FK20 CUDA
|
Go to the source code of this file.
Functions | |
__global__ void | fk20_poly2toeplitz_coefficients (fr_t *toeplitz_coefficients, const fr_t *polynomial) |
polynomial -> toeplitz_coefficients More... | |
__global__ void fk20_poly2toeplitz_coefficients | ( | fr_t * | toeplitz_coefficients, |
const fr_t * | polynomial | ||
) |
polynomial -> toeplitz_coefficients
[out] | toeplitz_coefficients | array with dimension [4096 * gridDim.x] |
[in] | polynomial | array with dimensions [rows * 16 * 512] |
Grid must be 1-D, 256 threads per block.
IMPORTANT: This function does not need shared memory. Making the kernel call with a dynamic shared memory allocation is known to cause some subtle bugs, that not always show during normal execution. Similar comment is present in fk20test_poly.cu and fk20_512test_poly.cu. In case this function changes and starts needing shared memory, correct the tests on those two files.
Definition at line 23 of file fk20_poly2toeplitz_coefficients.cu.