Parallelism

TileDB is fully parallelized internally, i.e., it uses multiple threads to process in parallel the most heavyweight tasks. In this tutorial we explain how TileDB parallelizes the queries and VFS operations, outlining the configuration parameters that you can use to control the amount of parallelization. Note that here we cover only the most important areas, as TileDB parallelizes numerous other internal tasks.

Queries

Internally, TileDB utilizes parallelism wherever possible to optimize read and write query performance and provide scalability for hardware with more available parallelism. Both read and write queries are parallelized. The array schema, query size, and layout all impact the available parallelization within the query. TileDB uses the Intel Threaded Building Blocks library for efficient scheduling and load balancing.

Filtering

For filtering (such as compression) during read queries, TileDB uses the following nested parallelism strategy:

parallel_for_each attribute being read:
  parallel_for_each tile of the attribute overlapping the subarray:
    parallel_for_each chunk of the tile:
      filter chunk

The “chunks” of a tile are controlled by a TileDB filter list parameter that defaults to 64KB. See the Filters tutorial to see how.

The nested parallelism in reads allows for maximum utilization of the available cores for filtering (e.g. decompression), in either the case where the query intersects few large tiles or many small tiles. For writes, TileDB uses the same strategy as reads:

parallel_for_each attribute being written:
  parallel_for_each tile of the attribute being written:
    parallel_for_each chunk of the tile:
      filter chunk

The configuration parameter sm.num_async_threads controls the number of threads available for use by asynchronous queries. It does not impact the parallelism used in the for loops illustrated above. Instead, the sm.num_tbb_threads parameter impacts the for loops above, although it is not recommended to modify this configuration parameter from its default setting.

I/O

After the tiles are filtered during a write query (or before filtering during read queries), I/O operations are issued to the underlying persistent storage via TileDB’s virtual filesystem (VFS) layer. For write queries, the configuration parameter sm.num_writer_threads controls the maximum number of write operations that will be dispatched in parallel to the VFS layer. For reads, the configuration parameter is sm.num_reader_threads. Both default to 1.

For flexibility in controlling the number of threads used for these parallel calls into the VFS layer, TileDB does not use TBB for these operations. Instead, a list of read or write tasks is accumulated during the query (one task per tile being read or written, per attribute) and dispatched to a thread pool with a work queue. The size of the thread pool is determined by the above configuration parameters.

VFS Operations

The VFS layer is additionally parallelized independently of the queries. Read VFS operations are parallelized internally using a private thread pool. This allows large reads to be split up and performed in parallel. Parallel VFS IO is enabled by default, but can be controlled with the following config parameters.

  • vfs.max_parallel_ops: The maximum number of parallel operations a single VFS read will be split into.
  • vfs.min_parallel_size: The minimum number of bytes per parallel VFS operation. TileDB will not split a single VFS read such that the parallel operations are responsible for less than vfs.min_parallel_size bytes (unless the operation size is not evenly divided by vfs.min_parallel_size, in which case the “last” thread will read less data).

All VFS backends are currently parallelized for reads using this strategy. In addition, POSIX and Win32 backends are parallelized for writes, controlled by the same config parameters.

VFS-S3

The S3 backend is also parallelized internally for writes. When a VFS write operation is issued to an S3 URI, TileDB will internally buffer the data in memory until enough has been buffered to dispatch a write operation to the S3 service. Any buffered data is flushed when the VFS file is synced or closed. The following config parameter controls this behavior.

  • vfs.s3.multipart_part_size: TileDB will buffer vfs.max_parallel_ops * vfs.s3.multipart_part_size bytes locally before issuing vfs.max_parallel_ops parallel “multipart” upload requests to S3.

VFS-HDFS

Currently, the HDFS backend is not parallelized for writes. Reads are parallelized using the generic VFS parallel read strategy outlined above.