Gopikrishnan
New Contributor
New ContributorMemory error when trying to simulate matrix multiplication using HLS math lib
component void test(ihc::stream_in<int> &matrixData,ihc::stream_in<int> &matrixData2,ihc::stream_out<int> &matrixout,bool ld,
int dim_in,int dim_out){
////////////input and output matrix definition/////////
hls_memory hls_singlepump hls_max_replicates(1) hls_bankwidth(sizeof(int))
int feat_matrix[N_FEAT][N_COLS_1];
hls_memory hls_singlepump hls_max_replicates(1) hls_bankwidth(sizeof(int))
static int weight_matrix[N_COLS_1][N_COLS_2];
int out_feat[N_FEAT][N_COLS_2];
///////////////////////////////////////////
//////////populate the feature and weight matrix//////////
for (int i=0; i <N_FEAT ; i++){
for(int j=0 ; j<N_COLS_1 ; j++){
feat_matrix[i][j] = matrixData.read();
}
}
if (ld){
for (int i=0; i <N_COLS_1 ; i++){
for(int j=0 ; j<N_COLS_2 ; j++){
bool success = false;
weight_matrix[i][j] = matrixData2.tryRead(success);
}
}
}
///////////////////////////////////////////////////////
////////MAT MUL AND WRITE BACK TO STREAM/////////////
matrix_multiply<int,N_FEAT,N_COLS_1,N_COLS_2,N_DSP,N_DSP>(feat_matrix,weight_matrix,out_feat);
for (int i=0; i <N_FEAT ; i++){
for(int j=0 ; j<dim_out ; j++){
matrixout.write(out_feat[i][j]);
}
}
////////////////////////////////////////////////////
////////////////////////////////////////////////////
}
When I'm trying to simulate this matrix multiplication for large dimensions it's giving me mem error
Fatal vsim4 memory allocation failure.
I figured out it's the loop unrolling that it's causing the issue. When I try to simulate the code using simple for loops with no unrolling it works fine. Is there any work around this or is there a better way of coding this
