Forum Discussion

Mickleman's avatar
Mickleman
Icon for Occasional Contributor rankOccasional Contributor
5 years ago

Achieving parallel execution of loop on FPGA

I am having difficulty persuading the compiler to execute an outer loop in parallel in an FPGA kernel. I have constructed a simple example to illustrate the issue. Here is a simple array initialisa...
MGRAV's avatar
MGRAV
Icon for New Contributor rankNew Contributor
5 years ago

Hi,

Did you try with "#pragma unroll 2" ?

I am working on something similar and I would say that something like that should almost work

const int ITEM_LENGTH = 1000;
const int GROUP_SIZE = 10;
uint16_t a[2][GROUP_SIZE][ITEM_LENGTH];

#pragma unroll 2
[[intel::ivdep]]
for (int block = 0; block < 2; block++)

  [[intel::ivdep]]
  for (int j = 0; j < GROUP_SIZE; j++)
    for (int i = 0; i < ITEM_LENGTH; i++)
      if ( i == 0 )
        a[block][j][i] = j;
      else
        a[block][j][i] = a[block][j][i-1] + i;

good luck

Mickleman's avatar
Mickleman
Icon for Occasional Contributor rankOccasional Contributor
5 years ago

Many thanks for replying MGRAV.

I have tried with the unroll pragma but as far as I can see it does not execute in parallel because of the memory dependency on the array - even though this is not a real dependency because each iteration accesses a distinct separate part of the array. I have only been able to confirm this by looking at the Graph Viewer as none of the FPGA runtime nodes is working at the moment.