SPRU514 User guide

SPRU514Z July 2001 – October 2023 SM320F28335-EP

3.15.3 __dmac Intrinsic

You can force the compiler to generate a DMAC instruction by using the __dmac intrinsic. In this case, you are fully responsible for ensuring that the data addresses are 32-bit aligned.

void __dmac(long *src1, long *src2, long &accum1, long &accum2, int shift);

Src1 and src2 must be 32-bit aligned addresses that point to int arrays.
Accum1 and accum2 are pass-by-reference longs for the two temporary accumulations. These must be added together after the intrinsic to compute the total sum.
Shift specifies how far to shift the products prior to each accumulation. Valid shift values are -1 for a left shift by 1, 0 for no shift, and 1-6 for a right shift by 1-6, respectively. Note that this argument is required, so you must pass 0 if you want no shift.

See Table 7-6 for details about the __dmac intrinsic.

Example 1:

int src1[2N], src2[2N]; // src1 and src2 are int arrays of at least size 2N
                        // You must ensure that both start on 32-bit 
                        // aligned boundaries.
{...}
int i;
long res = 0;
long temp = 0;
for (i=0; i < N; i++)            // N does not have to be a known constant
     __dmac(((long *)src1)[i], ((long *)src2)[i], res, temp, 0);
res += temp;

Example 2:

int *src1, *src2;     // src1 and src2 are pointers to int arrays of at 
                      // least size 2N. User must ensure that both are  
                      // 32-bit aligned addresses.
{...}
int i;
long res = 0;
long temp = 0;
long *ls1 = (long *)src1; 
long *ls2 = (long *)src2;
for (i=0; i < N; i++)     // N does not have to be a known constant
    __dmac(*ls1++, *ls2++, res, temp, 0);
res += temp;

In these examples, res holds the final sum of a multiply-accumulate operation on int arrays of length 2*N, with two computations being performed at a time.

Additionally, an optimization level >= -O2 must be used to generate efficient code. Moreover, if there is nothing else in the loop body as in these examples, the compiler generates a RPT || DMAC instruction, further improving performance.