// Run 16 parallel hashes
void sha2Hash(vecu16 pointers, int totalBlocks, vecu16 outHashes)
{
// Initial H values
vecu16 A = __builtin_vp_makevectori(0x6A09E667);
vecu16 B = __builtin_vp_makevectori(0xBB67AE85);
vecu16 C = __builtin_vp_makevectori(0x3C6EF372);
vecu16 D = __builtin_vp_makevectori(0xA54FF53A);
vecu16 E = __builtin_vp_makevectori(0x510E527F);
vecu16 F = __builtin_vp_makevectori(0x9B05688C);
vecu16 G = __builtin_vp_makevectori(0x1F83D9AB);
vecu16 H = __builtin_vp_makevectori(0x5BE0CD19);
for (int i = 0; i < totalBlocks; i++)
{
vecu16 W[64];
for (int index = 0; index < 16; index++)
{
W[index] = __builtin_vp_gather_loadi(pointers);
pointers += __builtin_vp_makevectori(4);
}
for (int index = 16; index < 64; index++)
W[index] = SIG1(W[index - 2]) + W[index - 7] + SIG0(W[index - 15]) + W[index - 16];
for (int round = 0; round < 64; round++)
{
vecu16 temp1 = H + SIG1(E) + CH(E, F, G) + __builtin_vp_makevectori(K[round]) + W[round];
vecu16 temp2 = SIG0(A) + MA(A, B, C);
H = G;
G = F;
F = E;
E = D + temp1;
D = C;
C = B;
B = A;
A = temp1 + temp2;
}
}
// doesn't add padding or length fields to end...
__builtin_vp_scatter_storei(outHashes, A);
__builtin_vp_scatter_storei(outHashes + __builtin_vp_makevectori(4), B);
__builtin_vp_scatter_storei(outHashes + __builtin_vp_makevectori(8), C);
__builtin_vp_scatter_storei(outHashes + __builtin_vp_makevectori(12), D);
__builtin_vp_scatter_storei(outHashes + __builtin_vp_makevectori(16), E);
__builtin_vp_scatter_storei(outHashes + __builtin_vp_makevectori(20), F);
__builtin_vp_scatter_storei(outHashes + __builtin_vp_makevectori(24), G);
__builtin_vp_scatter_storei(outHashes + __builtin_vp_makevectori(28), H);
}
void test_scatter_storei(veci16 ptr, veci16 value) // CHECK: test_scatter_storei
{
__builtin_vp_scatter_storei(ptr, value);
// CHECK: store_scat v1, (v0)
}
