github/concrete
1
1
Fork 0
mirror of https://github.com/zama-ai/concrete.git synced 2026-02-09 20:25:34 -05:00
Code Issues Packages Projects Releases Wiki Activity

fix(cuda): fix asynchronous behaviour for pbs and wop pbs

Browse Source
This commit is contained in:
Agnes Leroy
2023-01-23 15:55:57 +01:00
committed by Agnès Leroy
parent 2fcc5b2d0f
commit bd9cbbc7af
20 changed files with 81 additions and 259 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
include/bootstrap.h
View File

@@ -5,7 +5,8 @@
extern "C" {
void cuda_initialize_twiddles(uint32_t polynomial_size, uint32_t gpu_index);
void cuda_initialize_twiddles(uint32_t polynomial_size, void *v_stream,
uint32_t gpu_index);
void cuda_convert_lwe_bootstrap_key_32(void *dest, void *src, void *v_stream,
uint32_t gpu_index,

2
include/device.h
View File

@@ -33,4 +33,6 @@ int cuda_drop(void *ptr, uint32_t gpu_index);
int cuda_drop_async(void *ptr, cudaStream_t *stream, uint32_t gpu_index);
int cuda_get_max_shared_memory(uint32_t gpu_index);
int cuda_synchronize_stream(void *v_stream);
}

28
include/keyswitch.h
View File

@@ -15,23 +15,19 @@ void cuda_keyswitch_lwe_ciphertext_vector_64(
void *ksk, uint32_t lwe_dimension_in, uint32_t lwe_dimension_out,
uint32_t base_log, uint32_t level_count, uint32_t num_samples);
void cuda_fp_keyswitch_lwe_to_glwe_32(void *v_stream, void *glwe_array_out,
void *lwe_array_in, void *fp_ksk_array,
uint32_t input_lwe_dimension,
uint32_t output_glwe_dimension,
uint32_t output_polynomial_size,
uint32_t base_log, uint32_t level_count,
uint32_t number_of_input_lwe,
uint32_t number_of_keys);
void cuda_fp_keyswitch_lwe_to_glwe_32(
void *v_stream, uint32_t gpu_index, void *glwe_array_out,
void *lwe_array_in, void *fp_ksk_array, uint32_t input_lwe_dimension,
uint32_t output_glwe_dimension, uint32_t output_polynomial_size,
uint32_t base_log, uint32_t level_count, uint32_t number_of_input_lwe,
uint32_t number_of_keys);
void cuda_fp_keyswitch_lwe_to_glwe_64(void *v_stream, void *glwe_array_out,
void *lwe_array_in, void *fp_ksk_array,
uint32_t input_lwe_dimension,
uint32_t output_glwe_dimension,
uint32_t output_polynomial_size,
uint32_t base_log, uint32_t level_count,
uint32_t number_of_input_lwe,
uint32_t number_of_keys);
void cuda_fp_keyswitch_lwe_to_glwe_64(
void *v_stream, uint32_t gpu_index, void *glwe_array_out,
void *lwe_array_in, void *fp_ksk_array, uint32_t input_lwe_dimension,
uint32_t output_glwe_dimension, uint32_t output_polynomial_size,
uint32_t base_log, uint32_t level_count, uint32_t number_of_input_lwe,
uint32_t number_of_keys);
}
#endif // CNCRT_KS_H_