/**
 * Author......: See docs/credits.txt
 * License.....: MIT
 */

#define NEW_SIMD_CODE

#ifdef KERNEL_STATIC
#include M2S(INCLUDE_PATH/inc_vendor.h)
#include M2S(INCLUDE_PATH/inc_types.h)
#include M2S(INCLUDE_PATH/inc_platform.cl)
#include M2S(INCLUDE_PATH/inc_common.cl)
#include M2S(INCLUDE_PATH/inc_simd.cl)
#include M2S(INCLUDE_PATH/inc_hash_whirlpool.cl)
#endif

KERNEL_FQ void m06100_mxx (KERN_ATTR_VECTOR ())
{
  /**
   * modifier
   */

  const u64 gid = get_global_id (0);
  const u64 lid = get_local_id (0);
  const u64 lsz = get_local_size (0);

  /**
   * Whirlpool shared
   */

  #ifdef REAL_SHM

  LOCAL_VK u64 s_MT0[256];
  LOCAL_VK u64 s_MT1[256];
  LOCAL_VK u64 s_MT2[256];
  LOCAL_VK u64 s_MT3[256];
  LOCAL_VK u64 s_MT4[256];
  LOCAL_VK u64 s_MT5[256];
  LOCAL_VK u64 s_MT6[256];
  LOCAL_VK u64 s_MT7[256];

  for (u32 i = lid; i < 256; i += lsz)
  {
    s_MT0[i] = MT0[i];
    s_MT1[i] = MT1[i];
    s_MT2[i] = MT2[i];
    s_MT3[i] = MT3[i];
    s_MT4[i] = MT4[i];
    s_MT5[i] = MT5[i];
    s_MT6[i] = MT6[i];
    s_MT7[i] = MT7[i];
  }

  SYNC_THREADS ();

  #else

  CONSTANT_AS u64a *s_MT0 = MT0;
  CONSTANT_AS u64a *s_MT1 = MT1;
  CONSTANT_AS u64a *s_MT2 = MT2;
  CONSTANT_AS u64a *s_MT3 = MT3;
  CONSTANT_AS u64a *s_MT4 = MT4;
  CONSTANT_AS u64a *s_MT5 = MT5;
  CONSTANT_AS u64a *s_MT6 = MT6;
  CONSTANT_AS u64a *s_MT7 = MT7;

  #endif

  if (gid >= GID_CNT) return;

  /**
   * base
   */

  const u32 pw_len = pws[gid].pw_len;

  u32x w[64] = { 0 };

  for (u32 i = 0, idx = 0; i < pw_len; i += 4, idx += 1)
  {
    w[idx] = pws[gid].i[idx];
  }

  /**
   * loop
   */

  u32x w0l = w[0];

  for (u32 il_pos = 0; il_pos < IL_CNT; il_pos += VECT_SIZE)
  {
    const u32x w0r = words_buf_r[il_pos / VECT_SIZE];

    const u32x w0 = w0l | w0r;

    w[0] = w0;

    whirlpool_ctx_vector_t ctx;

    whirlpool_init_vector (&ctx, s_MT0, s_MT1, s_MT2, s_MT3, s_MT4, s_MT5, s_MT6, s_MT7);

    whirlpool_update_vector (&ctx, w, pw_len);

    whirlpool_final_vector (&ctx);

    const u32x r0 = ctx.h[DGST_R0];
    const u32x r1 = ctx.h[DGST_R1];
    const u32x r2 = ctx.h[DGST_R2];
    const u32x r3 = ctx.h[DGST_R3];

    COMPARE_M_SIMD (r0, r1, r2, r3);
  }
}

KERNEL_FQ void m06100_sxx (KERN_ATTR_VECTOR ())
{
  /**
   * modifier
   */

  const u64 gid = get_global_id (0);
  const u64 lid = get_local_id (0);
  const u64 lsz = get_local_size (0);

  /**
   * Whirlpool shared
   */

  #ifdef REAL_SHM

  LOCAL_VK u64 s_MT0[256];
  LOCAL_VK u64 s_MT1[256];
  LOCAL_VK u64 s_MT2[256];
  LOCAL_VK u64 s_MT3[256];
  LOCAL_VK u64 s_MT4[256];
  LOCAL_VK u64 s_MT5[256];
  LOCAL_VK u64 s_MT6[256];
  LOCAL_VK u64 s_MT7[256];

  for (u32 i = lid; i < 256; i += lsz)
  {
    s_MT0[i] = MT0[i];
    s_MT1[i] = MT1[i];
    s_MT2[i] = MT2[i];
    s_MT3[i] = MT3[i];
    s_MT4[i] = MT4[i];
    s_MT5[i] = MT5[i];
    s_MT6[i] = MT6[i];
    s_MT7[i] = MT7[i];
  }

  SYNC_THREADS ();

  #else

  CONSTANT_AS u64a *s_MT0 = MT0;
  CONSTANT_AS u64a *s_MT1 = MT1;
  CONSTANT_AS u64a *s_MT2 = MT2;
  CONSTANT_AS u64a *s_MT3 = MT3;
  CONSTANT_AS u64a *s_MT4 = MT4;
  CONSTANT_AS u64a *s_MT5 = MT5;
  CONSTANT_AS u64a *s_MT6 = MT6;
  CONSTANT_AS u64a *s_MT7 = MT7;

  #endif

  if (gid >= GID_CNT) return;

  /**
   * digest
   */

  const u32 search[4] =
  {
    digests_buf[DIGESTS_OFFSET_HOST].digest_buf[DGST_R0],
    digests_buf[DIGESTS_OFFSET_HOST].digest_buf[DGST_R1],
    digests_buf[DIGESTS_OFFSET_HOST].digest_buf[DGST_R2],
    digests_buf[DIGESTS_OFFSET_HOST].digest_buf[DGST_R3]
  };

  /**
   * base
   */

  const u32 pw_len = pws[gid].pw_len;

  u32x w[64] = { 0 };

  for (u32 i = 0, idx = 0; i < pw_len; i += 4, idx += 1)
  {
    w[idx] = pws[gid].i[idx];
  }

  /**
   * loop
   */

  u32x w0l = w[0];

  for (u32 il_pos = 0; il_pos < IL_CNT; il_pos += VECT_SIZE)
  {
    const u32x w0r = words_buf_r[il_pos / VECT_SIZE];

    const u32x w0 = w0l | w0r;

    w[0] = w0;

    whirlpool_ctx_vector_t ctx;

    whirlpool_init_vector (&ctx, s_MT0, s_MT1, s_MT2, s_MT3, s_MT4, s_MT5, s_MT6, s_MT7);

    whirlpool_update_vector (&ctx, w, pw_len);

    whirlpool_final_vector (&ctx);

    const u32x r0 = ctx.h[DGST_R0];
    const u32x r1 = ctx.h[DGST_R1];
    const u32x r2 = ctx.h[DGST_R2];
    const u32x r3 = ctx.h[DGST_R3];

    COMPARE_S_SIMD (r0, r1, r2, r3);
  }
}