD160S4B);
RIPEMD160_STEP (RIPEMD160_J , e1, a1, b1, c1, d1, w[11], RIPEMD160C40, RIPEMD160S4C);
RIPEMD160_STEP (RIPEMD160_J , d1, e1, a1, b1, c1, w[ 6], RIPEMD160C40, RIPEMD160S4D);
RIPEMD160_STEP (RIPEMD160_J , c1, d1, e1, a1, b1, w[15], RIPEMD160C40, RIPEMD160S4E);
RIPEMD160_STEP (RIPEMD160_J , b1, c1, d1, e1, a1, w[13], RIPEMD160C40, RIPEMD160S4F);
Cleanup -m 60xx kernels to latest standard
2016-04-13 10:58:22 +00:00
u32x a2 = dgst[0];
u32x b2 = dgst[1];
u32x c2 = dgst[2];
u32x d2 = dgst[3];
u32x e2 = dgst[4];
Initial commit
2015-12-04 14:47:52 +00:00
RIPEMD160_STEP_WORKAROUND_BUG (RIPEMD160_J , a2, b2, c2, d2, e2, w[ 5], RIPEMD160C50, RIPEMD160S50);
RIPEMD160_STEP (RIPEMD160_J , e2, a2, b2, c2, d2, w[14], RIPEMD160C50, RIPEMD160S51);
RIPEMD160_STEP (RIPEMD160_J , d2, e2, a2, b2, c2, w[ 7], RIPEMD160C50, RIPEMD160S52);
RIPEMD160_STEP (RIPEMD160_J , c2, d2, e2, a2, b2, w[ 0], RIPEMD160C50, RIPEMD160S53);
RIPEMD160_STEP (RIPEMD160_J , b2, c2, d2, e2, a2, w[ 9], RIPEMD160C50, RIPEMD160S54);
RIPEMD160_STEP (RIPEMD160_J , a2, b2, c2, d2, e2, w[ 2], RIPEMD160C50, RIPEMD160S55);
RIPEMD160_STEP (RIPEMD160_J , e2, a2, b2, c2, d2, w[11], RIPEMD160C50, RIPEMD160S56);
RIPEMD160_STEP (RIPEMD160_J , d2, e2, a2, b2, c2, w[ 4], RIPEMD160C50, RIPEMD160S57);
RIPEMD160_STEP (RIPEMD160_J , c2, d2, e2, a2, b2, w[13], RIPEMD160C50, RIPEMD160S58);
RIPEMD160_STEP (RIPEMD160_J , b2, c2, d2, e2, a2, w[ 6], RIPEMD160C50, RIPEMD160S59);
RIPEMD160_STEP (RIPEMD160_J , a2, b2, c2, d2, e2, w[15], RIPEMD160C50, RIPEMD160S5A);
RIPEMD160_STEP (RIPEMD160_J , e2, a2, b2, c2, d2, w[ 8], RIPEMD160C50, RIPEMD160S5B);
RIPEMD160_STEP (RIPEMD160_J , d2, e2, a2, b2, c2, w[ 1], RIPEMD160C50, RIPEMD160S5C);
RIPEMD160_STEP (RIPEMD160_J , c2, d2, e2, a2, b2, w[10], RIPEMD160C50, RIPEMD160S5D);
RIPEMD160_STEP (RIPEMD160_J , b2, c2, d2, e2, a2, w[ 3], RIPEMD160C50, RIPEMD160S5E);
RIPEMD160_STEP (RIPEMD160_J , a2, b2, c2, d2, e2, w[12], RIPEMD160C50, RIPEMD160S5F);
RIPEMD160_STEP (RIPEMD160_Io, e2, a2, b2, c2, d2, w[ 6], RIPEMD160C60, RIPEMD160S60);
RIPEMD160_STEP (RIPEMD160_Io, d2, e2, a2, b2, c2, w[11], RIPEMD160C60, RIPEMD160S61);
RIPEMD160_STEP (RIPEMD160_Io, c2, d2, e2, a2, b2, w[ 3], RIPEMD160C60, RIPEMD160S62);
RIPEMD160_STEP (RIPEMD160_Io, b2, c2, d2, e2, a2, w[ 7], RIPEMD160C60, RIPEMD160S63);
RIPEMD160_STEP (RIPEMD160_Io, a2, b2, c2, d2, e2, w[ 0], RIPEMD160C60, RIPEMD160S64);
RIPEMD160_STEP (RIPEMD160_Io, e2, a2, b2, c2, d2, w[13], RIPEMD160C60, RIPEMD160S65);
RIPEMD160_STEP (RIPEMD160_Io, d2, e2, a2, b2, c2, w[ 5], RIPEMD160C60, RIPEMD160S66);
RIPEMD160_STEP (RIPEMD160_Io, c2, d2, e2, a2, b2, w[10], RIPEMD160C60, RIPEMD160S67);
RIPEMD160_STEP (RIPEMD160_Io, b2, c2, d2, e2, a2, w[14], RIPEMD160C60, RIPEMD160S68);
RIPEMD160_STEP (RIPEMD160_Io, a2, b2, c2, d2, e2, w[15], RIPEMD160C60, RIPEMD160S69);
RIPEMD160_STEP (RIPEMD160_Io, e2, a2, b2, c2, d2, w[ 8], RIPEMD160C60, RIPEMD160S6A);
RIPEMD160_STEP (RIPEMD160_Io, d2, e2, a2, b2, c2, w[12], RIPEMD160C60, RIPEMD160S6B);
RIPEMD160_STEP (RIPEMD160_Io, c2, d2, e2, a2, b2, w[ 4], RIPEMD160C60, RIPEMD160S6C);
RIPEMD160_STEP (RIPEMD160_Io, b2, c2, d2, e2, a2, w[ 9], RIPEMD160C60, RIPEMD160S6D);
RIPEMD160_STEP (RIPEMD160_Io, a2, b2, c2, d2, e2, w[ 1], RIPEMD160C60, RIPEMD160S6E);
RIPEMD160_STEP (RIPEMD160_Io, e2, a2, b2, c2, d2, w[ 2], RIPEMD160C60, RIPEMD160S6F);
RIPEMD160_STEP (RIPEMD160_H , d2, e2, a2, b2, c2, w[15], RIPEMD160C70, RIPEMD160S70);
RIPEMD160_STEP (RIPEMD160_H , c2, d2, e2, a2, b2, w[ 5], RIPEMD160C70, RIPEMD160S71);
RIPEMD160_STEP (RIPEMD160_H , b2, c2, d2, e2, a2, w[ 1], RIPEMD160C70, RIPEMD160S72);
RIPEMD160_STEP (RIPEMD160_H , a2, b2, c2, d2, e2, w[ 3], RIPEMD160C70, RIPEMD160S73);
RIPEMD160_STEP (RIPEMD160_H , e2, a2, b2, c2, d2, w[ 7], RIPEMD160C70, RIPEMD160S74);
RIPEMD160_STEP (RIPEMD160_H , d2, e2, a2, b2, c2, w[14], RIPEMD160C70, RIPEMD160S75);
RIPEMD160_STEP (RIPEMD160_H , c2, d2, e2, a2, b2, w[ 6], RIPEMD160C70, RIPEMD160S76);
RIPEMD160_STEP (RIPEMD160_H , b2, c2, d2, e2, a2, w[ 9], RIPEMD160C70, RIPEMD160S77);
RIPEMD160_STEP (RIPEMD160_H , a2, b2, c2, d2, e2, w[11], RIPEMD160C70, RIPEMD160S78);
RIPEMD160_STEP (RIPEMD160_H , e2, a2, b2, c2, d2, w[ 8], RIPEMD160C70, RIPEMD160S79);
RIPEMD160_STEP (RIPEMD160_H , d2, e2, a2, b2, c2, w[12], RIPEMD160C70, RIPEMD160S7A);
RIPEMD160_STEP (RIPEMD160_H , c2, d2, e2, a2, b2, w[ 2], RIPEMD160C70, RIPEMD160S7B);
RIPEMD160_STEP (RIPEMD160_H , b2, c2, d2, e2, a2, w[10], RIPEMD160C70, RIPEMD160S7C);
RIPEMD160_STEP (RIPEMD160_H , a2, b2, c2, d2, e2, w[ 0], RIPEMD160C70, RIPEMD160S7D);
RIPEMD160_STEP (RIPEMD160_H , e2, a2, b2, c2, d2, w[ 4], RIPEMD160C70, RIPEMD160S7E);
RIPEMD160_STEP (RIPEMD160_H , d2, e2, a2, b2, c2, w[13], RIPEMD160C70, RIPEMD160S7F);
RIPEMD160_STEP (RIPEMD160_Go, c2, d2, e2, a2, b2, w[ 8], RIPEMD160C80, RIPEMD160S80);
RIPEMD160_STEP (RIPEMD160_Go, b2, c2, d2, e2, a2, w[ 6], RIPEMD160C80, RIPEMD160S81);
RIPEMD160_STEP (RIPEMD160_Go, a2, b2, c2, d2, e2, w[ 4], RIPEMD160C80, RIPEMD160S82);
RIPEMD160_STEP (RIPEMD160_Go, e2, a2, b2, c2, d2, w[ 1], RIPEMD160C80, RIPEMD160S83);
RIPEMD160_STEP (RIPEMD160_Go, d2, e2, a2, b2, c2, w[ 3], RIPEMD160C80, RIPEMD160S84);
RIPEMD160_STEP (RIPEMD160_Go, c2, d2, e2, a2, b2, w[11], RIPEMD160C80, RIPEMD160S85);
RIPEMD160_STEP (RIPEMD160_Go, b2, c2, d2, e2, a2, w[15], RIPEMD160C80, RIPEMD160S86);
RIPEMD160_STEP (RIPEMD160_Go, a2, b2, c2, d2, e2, w[ 0], RIPEMD160C80, RIPEMD160S87);
RIPEMD160_STEP (RIPEMD160_Go, e2, a2, b2, c2, d2, w[ 5], RIPEMD160C80, RIPEMD160S88);
RIPEMD160_STEP (RIPEMD160_Go, d2, e2, a2, b2, c2, w[12], RIPEMD160C80, RIPEMD160S89);
RIPEMD160_STEP (RIPEMD160_Go, c2, d2, e2, a2, b2, w[ 2], RIPEMD160C80, RIPEMD160S8A);
RIPEMD160_STEP (RIPEMD160_Go, b2, c2, d2, e2, a2, w[13], RIPEMD160C80, RIPEMD160S8B);
RIPEMD160_STEP (RIPEMD160_Go, a2, b2, c2, d2, e2, w[ 9], RIPEMD160C80, RIPEMD160S8C);
RIPEMD160_STEP (RIPEMD160_Go, e2, a2, b2, c2, d2, w[ 7], RIPEMD160C80, RIPEMD160S8D);
RIPEMD160_STEP (RIPEMD160_Go, d2, e2, a2, b2, c2, w[10], RIPEMD160C80, RIPEMD160S8E);
RIPEMD160_STEP (RIPEMD160_Go, c2, d2, e2, a2, b2, w[14], RIPEMD160C80, RIPEMD160S8F);
RIPEMD160_STEP (RIPEMD160_F , b2, c2, d2, e2, a2, w[12], RIPEMD160C90, RIPEMD160S90);
RIPEMD160_STEP (RIPEMD160_F , a2, b2, c2, d2, e2, w[15], RIPEMD160C90, RIPEMD160S91);
RIPEMD160_STEP (RIPEMD160_F , e2, a2, b2, c2, d2, w[10], RIPEMD160C90, RIPEMD160S92);
RIPEMD160_STEP (RIPEMD160_F , d2, e2, a2, b2, c2, w[ 4], RIPEMD160C90, RIPEMD160S93);
RIPEMD160_STEP (RIPEMD160_F , c2, d2, e2, a2, b2, w[ 1], RIPEMD160C90, RIPEMD160S94);
RIPEMD160_STEP (RIPEMD160_F , b2, c2, d2, e2, a2, w[ 5], RIPEMD160C90, RIPEMD160S95);
RIPEMD160_STEP (RIPEMD160_F , a2, b2, c2, d2, e2, w[ 8], RIPEMD160C90, RIPEMD160S96);
RIPEMD160_STEP (RIPEMD160_F , e2, a2, b2, c2, d2, w[ 7], RIPEMD160C90, RIPEMD160S97);
RIPEMD160_STEP (RIPEMD160_F , d2, e2, a2, b2, c2, w[ 6], RIPEMD160C90, RIPEMD160S98);
RIPEMD160_STEP (RIPEMD160_F , c2, d2, e2, a2, b2, w[ 2], RIPEMD160C90, RIPEMD160S99);
RIPEMD160_STEP (RIPEMD160_F , b2, c2, d2, e2, a2, w[13], RIPEMD160C90, RIPEMD160S9A);
RIPEMD160_STEP (RIPEMD160_F , a2, b2, c2, d2, e2, w[14], RIPEMD160C90, RIPEMD160S9B);
RIPEMD160_STEP (RIPEMD160_F , e2, a2, b2, c2, d2, w[ 0], RIPEMD160C90, RIPEMD160S9C);
RIPEMD160_STEP (RIPEMD160_F , d2, e2, a2, b2, c2, w[ 3], RIPEMD160C90, RIPEMD160S9D);
RIPEMD160_STEP (RIPEMD160_F , c2, d2, e2, a2, b2, w[ 9], RIPEMD160C90, RIPEMD160S9E);
RIPEMD160_STEP (RIPEMD160_F , b2, c2, d2, e2, a2, w[11], RIPEMD160C90, RIPEMD160S9F);
Cleanup -m 60xx kernels to latest standard
2016-04-13 10:58:22 +00:00
const u32x a = dgst[1] + c1 + d2;
const u32x b = dgst[2] + d1 + e2;
const u32x c = dgst[3] + e1 + a2;
const u32x d = dgst[4] + a1 + b2;
const u32x e = dgst[0] + b1 + c2;
Initial commit
2015-12-04 14:47:52 +00:00
dgst[0] = a;
dgst[1] = b;
dgst[2] = c;
dgst[3] = d;
dgst[4] = e;
}
Fixed two major problems 1) SIMD code for all attack-mode Macro vector_accessible() was not refactored and missing completely. Had to rename variables rules_cnt, combs_cnt and bfs_cnt into il_cnt which was a good thing anyway as with new SIMD code they all act in the same way. 2) SIMD code for attack-mode 0 With new SIMD code, apply_rules_vect() has to return u32 not u32x. This has massive impact on all *_a0 kernels. I've rewritten most of them. Deep testing using test.sh is still required. Some kernel need more fixes: - Some are kind of completely incompatible like m10400 but they still use old check_* includes, we should get rid of them as they are no longer neccessary as we have simd.c - Some have a chance but require additional effort like m11500. We can use commented out "#define NEW_SIMD_CODE" to find them This change can have negative impact on -a0 performance for device that require vectorization. That is mostly CPU devices. New GPU's are all scalar, so they wont get hurt by this. This change also proofes that there's no way to efficiently vectorize kernel rules with new SIMD code, but it enables the addition of the rule functions like @ that we were missing for some long time. This is a TODO.
2016-02-27 16:18:54 +00:00
__kernel void m06000_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 netntlm_t *netntlm_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)
Initial commit
2015-12-04 14:47:52 +00:00
{
/**
* modifier
*/
const u32 lid = get_local_id (0);
/**
* base
*/
const u32 gid = get_global_id (0);
if (gid >= gid_max) return;
Cleanup -m 60xx kernels to latest standard
2016-04-13 10:58:22 +00:00
u32 pw_buf0[4];
u32 pw_buf1[4];
Revert "Zero pws_buf before reuse" This reverts commit b409e5e9e162927b25af88e672326b246f17ec5f.
2016-02-22 20:32:38 +00:00
Cleanup -m 60xx kernels to latest standard
2016-04-13 10:58:22 +00:00
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];
Initial commit
2015-12-04 14:47:52 +00:00
const u32 pw_l_len = pws[gid].pw_len;
/**
* loop
*/
Cleanup -m 60xx kernels to latest standard
2016-04-13 10:58:22 +00:00
for (u32 il_pos = 0; il_pos < il_cnt; il_pos += VECT_SIZE)
Initial commit
2015-12-04 14:47:52 +00:00
{
Cleanup -m 60xx kernels to latest standard
2016-04-13 10:58:22 +00:00
const u32x pw_r_len = pwlenx_create_combt (combs_buf, il_pos);
const u32x pw_len = pw_l_len + pw_r_len;
/**
* concat password candidate
*/
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);
Initial commit
2015-12-04 14:47:52 +00:00
if (combs_mode == COMBINATOR_MODE_BASE_LEFT)
{
Cleanup -m 60xx kernels to latest standard
2016-04-13 10:58:22 +00:00
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);
Initial commit
2015-12-04 14:47:52 +00:00
}
Cleanup -m 60xx kernels to latest standard
2016-04-13 10:58:22 +00:00
u32x w0[4];
u32x w1[4];
u32x w2[4];
u32x w3[4];
Initial commit
2015-12-04 14:47:52 +00:00
w0[0] = wordl0[0] | wordr0[0];
w0[1] = wordl0[1] | wordr0[1];
w0[2] = wordl0[2] | wordr0[2];
w0[3] = wordl0[3] | wordr0[3];
w1[0] = wordl1[0] | wordr1[0];
w1[1] = wordl1[1] | wordr1[1];
w1[2] = wordl1[2] | wordr1[2];
w1[3] = wordl1[3] | wordr1[3];
w2[0] = wordl2[0] | wordr2[0];
w2[1] = wordl2[1] | wordr2[1];
w2[2] = wordl2[2] | wordr2[2];
w2[3] = wordl2[3] | wordr2[3];
w3[0] = wordl3[0] | wordr3[0];
w3[1] = wordl3[1] | wordr3[1];
Cleanup -m 60xx kernels to latest standard
2016-04-13 10:58:22 +00:00
w3[2] = wordl3[2] | wordr3[2];
w3[3] = wordl3[3] | wordr3[3];
/**
* RipeMD160
*/
u32x w[16];
w[ 0] = w0[0];
w[ 1] = w0[1];
w[ 2] = w0[2];
w[ 3] = w0[3];
w[ 4] = w1[0];
w[ 5] = w1[1];
w[ 6] = w1[2];
w[ 7] = w1[3];
w[ 8] = w2[0];
w[ 9] = w2[1];
w[10] = w2[2];
w[11] = w2[3];
w[12] = w3[0];
w[14] = w3[1];
w[14] = pw_len * 8;
w[15] = 0;
u32x dgst[5];
Initial commit
2015-12-04 14:47:52 +00:00
dgst[0] = RIPEMD160M_A;
dgst[1] = RIPEMD160M_B;
dgst[2] = RIPEMD160M_C;
dgst[3] = RIPEMD160M_D;
dgst[4] = RIPEMD160M_E;
Cleanup -m 60xx kernels to latest standard
2016-04-13 10:58:22 +00:00
ripemd160_transform (w, dgst);
Initial commit
2015-12-04 14:47:52 +00:00
Cleanup -m 60xx kernels to latest standard
2016-04-13 10:58:22 +00:00
COMPARE_M_SIMD (dgst[0], dgst[1], dgst[2], dgst[3]);
Initial commit
2015-12-04 14:47:52 +00:00
}
}
Fixed two major problems 1) SIMD code for all attack-mode Macro vector_accessible() was not refactored and missing completely. Had to rename variables rules_cnt, combs_cnt and bfs_cnt into il_cnt which was a good thing anyway as with new SIMD code they all act in the same way. 2) SIMD code for attack-mode 0 With new SIMD code, apply_rules_vect() has to return u32 not u32x. This has massive impact on all *_a0 kernels. I've rewritten most of them. Deep testing using test.sh is still required. Some kernel need more fixes: - Some are kind of completely incompatible like m10400 but they still use old check_* includes, we should get rid of them as they are no longer neccessary as we have simd.c - Some have a chance but require additional effort like m11500. We can use commented out "#define NEW_SIMD_CODE" to find them This change can have negative impact on -a0 performance for device that require vectorization. That is mostly CPU devices. New GPU's are all scalar, so they wont get hurt by this. This change also proofes that there's no way to efficiently vectorize kernel rules with new SIMD code, but it enables the addition of the rule functions like @ that we were missing for some long time. This is a TODO.
2016-02-27 16:18:54 +00:00
__kernel void m06000_m08 (__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)
Initial commit
2015-12-04 14:47:52 +00:00
{
}
Fixed two major problems 1) SIMD code for all attack-mode Macro vector_accessible() was not refactored and missing completely. Had to rename variables rules_cnt, combs_cnt and bfs_cnt into il_cnt which was a good thing anyway as with new SIMD code they all act in the same way. 2) SIMD code for attack-mode 0 With new SIMD code, apply_rules_vect() has to return u32 not u32x. This has massive impact on all *_a0 kernels. I've rewritten most of them. Deep testing using test.sh is still required. Some kernel need more fixes: - Some are kind of completely incompatible like m10400 but they still use old check_* includes, we should get rid of them as they are no longer neccessary as we have simd.c - Some have a chance but require additional effort like m11500. We can use commented out "#define NEW_SIMD_CODE" to find them This change can have negative impact on -a0 performance for device that require vectorization. That is mostly CPU devices. New GPU's are all scalar, so they wont get hurt by this. This change also proofes that there's no way to efficiently vectorize kernel rules with new SIMD code, but it enables the addition of the rule functions like @ that we were missing for some long time. This is a TODO.
2016-02-27 16:18:54 +00:00
__kernel void m06000_m16 (__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)
Initial commit
2015-12-04 14:47:52 +00:00
{
}
Fixed two major problems 1) SIMD code for all attack-mode Macro vector_accessible() was not refactored and missing completely. Had to rename variables rules_cnt, combs_cnt and bfs_cnt into il_cnt which was a good thing anyway as with new SIMD code they all act in the same way. 2) SIMD code for attack-mode 0 With new SIMD code, apply_rules_vect() has to return u32 not u32x. This has massive impact on all *_a0 kernels. I've rewritten most of them. Deep testing using test.sh is still required. Some kernel need more fixes: - Some are kind of completely incompatible like m10400 but they still use old check_* includes, we should get rid of them as they are no longer neccessary as we have simd.c - Some have a chance but require additional effort like m11500. We can use commented out "#define NEW_SIMD_CODE" to find them This change can have negative impact on -a0 performance for device that require vectorization. That is mostly CPU devices. New GPU's are all scalar, so they wont get hurt by this. This change also proofes that there's no way to efficiently vectorize kernel rules with new SIMD code, but it enables the addition of the rule functions like @ that we were missing for some long time. This is a TODO.
2016-02-27 16:18:54 +00:00
__kernel void m06000_s04 (__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 netntlm_t *netntlm_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)
Initial commit
2015-12-04 14:47:52 +00:00
{
/**
* modifier
*/
const u32 lid = get_local_id (0);
/**
* base
*/
const u32 gid = get_global_id (0);
if (gid >= gid_max) return;
Cleanup -m 60xx kernels to latest standard
2016-04-13 10:58:22 +00:00
u32 pw_buf0[4];
u32 pw_buf1[4];
Revert "Zero pws_buf before reuse" This reverts commit b409e5e9e162927b25af88e672326b246f17ec5f.
2016-02-22 20:32:38 +00:00
Cleanup -m 60xx kernels to latest standard
2016-04-13 10:58:22 +00:00
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];
Initial commit
2015-12-04 14:47:52 +00:00
const u32 pw_l_len = pws[gid].pw_len;
/**
* digest
*/
const u32 search[4] =
{
digests_buf[digests_offset].digest_buf[DGST_R0],
digests_buf[digests_offset].digest_buf[DGST_R1],
digests_buf[digests_offset].digest_buf[DGST_R2],
digests_buf[digests_offset].digest_buf[DGST_R3]
};
/**
* loop
*/
Cleanup -m 60xx kernels to latest standard
2016-04-13 10:58:22 +00:00
for (u32 il_pos = 0; il_pos < il_cnt; il_pos += VECT_SIZE)
Initial commit
2015-12-04 14:47:52 +00:00
{
Cleanup -m 60xx kernels to latest standard
2016-04-13 10:58:22 +00:00
const u32x pw_r_len = pwlenx_create_combt (combs_buf, il_pos);
const u32x pw_len = pw_l_len + pw_r_len;
/**
* concat password candidate
*/
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);
Initial commit
2015-12-04 14:47:52 +00:00
if (combs_mode == COMBINATOR_MODE_BASE_LEFT)
{
Cleanup -m 60xx kernels to latest standard
2016-04-13 10:58:22 +00:00
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);
Initial commit
2015-12-04 14:47:52 +00:00
}
Cleanup -m 60xx kernels to latest standard
2016-04-13 10:58:22 +00:00
u32x w0[4];
u32x w1[4];
u32x w2[4];
u32x w3[4];
Initial commit
2015-12-04 14:47:52 +00:00
w0[0] = wordl0[0] | wordr0[0];
w0[1] = wordl0[1] | wordr0[1];
w0[2] = wordl0[2] | wordr0[2];
w0[3] = wordl0[3] | wordr0[3];
w1[0] = wordl1[0] | wordr1[0];
w1[1] = wordl1[1] | wordr1[1];
w1[2] = wordl1[2] | wordr1[2];
w1[3] = wordl1[3] | wordr1[3];
w2[0] = wordl2[0] | wordr2[0];
w2[1] = wordl2[1] | wordr2[1];
w2[2] = wordl2[2] | wordr2[2];
w2[3] = wordl2[3] | wordr2[3];
w3[0] = wordl3[0] | wordr3[0];
w3[1] = wordl3[1] | wordr3[1];
Cleanup -m 60xx kernels to latest standard
2016-04-13 10:58:22 +00:00
w3[2] = wordl3[2] | wordr3[2];
w3[3] = wordl3[3] | wordr3[3];
/**
* RipeMD160
*/
u32x w[16];
w[ 0] = w0[0];
w[ 1] = w0[1];
w[ 2] = w0[2];
w[ 3] = w0[3];
w[ 4] = w1[0];
w[ 5] = w1[1];
w[ 6] = w1[2];
w[ 7] = w1[3];
w[ 8] = w2[0];
w[ 9] = w2[1];
w[10] = w2[2];
w[11] = w2[3];
w[12] = w3[0];
w[14] = w3[1];
w[14] = pw_len * 8;
w[15] = 0;
u32x dgst[5];
Initial commit
2015-12-04 14:47:52 +00:00
dgst[0] = RIPEMD160M_A;
dgst[1] = RIPEMD160M_B;
dgst[2] = RIPEMD160M_C;
dgst[3] = RIPEMD160M_D;
dgst[4] = RIPEMD160M_E;
Cleanup -m 60xx kernels to latest standard
2016-04-13 10:58:22 +00:00
ripemd160_transform (w, dgst);
Initial commit
2015-12-04 14:47:52 +00:00
Cleanup -m 60xx kernels to latest standard
2016-04-13 10:58:22 +00:00
COMPARE_S_SIMD (dgst[0], dgst[1], dgst[2], dgst[3]);
Initial commit
2015-12-04 14:47:52 +00:00
}
}
Fixed two major problems 1) SIMD code for all attack-mode Macro vector_accessible() was not refactored and missing completely. Had to rename variables rules_cnt, combs_cnt and bfs_cnt into il_cnt which was a good thing anyway as with new SIMD code they all act in the same way. 2) SIMD code for attack-mode 0 With new SIMD code, apply_rules_vect() has to return u32 not u32x. This has massive impact on all *_a0 kernels. I've rewritten most of them. Deep testing using test.sh is still required. Some kernel need more fixes: - Some are kind of completely incompatible like m10400 but they still use old check_* includes, we should get rid of them as they are no longer neccessary as we have simd.c - Some have a chance but require additional effort like m11500. We can use commented out "#define NEW_SIMD_CODE" to find them This change can have negative impact on -a0 performance for device that require vectorization. That is mostly CPU devices. New GPU's are all scalar, so they wont get hurt by this. This change also proofes that there's no way to efficiently vectorize kernel rules with new SIMD code, but it enables the addition of the rule functions like @ that we were missing for some long time. This is a TODO.
2016-02-27 16:18:54 +00:00
__kernel void m06000_s08 (__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)
Initial commit
2015-12-04 14:47:52 +00:00
{
}
Fixed two major problems 1) SIMD code for all attack-mode Macro vector_accessible() was not refactored and missing completely. Had to rename variables rules_cnt, combs_cnt and bfs_cnt into il_cnt which was a good thing anyway as with new SIMD code they all act in the same way. 2) SIMD code for attack-mode 0 With new SIMD code, apply_rules_vect() has to return u32 not u32x. This has massive impact on all *_a0 kernels. I've rewritten most of them. Deep testing using test.sh is still required. Some kernel need more fixes: - Some are kind of completely incompatible like m10400 but they still use old check_* includes, we should get rid of them as they are no longer neccessary as we have simd.c - Some have a chance but require additional effort like m11500. We can use commented out "#define NEW_SIMD_CODE" to find them This change can have negative impact on -a0 performance for device that require vectorization. That is mostly CPU devices. New GPU's are all scalar, so they wont get hurt by this. This change also proofes that there's no way to efficiently vectorize kernel rules with new SIMD code, but it enables the addition of the rule functions like @ that we were missing for some long time. This is a TODO.
2016-02-27 16:18:54 +00:00
__kernel void m06000_s16 (__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)
Initial commit
2015-12-04 14:47:52 +00:00
{
}
Reference in New Issue Copy Permalink