void _tile_dpbusd (__tile dst, __tile a, __tile b)
| Type | Value |
|---|---|
| Type | Tile |
| Header file | #include <immintrin.h> |
| Instruction | TDPBUSD tmm, tmm, tmm |
| CPUID flags | AMXINT8 |
Compute dot-product of bytes in tiles with a source/destination accumulator. Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in "a" with corresponding signed 8-bit integers in "b", producing 4 intermediate 32-bit results. Sum these 4 results with the corresponding 32-bit integer in "dst", and store the 32-bit result back to tile "dst".
AMX
Application-Targeted
DEFINE DPBD(c, x, y) {
tmp1 := ZeroExtend32(x.byte[0]) * SignExtend32(y.byte[0])
tmp2 := ZeroExtend32(x.byte[1]) * SignExtend32(y.byte[1])
tmp3 := ZeroExtend32(x.byte[2]) * SignExtend32(y.byte[2])
tmp4 := ZeroExtend32(x.byte[3]) * SignExtend32(y.byte[3])
RETURN c + tmp1 + tmp2 + tmp3 + tmp4
}
FOR m := 0 TO dst.rows - 1
tmp := dst.row[m]
FOR k := 0 TO (a.colsb / 4) - 1
FOR n := 0 TO (dst.colsb / 4) - 1
tmp.dword[n] := DPBD(tmp.dword[n], a.row[m].dword[k], b.row[k].dword[n])
ENDFOR
ENDFOR
write_row_and_zero(dst, m, tmp, dst.colsb)
ENDFOR
zero_upper_rows(dst, dst.rows)
zero_tileconfig_start()