Go to the source code of this file.
Functions | |
void | cs_halo_cuda_pack_send_buffer_real (const cs_halo_t *halo, cs_halo_type_t sync_mode, cs_lnum_t stride, const cs_real_t var[], cs_real_t send_buffer[]) |
Pack cs_real_t halo data to send into dense buffer, using CUDA. More... | |
void cs_halo_cuda_pack_send_buffer_real | ( | const cs_halo_t * | halo, |
cs_halo_type_t | sync_mode, | ||
cs_lnum_t | stride, | ||
const cs_real_t | var[], | ||
cs_real_t | send_buffer[] | ||
) |
Pack cs_real_t halo data to send into dense buffer, using CUDA.
A local state and/or buffer may be provided, or the default (global) state and buffer will be used. If provided explicitely, the buffer must be of sufficient size.
[in] | halo | pointer to halo structure |
[in] | sync_mode | synchronization mode (standard or extended) |
[in] | stride | number of (interlaced) values by entity |
[in] | var | pointer to value array (device) |
[out] | send_buffer | pointer to send buffer, NULL for global buffer |