Fast query processing in columnar databases with GPUs
According to one exemplary embodiment, a method for processing a query associated with a database is provided. The method may include receiving the query. The method may include estimating a number of groups. The method may include copying a plurality of data from the database to graphics processing unit (GPU) memory. The method may include creating a hash table in GPU memory. The method may include determining if a group associated with the database is present in the hash table. The method may include adding the group to the hash table based on determining that the group is not present in the hash table. The method may include aggregating a value associated with the group in the hash table based on determining that the group is present in the hash table. The method may include determining a plurality of results. The method may then include retrieving the plurality of results.
1. A method comprising:
receiving a group by/aggregate query;
estimating a number of groups based on the received query and a database;
copying data tuples of columnar data from the database to graphics processing unit (GPU) memory associated with a GPU;
creating a global hash table in GPU memory based on the estimated number of groups;
determining a threshold value, wherein the threshold value is determined based on one or more of: a finite number, a data type attribute, and a number of aggregation functions, based on determining that the number of groups in the global hash table does not exceed the threshold value;
creating a plurality of hash tables in shared memory of stream multiprocessors;
inserting groups into the global hash table using a plurality of execution threads;
aggregating the group values using the GPU's atomic operations in parallel;
inserting data into each of the multiple hash tables via the multiple execution threads, wherein the inserting comprises creating partial data;
merging the multiple hash tables;
inserting the merged, multiple hash tables into the global hash table;
scanning the global hash table;
and
retrieving a plurality of results.
2. The method of claim 1 , wherein aggregating the group values is based on the received query.
3. The method of claim 1 , wherein estimating the number of groups comprises using an optimizer executed in a central processing unit (CPU).