Community
cancel
Showing results for 
Search instead for 
Did you mean: 
Highlighted
Beginner
39 Views

Building opencl kernels clCreateProgramWithSource/clBuildProgram

I'm trying to build the following kernel:


#pragma OPENCL EXTENSION cl_amd_media_ops2 : enable


#include "wolf-aes.cl"
#include "wolf-skein.cl"
#include "jh.cl"
#include "blake256.cl"
#include "groestl256.cl"

 

static const __constant ulong keccakf_rndc[24] =
{
    0x0000000000000001, 0x0000000000008082, 0x800000000000808a,
    0x8000000080008000, 0x000000000000808b, 0x0000000080000001,
    0x8000000080008081, 0x8000000000008009, 0x000000000000008a,
    0x0000000000000088, 0x0000000080008009, 0x000000008000000a,
    0x000000008000808b, 0x800000000000008b, 0x8000000000008089,
    0x8000000000008003, 0x8000000000008002, 0x8000000000000080,
    0x000000000000800a, 0x800000008000000a, 0x8000000080008081,
    0x8000000000008080, 0x0000000080000001, 0x8000000080008008
};

static const __constant uchar sbox[256] =
{
    0x63, 0x7C, 0x77, 0x7B, 0xF2, 0x6B, 0x6F, 0xC5, 0x30, 0x01, 0x67, 0x2B, 0xFE, 0xD7, 0xAB, 0x76,
    0xCA, 0x82, 0xC9, 0x7D, 0xFA, 0x59, 0x47, 0xF0, 0xAD, 0xD4, 0xA2, 0xAF, 0x9C, 0xA4, 0x72, 0xC0,
    0xB7, 0xFD, 0x93, 0x26, 0x36, 0x3F, 0xF7, 0xCC, 0x34, 0xA5, 0xE5, 0xF1, 0x71, 0xD8, 0x31, 0x15,
    0x04, 0xC7, 0x23, 0xC3, 0x18, 0x96, 0x05, 0x9A, 0x07, 0x12, 0x80, 0xE2, 0xEB, 0x27, 0xB2, 0x75,
    0x09, 0x83, 0x2C, 0x1A, 0x1B, 0x6E, 0x5A, 0xA0, 0x52, 0x3B, 0xD6, 0xB3, 0x29, 0xE3, 0x2F, 0x84,
    0x53, 0xD1, 0x00, 0xED, 0x20, 0xFC, 0xB1, 0x5B, 0x6A, 0xCB, 0xBE, 0x39, 0x4A, 0x4C, 0x58, 0xCF,
    0xD0, 0xEF, 0xAA, 0xFB, 0x43, 0x4D, 0x33, 0x85, 0x45, 0xF9, 0x02, 0x7F, 0x50, 0x3C, 0x9F, 0xA8,
    0x51, 0xA3, 0x40, 0x8F, 0x92, 0x9D, 0x38, 0xF5, 0xBC, 0xB6, 0xDA, 0x21, 0x10, 0xFF, 0xF3, 0xD2,
    0xCD, 0x0C, 0x13, 0xEC, 0x5F, 0x97, 0x44, 0x17, 0xC4, 0xA7, 0x7E, 0x3D, 0x64, 0x5D, 0x19, 0x73,
    0x60, 0x81, 0x4F, 0xDC, 0x22, 0x2A, 0x90, 0x88, 0x46, 0xEE, 0xB8, 0x14, 0xDE, 0x5E, 0x0B, 0xDB,
    0xE0, 0x32, 0x3A, 0x0A, 0x49, 0x06, 0x24, 0x5C, 0xC2, 0xD3, 0xAC, 0x62, 0x91, 0x95, 0xE4, 0x79,
    0xE7, 0xC8, 0x37, 0x6D, 0x8D, 0xD5, 0x4E, 0xA9, 0x6C, 0x56, 0xF4, 0xEA, 0x65, 0x7A, 0xAE, 0x08,
    0xBA, 0x78, 0x25, 0x2E, 0x1C, 0xA6, 0xB4, 0xC6, 0xE8, 0xDD, 0x74, 0x1F, 0x4B, 0xBD, 0x8B, 0x8A,
    0x70, 0x3E, 0xB5, 0x66, 0x48, 0x03, 0xF6, 0x0E, 0x61, 0x35, 0x57, 0xB9, 0x86, 0xC1, 0x1D, 0x9E,
    0xE1, 0xF8, 0x98, 0x11, 0x69, 0xD9, 0x8E, 0x94, 0x9B, 0x1E, 0x87, 0xE9, 0xCE, 0x55, 0x28, 0xDF,
    0x8C, 0xA1, 0x89, 0x0D, 0xBF, 0xE6, 0x42, 0x68, 0x41, 0x99, 0x2D, 0x0F, 0xB0, 0x54, 0xBB, 0x16
};


void keccakf1600(ulong *s)
{
    for(int i = 0; i < 24; ++i)
    {
        ulong bc[5], tmp1, tmp2;
        bc[0] = s[0] ^ s[5] ^ s[10] ^ s[15] ^ s[20] ^ rotate(s[2] ^ s[7] ^ s[12] ^ s[17] ^ s[22], 1UL);
        bc[1] = s[1] ^ s[6] ^ s[11] ^ s[16] ^ s[21] ^ rotate(s[3] ^ s[8] ^ s[13] ^ s[18] ^ s[23], 1UL);
        bc[2] = s[2] ^ s[7] ^ s[12] ^ s[17] ^ s[22] ^ rotate(s[4] ^ s[9] ^ s[14] ^ s[19] ^ s[24], 1UL);
        bc[3] = s[3] ^ s[8] ^ s[13] ^ s[18] ^ s[23] ^ rotate(s[0] ^ s[5] ^ s[10] ^ s[15] ^ s[20], 1UL);
        bc[4] = s[4] ^ s[9] ^ s[14] ^ s[19] ^ s[24] ^ rotate(s[1] ^ s[6] ^ s[11] ^ s[16] ^ s[21], 1UL);
        
        tmp1 = s[1] ^ bc[0];
        
        s[0] ^= bc[4];
        s[1] = rotate(s[6] ^ bc[0], 44UL);
        s[6] = rotate(s[9] ^ bc[3], 20UL);
        s[9] = rotate(s[22] ^ bc[1], 61UL);
        s[22] = rotate(s[14] ^ bc[3], 39UL);
        s[14] = rotate(s[20] ^ bc[4], 18UL);
        s[20] = rotate(s[2] ^ bc[1], 62UL);
        s[2] = rotate(s[12] ^ bc[1], 43UL);
        s[12] = rotate(s[13] ^ bc[2], 25UL);
        s[13] = rotate(s[19] ^ bc[3], 8UL);
        s[19] = rotate(s[23] ^ bc[2], 56UL);
        s[23] = rotate(s[15] ^ bc[4], 41UL);
        s[15] = rotate(s[4] ^ bc[3], 27UL);
        s[4] = rotate(s[24] ^ bc[3], 14UL);
        s[24] = rotate(s[21] ^ bc[0], 2UL);
        s[21] = rotate(s[8] ^ bc[2], 55UL);
        s[8] = rotate(s[16] ^ bc[0], 35UL);
        s[16] = rotate(s[5] ^ bc[4], 36UL);
        s[5] = rotate(s[3] ^ bc[2], 28UL);
        s[3] = rotate(s[18] ^ bc[2], 21UL);
        s[18] = rotate(s[17] ^ bc[1], 15UL);
        s[17] = rotate(s[11] ^ bc[0], 10UL);
        s[11] = rotate(s[7] ^ bc[1], 6UL);
        s[7] = rotate(s[10] ^ bc[4], 3UL);
        s[10] = rotate(tmp1, 1UL);
        
        tmp1 = s[0]; tmp2 = s[1]; s[0] = bitselect(s[0] ^ s[2], s[0], s[1]); s[1] = bitselect(s[1] ^ s[3], s[1], s[2]); s[2] = bitselect(s[2] ^ s[4], s[2], s[3]); s[3] = bitselect(s[3] ^ tmp1, s[3], s[4]); s[4] = bitselect(s[4] ^ tmp2, s[4], tmp1);
        tmp1 = s[5]; tmp2 = s[6]; s[5] = bitselect(s[5] ^ s[7], s[5], s[6]); s[6] = bitselect(s[6] ^ s[8], s[6], s[7]); s[7] = bitselect(s[7] ^ s[9], s[7], s[8]); s[8] = bitselect(s[8] ^ tmp1, s[8], s[9]); s[9] = bitselect(s[9] ^ tmp2, s[9], tmp1);
        tmp1 = s[10]; tmp2 = s[11]; s[10] = bitselect(s[10] ^ s[12], s[10], s[11]); s[11] = bitselect(s[11] ^ s[13], s[11], s[12]); s[12] = bitselect(s[12] ^ s[14], s[12], s[13]); s[13] = bitselect(s[13] ^ tmp1, s[13], s[14]); s[14] = bitselect(s[14] ^ tmp2, s[14], tmp1);
        tmp1 = s[15]; tmp2 = s[16]; s[15] = bitselect(s[15] ^ s[17], s[15], s[16]); s[16] = bitselect(s[16] ^ s[18], s[16], s[17]); s[17] = bitselect(s[17] ^ s[19], s[17], s[18]); s[18] = bitselect(s[18] ^ tmp1, s[18], s[19]); s[19] = bitselect(s[19] ^ tmp2, s[19], tmp1);
        tmp1 = s[20]; tmp2 = s[21]; s[20] = bitselect(s[20] ^ s[22], s[20], s[21]); s[21] = bitselect(s[21] ^ s[23], s[21], s[22]); s[22] = bitselect(s[22] ^ s[24], s[22], s[23]); s[23] = bitselect(s[23] ^ tmp1, s[23], s[24]); s[24] = bitselect(s[24] ^ tmp2, s[24], tmp1);
        s[0] ^= keccakf_rndc;
    }
}

static const __constant uint keccakf_rotc[24] =
{
    1,  3,  6,  10, 15, 21, 28, 36, 45, 55, 2,  14,
    27, 41, 56, 8,  25, 43, 62, 18, 39, 61, 20, 44
};

static const __constant uint keccakf_piln[24] =
{
    10, 7,  11, 17, 18, 3, 5,  16, 8,  21, 24, 4,
    15, 23, 19, 13, 12, 2, 20, 14, 22, 9,  6,  1
};

void keccakf1600_1(ulong *st)
{
    int i, round;
    ulong t, bc[5];
    
    #pragma unroll 1
    for(round = 0; round < 24; ++round)
    {

        // Theta
        bc[0] = st[0] ^ st[5] ^ st[10] ^ st[15] ^ st[20];
        bc[1] = st[1] ^ st[6] ^ st[11] ^ st[16] ^ st[21];
        bc[2] = st[2] ^ st[7] ^ st[12] ^ st[17] ^ st[22];
        bc[3] = st[3] ^ st[8] ^ st[13] ^ st[18] ^ st[23];
        bc[4] = st[4] ^ st[9] ^ st[14] ^ st[19] ^ st[24];
        
        #pragma unroll 1
        for (i = 0; i < 5; ++i) {
            t = bc[(i + 4) % 5] ^ rotate(bc[(i + 1) % 5], 1UL);
            st[i     ] ^= t;
            st[i +  5] ^= t;
            st[i + 10] ^= t;
            st[i + 15] ^= t;
            st[i + 20] ^= t;
        }

        // Rho Pi
        t = st[1];
        #pragma unroll
        for (i = 0; i < 24; ++i) {
            bc[0] = st[keccakf_piln];
            st[keccakf_piln] = rotate(t, (ulong)keccakf_rotc);
            t = bc[0];
        }

        //ulong tmp1 = st[0]; ulong tmp2 = st[1]; st[0] = bitselect(st[0] ^ st[2], st[0], st[1]); st[1] = bitselect(st[1] ^ st[3], st[1], st[2]); st[2] = bitselect(st[2] ^ st[4], st[2], st[3]); st[3] = bitselect(st[3] ^ tmp1, st[3], st[4]); st[4] = bitselect(st[4] ^ tmp2, st[4], tmp1);
        //tmp1 = st[5]; tmp2 = st[6]; st[5] = bitselect(st[5] ^ st[7], st[5], st[6]); st[6] = bitselect(st[6] ^ st[8], st[6], st[7]); st[7] = bitselect(st[7] ^ st[9], st[7], st[8]); st[8] = bitselect(st[8] ^ tmp1, st[8], st[9]); st[9] = bitselect(st[9] ^ tmp2, st[9], tmp1);
        //tmp1 = st[10]; tmp2 = st[11]; st[10] = bitselect(st[10] ^ st[12], st[10], st[11]); st[11] = bitselect(st[11] ^ st[13], st[11], st[12]); st[12] = bitselect(st[12] ^ st[14], st[12], st[13]); st[13] = bitselect(st[13] ^ tmp1, st[13], st[14]); st[14] = bitselect(st[14] ^ tmp2, st[14], tmp1);
        //tmp1 = st[15]; tmp2 = st[16]; st[15] = bitselect(st[15] ^ st[17], st[15], st[16]); st[16] = bitselect(st[16] ^ st[18], st[16], st[17]); st[17] = bitselect(st[17] ^ st[19], st[17], st[18]); st[18] = bitselect(st[18] ^ tmp1, st[18], st[19]); st[19] = bitselect(st[19] ^ tmp2, st[19], tmp1);
        //tmp1 = st[20]; tmp2 = st[21]; st[20] = bitselect(st[20] ^ st[22], st[20], st[21]); st[21] = bitselect(st[21] ^ st[23], st[21], st[22]); st[22] = bitselect(st[22] ^ st[24], st[22], st[23]); st[23] = bitselect(st[23] ^ tmp1, st[23], st[24]); st[24] = bitselect(st[24] ^ tmp2, st[24], tmp1);
        
        #pragma unroll 1
        for(int i = 0; i < 25; i += 5)
        {    
            ulong tmp[5];
            
            #pragma unroll 1
            for(int x = 0; x < 5; ++x)
                tmp = bitselect(st[i + x] ^ st[i + ((x + 2) % 5)], st[i + x], st[i + ((x + 1) % 5)]);
            
            #pragma unroll 1
            for(int x = 0; x < 5; ++x) st[i + x] = tmp;
        }
        
        //  Iota
        st[0] ^= keccakf_rndc[round];
    }
}

void keccakf1600_2(ulong *st)
{
    int i, round;
    ulong t, bc[5];
    
    #pragma unroll 1
    for(round = 0; round < 24; ++round)
    {

        // Theta
        //bc[0] = st[0] ^ st[5] ^ st[10] ^ st[15] ^ st[20];
        //bc[1] = st[1] ^ st[6] ^ st[11] ^ st[16] ^ st[21];
        //bc[2] = st[2] ^ st[7] ^ st[12] ^ st[17] ^ st[22];
        //bc[3] = st[3] ^ st[8] ^ st[13] ^ st[18] ^ st[23];
        //bc[4] = st[4] ^ st[9] ^ st[14] ^ st[19] ^ st[24];
        
        /*
        #pragma unroll
        for (i = 0; i < 5; ++i) {
            t = bc[(i + 4) % 5] ^ rotate(bc[(i + 1) % 5], 1UL);
            st[i     ] ^= t;
            st[i +  5] ^= t;
            st[i + 10] ^= t;
            st[i + 15] ^= t;
            st[i + 20] ^= t;
        }
        */
        
        bc[0] = st[0] ^ st[5] ^ st[10] ^ st[15] ^ st[20] ^ rotate(st[2] ^ st[7] ^ st[12] ^ st[17] ^ st[22], 1UL);
        bc[1] = st[1] ^ st[6] ^ st[11] ^ st[16] ^ st[21] ^ rotate(st[3] ^ st[8] ^ st[13] ^ st[18] ^ st[23], 1UL);
        bc[2] = st[2] ^ st[7] ^ st[12] ^ st[17] ^ st[22] ^ rotate(st[4] ^ st[9] ^ st[14] ^ st[19] ^ st[24], 1UL);
        bc[3] = st[3] ^ st[8] ^ st[13] ^ st[18] ^ st[23] ^ rotate(st[0] ^ st[5] ^ st[10] ^ st[15] ^ st[20], 1UL);
        bc[4] = st[4] ^ st[9] ^ st[14] ^ st[19] ^ st[24] ^ rotate(st[1] ^ st[6] ^ st[11] ^ st[16] ^ st[21], 1UL);
        
        st[0] ^= bc[4];
        st[5] ^= bc[4];
        st[10] ^= bc[4];
        st[15] ^= bc[4];
        st[20] ^= bc[4];
        
        st[1] ^= bc[0];
        st[6] ^= bc[0];
        st[11] ^= bc[0];
        st[16] ^= bc[0];
        st[21] ^= bc[0];
        
        st[2] ^= bc[1];
        st[7] ^= bc[1];
        st[12] ^= bc[1];
        st[17] ^= bc[1];
        st[22] ^= bc[1];
        
        st[3] ^= bc[2];
        st[8] ^= bc[2];
        st[13] ^= bc[2];
        st[18] ^= bc[2];
        st[23] ^= bc[2];
        
        st[4] ^= bc[3];
        st[9] ^= bc[3];
        st[14] ^= bc[3];
        st[19] ^= bc[3];
        st[24] ^= bc[3];
        
        // Rho Pi
        t = st[1];
        #pragma unroll
        for (i = 0; i < 24; ++i) {
            bc[0] = st[keccakf_piln];
            st[keccakf_piln] = rotate(t, (ulong)keccakf_rotc);
            t = bc[0];
        }
        
        
        
        /*ulong tmp1 = st[1] ^ bc[0];
        
        st[0] ^= bc[4];
        st[1] = rotate(st[6] ^ bc[0], 44UL);
        st[6] = rotate(st[9] ^ bc[3], 20UL);
        st[9] = rotate(st[22] ^ bc[1], 61UL);
        st[22] = rotate(st[14] ^ bc[3], 39UL);
        st[14] = rotate(st[20] ^ bc[4], 18UL);
        st[20] = rotate(st[2] ^ bc[1], 62UL);
        st[2] = rotate(st[12] ^ bc[1], 43UL);
        st[12] = rotate(st[13] ^ bc[2], 25UL);
        st[13] = rotate(st[19] ^ bc[3], 8UL);
        st[19] = rotate(st[23] ^ bc[2], 56UL);
        st[23] = rotate(st[15] ^ bc[4], 41UL);
        st[15] = rotate(st[4] ^ bc[3], 27UL);
        st[4] = rotate(st[24] ^ bc[3], 14UL);
        st[24] = rotate(st[21] ^ bc[0], 2UL);
        st[21] = rotate(st[8] ^ bc[2], 55UL);
        st[8] = rotate(st[16] ^ bc[0], 35UL);
        st[16] = rotate(st[5] ^ bc[4], 36UL);
        st[5] = rotate(st[3] ^ bc[2], 28UL);
        st[3] = rotate(st[18] ^ bc[2], 21UL);
        st[18] = rotate(st[17] ^ bc[1], 15UL);
        st[17] = rotate(st[11] ^ bc[0], 10UL);
        st[11] = rotate(st[7] ^ bc[1], 6UL);
        st[7] = rotate(st[10] ^ bc[4], 3UL);
        st[10] = rotate(tmp1, 1UL);
        */
        
        
        //ulong tmp1 = st[0]; ulong tmp2 = st[1]; st[0] = bitselect(st[0] ^ st[2], st[0], st[1]); st[1] = bitselect(st[1] ^ st[3], st[1], st[2]); st[2] = bitselect(st[2] ^ st[4], st[2], st[3]); st[3] = bitselect(st[3] ^ tmp1, st[3], st[4]); st[4] = bitselect(st[4] ^ tmp2, st[4], tmp1);
        //tmp1 = st[5]; tmp2 = st[6]; st[5] = bitselect(st[5] ^ st[7], st[5], st[6]); st[6] = bitselect(st[6] ^ st[8], st[6], st[7]); st[7] = bitselect(st[7] ^ st[9], st[7], st[8]); st[8] = bitselect(st[8] ^ tmp1, st[8], st[9]); st[9] = bitselect(st[9] ^ tmp2, st[9], tmp1);
        //tmp1 = st[10]; tmp2 = st[11]; st[10] = bitselect(st[10] ^ st[12], st[10], st[11]); st[11] = bitselect(st[11] ^ st[13], st[11], st[12]); st[12] = bitselect(st[12] ^ st[14], st[12], st[13]); st[13] = bitselect(st[13] ^ tmp1, st[13], st[14]); st[14] = bitselect(st[14] ^ tmp2, st[14], tmp1);
        //tmp1 = st[15]; tmp2 = st[16]; st[15] = bitselect(st[15] ^ st[17], st[15], st[16]); st[16] = bitselect(st[16] ^ st[18], st[16], st[17]); st[17] = bitselect(st[17] ^ st[19], st[17], st[18]); st[18] = bitselect(st[18] ^ tmp1, st[18], st[19]); st[19] = bitselect(st[19] ^ tmp2, st[19], tmp1);
        //tmp1 = st[20]; tmp2 = st[21]; st[20] = bitselect(st[20] ^ st[22], st[20], st[21]); st[21] = bitselect(st[21] ^ st[23], st[21], st[22]); st[22] = bitselect(st[22] ^ st[24], st[22], st[23]); st[23] = bitselect(st[23] ^ tmp1, st[23], st[24]); st[24] = bitselect(st[24] ^ tmp2, st[24], tmp1);
        
        #pragma unroll
        for(int i = 0; i < 25; i += 5)
        {
            ulong tmp1 = st, tmp2 = st[i + 1];
            
            st = bitselect(st ^ st[i + 2], st, st[i + 1]);
            st[i + 1] = bitselect(st[i + 1] ^ st[i + 3], st[i + 1], st[i + 2]);
            st[i + 2] = bitselect(st[i + 2] ^ st[i + 4], st[i + 2], st[i + 3]);
            st[i + 3] = bitselect(st[i + 3] ^ tmp1, st[i + 3], st[i + 4]);
            st[i + 4] = bitselect(st[i + 4] ^ tmp2, st[i + 4], tmp1);
        }
        
        //  Iota
        st[0] ^= keccakf_rndc[round];
    }
}

void CNKeccak(ulong *output, ulong *input)
{
    ulong st[25];
    
    // Copy 72 bytes
    for(int i = 0; i < 9; ++i) st = input;
    
    // Last four and '1' bit for padding
    //st[9] = as_ulong((uint2)(((uint *)input)[18], 0x00000001U));
    
    st[9] = (input[9] & 0x00000000FFFFFFFFUL) | 0x0000000100000000UL;
    
    for(int i = 10; i < 25; ++i) st = 0x00UL;
    
    // Last bit of padding
    st[16] = 0x8000000000000000UL;
    
    keccakf1600_1(st);
    
    for(int i = 0; i < 25; ++i) output = st;
}

static const __constant uchar rcon[8] = { 0x8d, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40 };

#pragma OPENCL EXTENSION cl_amd_media_ops2 : enable

#define BYTE(x, y)    (amd_bfe((x), (y) << 3U, 8U))

#define SubWord(inw)        ((sbox[BYTE(inw, 3)] << 24) | (sbox[BYTE(inw, 2)] << 16) | (sbox[BYTE(inw, 1)] << 8) | sbox[BYTE(inw, 0)])

void AESExpandKey256(uint *keybuf)
{
    //#pragma unroll 4
    for(uint c = 8, i = 1; c < 60; ++c)
    {
        // For 256-bit keys, an sbox permutation is done every other 4th uint generated, AND every 8th
        uint t = ((!(c & 7)) || ((c & 7) == 4)) ? SubWord(keybuf[c - 1]) : keybuf[c - 1];
        
        // If the uint we're generating has an index that is a multiple of 8, rotate and XOR with the round constant,
        // then XOR this with previously generated uint. If it's 4 after a multiple of 8, only the sbox permutation
        // is done, followed by the XOR. If neither are true, only the XOR with the previously generated uint is done.
        keybuf = keybuf[c - 8] ^ ((!(c & 7)) ? rotate(t, 24U) ^ as_uint((uchar4)(rcon[i++], 0U, 0U, 0U)) : t);
    }
}

#define IDX(x)    ((x) * (get_global_size(0)))

__attribute__((reqd_work_group_size(WORKSIZE, 8, 1)))
__kernel void cn0(__global ulong *input, __global uint4 *Scratchpad, __global ulong *states)
{
    ulong State[25];
    uint ExpandedKey1[256];
    __local uint AES0[256], AES1[256], AES2[256], AES3[256];
    ulong inbuf[10];
    uint4 text;
    
    states += (25 * (get_global_id(0) - get_global_offset(0)));
    Scratchpad += ((get_global_id(0) - get_global_offset(0)));
    
    for(int i = get_local_id(0); i < 256; i += WORKSIZE)
    {
        const uint tmp = AES0_C;
        AES0 = tmp;
        AES1 = rotate(tmp, 8U);
        AES2 = rotate(tmp, 16U);
        AES3 = rotate(tmp, 24U);
    }
    
    ((ulong8 *)inbuf)[0] = vload8(0, input);
    inbuf[8] = input[8];
    inbuf[9] = (ulong)((__global uint *)input)[18];
    
    ((uint *)(((uchar *)inbuf) + 39))[0] = get_global_id(0);
    CNKeccak(State, inbuf);
    
    #pragma unroll
    for(int i = 0; i < 25; ++i) states = State;
    
    text = vload4(get_local_id(1) + 4, (uint *)(State));
    
    #pragma unroll
    for(int i = 0; i < 4; ++i) ((ulong *)ExpandedKey1) = State;
    
    AESExpandKey256(ExpandedKey1);
    
    mem_fence(CLK_LOCAL_MEM_FENCE);
    
    #pragma unroll 2
    for(int i = 0; i < 0x4000; ++i)
    {
        #pragma unroll
        for(int j = 0; j < 10; ++j)
            text = AES_Round(AES0, AES1, AES2, AES3, text, ((uint4 *)ExpandedKey1));
        
        Scratchpad[IDX((i << 3) + get_local_id(1))] = text;
    }
    
    mem_fence(CLK_GLOBAL_MEM_FENCE);
}

__attribute__((reqd_work_group_size(WORKSIZE, 1, 1)))
__kernel void cn1(__global uint4 *Scratchpad, __global ulong *states)
{
    ulong a[2], b[2];
    __local uint AES0[256], AES1[256], AES2[256], AES3[256];
    
    Scratchpad += ((get_global_id(0) - get_global_offset(0)));
    states += (25 * (get_global_id(0) - get_global_offset(0)));
    
    for(int i = get_local_id(0); i < 256; i += WORKSIZE)
    {
        const uint tmp = AES0_C;
        AES0 = tmp;
        AES1 = rotate(tmp, 8U);
        AES2 = rotate(tmp, 16U);
        AES3 = rotate(tmp, 24U);
    }
    
    a[0] = states[0] ^ states[4];
    b[0] = states[2] ^ states[6];
    a[1] = states[1] ^ states[5];
    b[1] = states[3] ^ states[7];
    
    uint4 b_x = ((uint4 *)b)[0];
    
    mem_fence(CLK_LOCAL_MEM_FENCE);
    
    #pragma unroll 8
    for(int i = 0; i < 0x80000; ++i)
    {
        ulong c[2];
        
        ((uint4 *)c)[0] = Scratchpad[IDX((a[0] & 0x1FFFF0) >> 4)];
        ((uint4 *)c)[0] = AES_Round(AES0, AES1, AES2, AES3, ((uint4 *)c)[0], ((uint4 *)a)[0]);
        //b_x ^= ((uint4 *)c)[0];
        
        Scratchpad[IDX((a[0] & 0x1FFFF0) >> 4)] = b_x ^ ((uint4 *)c)[0];
        
        uint4 tmp;
        tmp = Scratchpad[IDX((c[0] & 0x1FFFF0) >> 4)];
        
        a[1] += c[0] * as_ulong2(tmp).s0;
        a[0] += mul_hi(c[0], as_ulong2(tmp).s0);
        
        Scratchpad[IDX((c[0] & 0x1FFFF0) >> 4)] = ((uint4 *)a)[0];
        
        ((uint4 *)a)[0] ^= tmp;
        
        b_x = ((uint4 *)c)[0];
    }
    
    mem_fence(CLK_GLOBAL_MEM_FENCE);
}

__attribute__((reqd_work_group_size(WORKSIZE, 8, 1)))
__kernel void cn2(__global uint4 *Scratchpad, __global ulong *states, __global uint *Branch0, __global uint *Branch1, __global uint *Branch2, __global uint *Branch3)
{
    __local uint AES0[256], AES1[256], AES2[256], AES3[256];
    uint ExpandedKey2[256];
    ulong State[25];
    uint4 text;
    
    Scratchpad += ((get_global_id(0) - get_global_offset(0)));
    states += (25 * (get_global_id(0) - get_global_offset(0)));
    
    for(int i = get_local_id(0); i < 256; i += WORKSIZE)
    {
        const uint tmp = AES0_C;
        AES0 = tmp;
        AES1 = rotate(tmp, 8U);
        AES2 = rotate(tmp, 16U);
        AES3 = rotate(tmp, 24U);
    }
    
    #if defined(__Tahiti__) || defined(__Pitcairn__)
    
    for(int i = 0; i < 4; ++i) ((ulong *)ExpandedKey2) = states[i + 4];
    text = vload4(get_local_id(1) + 4, (__global uint *)states);
    
    #else
    
    text = vload4(get_local_id(1) + 4, (__global uint *)states);
    ((uint8 *)ExpandedKey2)[0] = vload8(1, (__global uint *)states);
    
    #endif
    
    AESExpandKey256(ExpandedKey2);
    
    barrier(CLK_LOCAL_MEM_FENCE);
    
    #pragma unroll 2
    for(int i = 0; i < 0x4000; ++i)
    {        
        text ^= Scratchpad[IDX((i << 3) + get_local_id(1))];
        
        #pragma unroll
        for(int j = 0; j < 10; ++j)
            text = AES_Round(AES0, AES1, AES2, AES3, text, ((uint4 *)ExpandedKey2));
    }
    
    vstore2(as_ulong2(text), get_local_id(1) + 4, states);
    
    barrier(CLK_GLOBAL_MEM_FENCE);
    
    if(!get_local_id(1))
    {
        for(int i = 0; i < 25; ++i) State = states;
        
        keccakf1600_2(State);
        
        for(int i = 0; i < 25; ++i) states = State;
        
        switch(State[0] & 3)
        {
            case 0:
                Branch0[atomic_inc(Branch0 + get_global_size(0))] = get_global_id(0) - get_global_offset(0);
                break;
            case 1:
                Branch1[atomic_inc(Branch1 + get_global_size(0))] = get_global_id(0) - get_global_offset(0);
                break;
            case 2:
                Branch2[atomic_inc(Branch2 + get_global_size(0))] = get_global_id(0) - get_global_offset(0);
                break;
            case 3:
                Branch3[atomic_inc(Branch3 + get_global_size(0))] = get_global_id(0) - get_global_offset(0);
                break;
        }
    }
    
    mem_fence(CLK_GLOBAL_MEM_FENCE);
}

/*
__kernel void cryptonight(__global ulong *input, __global uint4 *Scratchpad, __global ulong *states, __global uint *Branch0, __global uint *Branch1, __global uint *Branch2, __global uint *Branch3, ulong ThreadCount)
{
    uchar State[200];
    __local uint AES0[256], AES1[256], AES2[256], AES3[256];
    uchar ExpandedKey1[256], ExpandedKey2[256];
    ulong inbuf[10], a[2], b[2];
    uint4 text[8];
    
    for(int i = 0; i < 256; ++i)
    {
        const uint tmp = AES0_C;
        AES0 = tmp;
        AES1 = rotate(tmp, 8U);
        AES2 = rotate(tmp, 16U);
        AES3 = rotate(tmp, 24U);
    }
    
    ((ulong8 *)inbuf)[0] = vload8(0, input);
    inbuf[8] = input[8];
    inbuf[9] = (ulong)((__global uint *)input)[18];
    
    ((uint *)(((uchar *)inbuf) + 39))[0] = get_global_id(0);
    CNKeccak((ulong *)State, inbuf);
    
    a[0] = ((ulong *)State)[0] ^ ((ulong *)State)[4];
    b[0] = ((ulong *)State)[2] ^ ((ulong *)State)[6];
    a[1] = ((ulong *)State)[1] ^ ((ulong *)State)[5];
    b[1] = ((ulong *)State)[3] ^ ((ulong *)State)[7];
    
    for(uint i = 0; i < 8; ++i) text = vload4(i + 4, (uint *)(State));
    
    for(int i = 0; i < 4; ++i) ((ulong *)ExpandedKey1) = ((ulong *)State);
    for(int i = 0; i < 4; ++i) ((ulong *)ExpandedKey2) = ((ulong *)State)[i + 4];
        
    AESExpandKey256(ExpandedKey1);
    AESExpandKey256(ExpandedKey2);
    
    mem_fence(CLK_LOCAL_MEM_FENCE);
    
    Scratchpad += ((1 << 17) * (get_global_id(0) - get_global_offset(0)));
    
    //#pragma unroll 1
    for(int i = 0; i < (1 << 17); i += 8)
    {
        #pragma unroll
        for(int j = 0; j < 10; ++j)
        {
            #pragma unroll
            for(int x = 0; x < 8; ++x)
                text = AES_Round(AES0, AES1, AES2, AES3, text, ((uint4 *)ExpandedKey1));
        }
        
        for(int j = 0; j < 8; ++j) *(Scratchpad + i + j) = text;
    }
    
    
    uint4 b_x = ((uint4 *)b)[0];
    
    //#pragma unroll 1
    for(int i = 0; i < 0x80000; ++i)
    {
        ulong c[2];
        
        ((uint4 *)c)[0] = Scratchpad[(a[0] & 0x1FFFF0) >> 4];
        ((uint4 *)c)[0] = AES_Round(AES0, AES1, AES2, AES3, ((uint4 *)c)[0], ((uint4 *)a)[0]);
        b_x ^= ((uint4 *)c)[0];
        
        Scratchpad[(a[0] & 0x1FFFF0) >> 4] = b_x;
        
        uint4 tmp;
        tmp = Scratchpad[(c[0] & 0x1FFFF0) >> 4];
        
        a[1] += c[0] * as_ulong2(tmp).s0;
        a[0] += mul_hi(c[0], as_ulong2(tmp).s0);
        
        Scratchpad[(c[0] & 0x1FFFF0) >> 4] = ((uint4 *)a)[0];
        
        ((uint4 *)a)[0] ^= tmp;
        
        b_x = ((uint4 *)c)[0];
    }
    
    for(uint i = 0; i < 8; ++i) text = vload4(i + 4, (uint *)(State));
    
    for(int i = 0; i < (1 << 17); i += 8)
    {
        #pragma unroll
        for(int j = 0; j < 8; ++j) text ^= Scratchpad[i + j];
        
        #pragma unroll 1
        for(int j = 0; j < 10; ++j)
        {
            #pragma unroll
            for(int x = 0; x < 8; ++x)
                text = AES_Round(AES0, AES1, AES2, AES3, text, ((uint4 *)ExpandedKey2));
        }
    }
    
    for(uint i = 0; i < 8; ++i) vstore4(text, i + 4, (uint *)(State));
    
    keccakf1600((ulong *)State);
        
    states += (25 * (get_global_id(0) - get_global_offset(0)));
    
    for(int i = 0; i < 25; ++i) states = ((ulong *)State);
    
    switch(State[0] & 3)
    {
        case 0:
            Branch0[atomic_inc(Branch0 + ThreadCount)] = get_global_id(0) - get_global_offset(0);
            break;
        case 1:
            Branch1[atomic_inc(Branch1 + ThreadCount)] = get_global_id(0) - get_global_offset(0);
            break;
        case 2:
            Branch2[atomic_inc(Branch2 + ThreadCount)] = get_global_id(0) - get_global_offset(0);
            break;
        case 3:
            Branch3[atomic_inc(Branch3 + ThreadCount)] = get_global_id(0) - get_global_offset(0);
            break;
    }    
}
*/

#define VSWAP8(x)    (((x) >> 56) | (((x) >> 40) & 0x000000000000FF00UL) | (((x) >> 24) & 0x0000000000FF0000UL) \
          | (((x) >>  8) & 0x00000000FF000000UL) | (((x) <<  8) & 0x000000FF00000000UL) \
          | (((x) << 24) & 0x0000FF0000000000UL) | (((x) << 40) & 0x00FF000000000000UL) | (((x) << 56) & 0xFF00000000000000UL))

#define VSWAP4(x)    ((((x) >> 24) & 0xFFU) | (((x) >> 8) & 0xFF00U) | (((x) << 8) & 0xFF0000U) | (((x) << 24) & 0xFF000000U))

__kernel void Skein(__global ulong *states, __global uint *BranchBuf, __global uint *output, uint Target)
{
    const ulong idx = get_global_id(0) - get_global_offset(0);
    states += 25 * BranchBuf[idx];
    
    // skein
    ulong8 h = vload8(0, SKEIN512_256_IV);
    
    // Type field begins with final bit, first bit, then six bits of type; the last 96
    // bits are input processed (including in the block to be processed with that tweak)
    // The output transform is only one run of UBI, since we need only 256 bits of output
    // The tweak for the output transform is Type = Output with the Final bit set
    // T[0] for the output is 8, and I don't know why - should be message size...
    ulong t[3] = { 0x00UL, 0x7000000000000000UL, 0x00UL };
    ulong8 p, m;
    
    for(uint i = 0; i < 4; ++i)
    {
        if(i < 3) t[0] += 0x40UL;
        else t[0] += 0x08UL;
        
        t[2] = t[0] ^ t[1];
        
        m = (i < 3) ? vload8(i, states) : (ulong8)(states[24], 0UL, 0UL, 0UL, 0UL, 0UL, 0UL, 0UL);
        const ulong h8 = h.s0 ^ h.s1 ^ h.s2 ^ h.s3 ^ h.s4 ^ h.s5 ^ h.s6 ^ h.s7 ^ SKEIN_KS_PARITY;
        p = Skein512Block(m, h, h8, t);
        
        h = m ^ p;
        
        if(i < 2) t[1] = 0x3000000000000000UL;
        else t[1] = 0xB000000000000000UL;
    }
    
    t[0] = 0x08UL;
    t[1] = 0xFF00000000000000UL;
    t[2] = t[0] ^ t[1];
    
    p = (ulong8)(0);
    const ulong h8 = h.s0 ^ h.s1 ^ h.s2 ^ h.s3 ^ h.s4 ^ h.s5 ^ h.s6 ^ h.s7 ^ SKEIN_KS_PARITY;
    
    p = Skein512Block(p, h, h8, t);
    
    //vstore8(p, 0, output);
    
    if(as_uint16(p).s7 <= Target) output[atomic_inc(output + 0xFF)] = BranchBuf[idx] + get_global_offset(0);
    
    mem_fence(CLK_GLOBAL_MEM_FENCE);    
}

#define SWAP8(x)    as_ulong(as_uchar8(x).s76543210)

__kernel void JH(__global ulong *states, __global uint *BranchBuf, __global uint *output, uint Target)
{
    const uint idx = get_global_id(0) - get_global_offset(0);
    states += 25 * BranchBuf[idx];
    
    sph_u64 h0h = 0xEBD3202C41A398EBUL, h0l = 0xC145B29C7BBECD92UL, h1h = 0xFAC7D4609151931CUL, h1l = 0x038A507ED6820026UL, h2h = 0x45B92677269E23A4UL, h2l = 0x77941AD4481AFBE0UL, h3h = 0x7A176B0226ABB5CDUL, h3l = 0xA82FFF0F4224F056UL;
    sph_u64 h4h = 0x754D2E7F8996A371UL, h4l = 0x62E27DF70849141DUL, h5h = 0x948F2476F7957627UL, h5l = 0x6C29804757B6D587UL, h6h = 0x6C0D8EAC2D275E5CUL, h6l = 0x0F7A0557C6508451UL, h7h = 0xEA12247067D3E47BUL, h7l = 0x69D71CD313ABE389UL;
    sph_u64 tmp;
    
    for(int i = 0; i < 5; ++i)
    {
        ulong input[8];
        
        if(i < 3)
        {
            for(int x = 0; x < 8; ++x) input = (states[(i << 3) + x]);
        }
        else if(i == 3)
        {
            input[0] = (states[24]);
            input[1] = 0x80UL;
            for(int x = 2; x < 8; ++x) input = 0x00UL;
        }
        else
        {
            input[7] = 0x4006000000000000UL;
                        
            for(int x = 0; x < 7; ++x) input = 0x00UL;
        }
        
        h0h ^= input[0];
        h0l ^= input[1];
        h1h ^= input[2];
        h1l ^= input[3];
        h2h ^= input[4];
        h2l ^= input[5];
        h3h ^= input[6];
        h3l ^= input[7];
        
        E8;
        
        h4h ^= input[0];
        h4l ^= input[1];
        h5h ^= input[2];
        h5l ^= input[3];
        h6h ^= input[4];
        h6l ^= input[5];
        h7h ^= input[6];
        h7l ^= input[7];
    }
    
    //output[0] = h6h;
    //output[1] = h6l;
    //output[2] = h7h;
    //output[3] = h7l;
    
    if(as_uint2(h7l).s1 <= Target) output[atomic_inc(output + 0xFF)] = BranchBuf[idx] + get_global_offset(0);
}

#define SWAP4(x)    as_uint(as_uchar4(x).s3210)

__kernel void Blake(__global ulong *states, __global uint *BranchBuf, __global uint *output, uint Target)
{
    const uint idx = get_global_id(0) - get_global_offset(0);
    states += 25 * BranchBuf[idx];
    
    unsigned int m[16];
    unsigned int v[16];
    uint h[8];
    
    ((uint8 *)h)[0] = vload8(0U, c_IV256);
    
    for(uint i = 0, bitlen = 0; i < 4; ++i)
    {
        if(i < 3)
        {
            ((uint16 *)m)[0] = vload16(i, (__global uint *)states);
            for(int i = 0; i < 16; ++i) m = SWAP4(m);
            bitlen += 512;
        }
        else
        {
            m[0] = SWAP4(((__global uint *)states)[48]);
            m[1] = SWAP4(((__global uint *)states)[49]);
            m[2] = 0x80000000U;
            
            for(int i = 3; i < 13; ++i) m = 0x00U;
            
            m[13] = 1U;
            m[14] = 0U;
            m[15] = 0x640;
            bitlen += 64;
        }
        
        ((uint16 *)v)[0].lo = ((uint8 *)h)[0];
        ((uint16 *)v)[0].hi = vload8(0U, c_u256);
        
        //v[12] ^= (i < 3) ? (i + 1) << 9 : 1600U;
        //v[13] ^= (i < 3) ? (i + 1) << 9 : 1600U;
        
        v[12] ^= bitlen;
        v[13] ^= bitlen;
        
        for(int r = 0; r < 14; r++)
        {    
            GS(0, 4, 0x8, 0xC, 0x0);
            GS(1, 5, 0x9, 0xD, 0x2);
            GS(2, 6, 0xA, 0xE, 0x4);
            GS(3, 7, 0xB, 0xF, 0x6);
            GS(0, 5, 0xA, 0xF, 0x8);
            GS(1, 6, 0xB, 0xC, 0xA);
            GS(2, 7, 0x8, 0xD, 0xC);
            GS(3, 4, 0x9, 0xE, 0xE);
        }
        
        ((uint8 *)h)[0] ^= ((uint8 *)v)[0] ^ ((uint8 *)v)[1];
    }
    
    for(int i = 0; i < 8; ++i) h = SWAP4(h);
    
    //for(int i = 0; i < 4; ++i) output = ((ulong *)h);
    if(h[7] <= Target) output[atomic_inc(output + 0xFF)] = BranchBuf[idx] + get_global_offset(0);
}

__kernel void Groestl(__global ulong *states, __global uint *BranchBuf, __global uint *output, uint Target)
{    
    const uint idx = get_global_id(0) - get_global_offset(0);
    states += 25 * BranchBuf[idx];
    
    ulong State[8];
    
    for(int i = 0; i < 7; ++i) State = 0UL;
    
    State[7] = 0x0001000000000000UL;
    
    for(uint i = 0; i < 4; ++i)
    {
        ulong H[8], M[8];
        
        if(i < 3)
        {
            ((ulong8 *)M)[0] = vload8(i, states);
        }
        else
        {
            M[0] = states[24];
            M[1] = 0x80UL;
                        
            for(int x = 2; x < 7; ++x) M = 0UL;
            
            M[7] = 0x0400000000000000UL;
        }
        
        for(int x = 0; x < 8; ++x) H = M ^ State;
        
        PERM_SMALL_P(H);
        PERM_SMALL_Q(M);
        
        for(int x = 0; x < 8; ++x) State ^= H ^ M;
    }
    
    ulong tmp[8];
    
    for(int i = 0; i < 8; ++i) tmp = State;
    
    PERM_SMALL_P(State);
    
    for(int i = 0; i < 8; ++i) State ^= tmp;
    
    //for(int i = 0; i < 4; ++i) output = State[i + 4];
    if(as_uint2(State[7]).s1 <= Target) output[atomic_inc(output + 0xFF)] = BranchBuf[idx] + get_global_offset(0);
}

on an Nvidia device running CUDA 7.5, on debian:

$ lsb_release  -da
No LSB modules are available.
Distributor ID:    Debian
Description:    Debian GNU/Linux 8.4 (jessie)
Release:    8.4
Codename:    jessie

opencl was installed together with CUDA (following the guide on the download page (I cannot post links yet)).

The code that calls the read and build function is:

HashData->Program = clCreateProgramWithSource(OCL->Context, 1, (const char **)&KernelSource, NULL, &retval);
    
    if(retval != CL_SUCCESS)
    {
        Log(LOG_CRITICAL, "Error %d when calling clCreateProgramWithSource on the contents of %s.", retval, "cryptonight.cl");
        return(ERR_OCL_API);
    }
    
    Options = (char *)malloc(sizeof(char) * 32);
    
    snprintf(Options, 31, "-I. -DWORKSIZE=%d", LocalThreads);
    
    retval = clBuildProgram(HashData->Program, 1, &OCL->Devices[DeviceIdx].DeviceID, Options, NULL, NULL);
    
    if(retval != CL_SUCCESS)
    {
        Log(LOG_CRITICAL, "Error %d when calling clBuildProgram.", retval);
        
        retval = clGetProgramBuildInfo(HashData->Program, OCL->Devices[DeviceIdx].DeviceID, CL_PROGRAM_BUILD_LOG, 0, NULL, &len);
    
        if(retval != CL_SUCCESS)
        {
            Log(LOG_CRITICAL, "Error %d when calling clGetProgramBuildInfo for length of build log output.", retval);
            return(ERR_OCL_API);
        }
        
        BuildLog = (char *)malloc(sizeof(char) * (len + 2));
        
        retval = clGetProgramBuildInfo(HashData->Program, OCL->Devices[DeviceIdx].DeviceID, CL_PROGRAM_BUILD_LOG, len, BuildLog, NULL);
        
        if(retval != CL_SUCCESS)
        {
            Log(LOG_CRITICAL, "Error %d when calling clGetProgramBuildInfo for build log.", retval);
            return(ERR_OCL_API);
        }
        
        Log(LOG_CRITICAL, "Build Log:\n%s", BuildLog);
        
        free(BuildLog);
        
        return(ERR_OCL_API);
    }


clCreateProgramWithSource()
completes successfully but clBuildProgram() fails with the following logs:

<kernel>:1:26: warning: unknown OpenCL extension 'cl_amd_media_ops2' - ignoring
#pragma OPENCL EXTENSION cl_amd_media_ops2 : enable
                         ^
In file included from <kernel>:4:
./wolf-aes.cl:79:14: warning: implicit declaration of function 'amd_bfe' is invalid in C99
        Y.s0 = AES0[BYTE(X.s0, 0)] ^ AES1[BYTE(X.s1, 1)] ^ AES2[BYTE(X.s2, 2)] ^ AES3[BYTE(X.s3, 3)];
                    ^
./wolf-aes.cl:74:21: note: expanded from macro 'BYTE'
#define BYTE(x, y)      (amd_bfe((x), (y) << 3U, 8U))
                         ^
In file included from <kernel>:5:
./wolf-skein.cl:33:29: warning: implicit declaration of function 'amd_bitalign' is invalid in C99
        if(y < 32) return(as_ulong(amd_bitalign(x, x.s10, 32 - y)));
                                   ^
./wolf-skein.cl:33:20: error: call to 'as_ulong' is ambiguous
        if(y < 32) return(as_ulong(amd_bitalign(x, x.s10, 32 - y)));
                          ^~~~~~~~
cl_kernel.h:14824:24: note: candidate function
ulong __OVERLOADABLE__ as_ulong (long);
                       ^
cl_kernel.h:14825:24: note: candidate function
ulong __OVERLOADABLE__ as_ulong (ulong);
                       ^
cl_kernel.h:14826:24: note: candidate function
ulong __OVERLOADABLE__ as_ulong (double);
                       ^
cl_kernel.h:15502:24: note: candidate function
ulong __OVERLOADABLE__ as_ulong (char8);
                       ^
cl_kernel.h:15503:24: note: candidate function
ulong __OVERLOADABLE__ as_ulong (uchar8);
                       ^
cl_kernel.h:15505:24: note: candidate function
ulong __OVERLOADABLE__ as_ulong (short3);
                       ^
cl_kernel.h:15507:24: note: candidate function
ulong __OVERLOADABLE__ as_ulong (short4);
                       ^
cl_kernel.h:15509:24: note: candidate function
ulong __OVERLOADABLE__ as_ulong (ushort3);
                       ^
cl_kernel.h:15511:24: note: candidate function
ulong __OVERLOADABLE__ as_ulong (ushort4);
                       ^
cl_kernel.h:15512:24: note: candidate function
ulong __OVERLOADABLE__ as_ulong (int2);
                       ^
cl_kernel.h:15513:24: note: candidate function
ulong __OVERLOADABLE__ as_ulong (uint2);
                       ^
cl_kernel.h:15514:24: note: candidate function
ulong __OVERLOADABLE__ as_ulong (float2);
                       ^
In file included from <kernel>:5:
./wolf-skein.cl:34:14: error: call to 'as_ulong' is ambiguous
        else return(as_ulong(amd_bitalign(x.s10, x, 32 - (y - 32))));
                    ^~~~~~~~
cl_kernel.h:14824:24: note: candidate function
ulong __OVERLOADABLE__ as_ulong (long);
                       ^
cl_kernel.h:14825:24: note: candidate function
ulong __OVERLOADABLE__ as_ulong (ulong);
                       ^
cl_kernel.h:14826:24: note: candidate function
ulong __OVERLOADABLE__ as_ulong (double);
                       ^
cl_kernel.h:15502:24: note: candidate function
ulong __OVERLOADABLE__ as_ulong (char8);
                       ^
cl_kernel.h:15503:24: note: candidate function
ulong __OVERLOADABLE__ as_ulong (uchar8);
                       ^
cl_kernel.h:15505:24: note: candidate function
ulong __OVERLOADABLE__ as_ulong (short3);
                       ^
cl_kernel.h:15507:24: note: candidate function
ulong __OVERLOADABLE__ as_ulong (short4);
                       ^
cl_kernel.h:15509:24: note: candidate function
ulong __OVERLOADABLE__ as_ulong (ushort3);
                       ^
cl_kernel.h:15511:24: note: candidate function
ulong __OVERLOADABLE__ as_ulong (ushort4);
                       ^
cl_kernel.h:15512:24: note: candidate function
ulong __OVERLOADABLE__ as_ulong (int2);
                       ^
cl_kernel.h:15513:24: note: candidate function
ulong __OVERLOADABLE__ as_ulong (uint2);
                       ^
cl_kernel.h:15514:24: note: candidate function
ulong __OVERLOADABLE__ as_ulong (float2);
                       ^
<kernel>:316:26: warning: unknown OpenCL extension 'cl_amd_media_ops2' - ignoring
#pragma OPENCL EXTENSION cl_amd_media_ops2 :

Is it possible to define an emulated context for AMD GPUs on Nvidia systems so that the code can be run?

0 Kudos
2 Replies
Highlighted
39 Views

This forum only covers Intel(R) SDK for OpenCL applications.  I see that this code makes extensive use of the cl_amd_media_ops2 extension functions.  These are a vendor-specific extension to the Khronos OpenCL standard.  You can check which extensions are available in your environment when running with query code like in the Platform/Device Capabilities Viewer sample from our sample page.  https://software.intel.com/en-us/intel-opencl-support/code-samples.

 

0 Kudos
Highlighted
Beginner
39 Views

This is the output of CapsBasic:

Number of available platforms: 1
Platform names:
    [0] NVIDIA CUDA
There is no found platform with name containing "Intel" as a substring.

I guess there are no emulators for AMD-specific functions.

0 Kudos