Skip to content

Commit

Permalink
fixed shared memory launch
Browse files Browse the repository at this point in the history
  • Loading branch information
Tonglin Chen committed May 2, 2019
1 parent 53dd65d commit 53add47
Showing 1 changed file with 1 addition and 1 deletion.
2 changes: 1 addition & 1 deletion groupby_hash.cu
Original file line number Diff line number Diff line change
Expand Up @@ -70,7 +70,7 @@ void groupby_hash_GPU(const int hash_size, const int* key_columns_h, int num_key
# ifdef TESLA
sharedMemPerBlock = 32 * 1024;
# endif
size_t max_capacity = sharedMemPerBlock - sizeof(unsigned int);
size_t max_capacity = sharedMemPerBlock - 48; // for some reason 48 is required for reserved variable
size_t s_len_table = max_capacity / (2*sizeof(int) + sizeof(Tval)*num_ops);
size_t sharedMemorySize = 0;
while (true) { // calculate the suitable length of shared memory table
Expand Down

0 comments on commit 53add47

Please sign in to comment.