72 lines
2.1 KiB
Common Lisp
72 lines
2.1 KiB
Common Lisp
|
#define TS 32
|
||
|
|
#define WPT 8
|
||
|
|
#define RTS (TS/WPT)
|
||
|
|
|
||
|
|
// super super slow sgemm kernel by heehoon
|
||
|
|
__kernel void sgemm(__global float *A, __global float *B, __global float *C, int M, int N, int K) {
|
||
|
|
|
||
|
|
// Thread identifiers
|
||
|
|
const int row = get_local_id(0); // Local row ID (max: TS)
|
||
|
|
const int col = get_local_id(1); // Local col ID (max: TS/WPT == RTS)
|
||
|
|
const int globalRow = TS*get_group_id(0) + row; // Row ID of C (0..M)
|
||
|
|
const int globalCol = TS*get_group_id(1) + col; // Col ID of C (0..N)
|
||
|
|
|
||
|
|
// Local memory to fit a tile of TS*TS elements of A and B
|
||
|
|
__local float Asub[TS][TS];
|
||
|
|
__local float Bsub[TS][TS];
|
||
|
|
|
||
|
|
// Initialise the accumulation registers
|
||
|
|
float acc[WPT];
|
||
|
|
for (int w=0; w<WPT; w++) {
|
||
|
|
acc[w] = 0.0f;
|
||
|
|
}
|
||
|
|
|
||
|
|
// Loop over all tiles
|
||
|
|
const int numTiles = K/TS;
|
||
|
|
//const int numTiles = (K+TS-1)/TS;
|
||
|
|
for (int t=0; t<numTiles+1 ; t++) {
|
||
|
|
|
||
|
|
// Load one tile of A and B into local memory
|
||
|
|
for (int w=0; w<WPT; w++) {
|
||
|
|
const int tiledRow = TS*t + row;
|
||
|
|
const int tiledCol = TS*t + col;
|
||
|
|
//if(row+w*RTS>=32) printf("That is out!");
|
||
|
|
//if(col>=32) printf("That is out!");
|
||
|
|
|
||
|
|
if(((globalRow + w*RTS) >= M) || (tiledCol >= K)){
|
||
|
|
Asub[row + w*RTS][col] = 0.0f;
|
||
|
|
}
|
||
|
|
else{
|
||
|
|
Asub[row + w*RTS][col] = A[(globalRow + w*RTS)*K + tiledCol];
|
||
|
|
}
|
||
|
|
if(((tiledRow + w*RTS) >= K) || (globalCol >= N)){
|
||
|
|
Bsub[row + w*RTS][col] = 0.0f;
|
||
|
|
}
|
||
|
|
else{
|
||
|
|
Bsub[row + w*RTS][col] = B[(tiledRow + w*RTS)*N + globalCol];
|
||
|
|
}
|
||
|
|
}
|
||
|
|
|
||
|
|
// Synchronise to make sure the tile is loaded
|
||
|
|
barrier(CLK_LOCAL_MEM_FENCE);
|
||
|
|
|
||
|
|
// Perform the computation for a single tile
|
||
|
|
for (int k=0; k<TS; k++) {
|
||
|
|
for (int w=0; w<WPT; w++) {
|
||
|
|
acc[w] += Asub[row + w*RTS][k] * Bsub[k][col];
|
||
|
|
}
|
||
|
|
}
|
||
|
|
|
||
|
|
// Synchronise before loading the next tile
|
||
|
|
barrier(CLK_LOCAL_MEM_FENCE);
|
||
|
|
}
|
||
|
|
|
||
|
|
// Store the final results in C
|
||
|
|
for (int w=0; w<WPT; w++) {
|
||
|
|
if(((globalRow + w*RTS) <M) && (globalCol < N)){
|
||
|
|
C[(globalRow + w*RTS)*N + globalCol] = acc[w];
|
||
|
|
}
|
||
|
|
}
|
||
|
|
|
||
|
|
}
|