MapD Core

A SQL engine for the fastest queries on large datasets

Millisecond Queries on Billions of Rows

MapD Core is an open-source SQL engine architected on GPUs. MapD Core exploits the parallel processing power of GPU hardware to query billions of rows in milliseconds using standard SQL. Because the MapD Platform leverages the full graphics pipeline for server-side visual rendering, its visual analytics are far faster and more responsive than technologies that execute queries on the GPU but then render them on CPUs.

Orders of Magnitude Faster

In third-party benchmark tests, the MapD Core SQL engine is orders of magnitude faster than even the fastest CPU solutions. That speed gives analysts interactive insights, enabling them to ask more questions and dig deeper into the data. MapD Core’s comparative query speeds are shown in the following chart:

See Details
Hide Details
Data source: Summary of the 1.1 Billion Taxi Rides Benchmarks at
Query 1 SELECT cab_type, count() FROM trips GROUP BY cab_type;
Query 2 SELECT passenger_count, avg(total_amount) FROM trips GROUP BY passenger_count;
Query 3 SELECT passenger_count, extract(year from pickup_datetime) AS pickup_year, count() FROM trips GROUP BY passenger_count, pickup_year;
Query 4 SELECT passenger_count, extract(year from pickup_datetime) AS pickup_year, cast(trip_distance as int) AS distance, count(*) AS the_count FROM trips GROUP BY passenger_count, pickup_year, distance ORDER BY pickup_year, the_count desc;
System configurations
MapD: 1 machine (16 cores, 512 GB RAM, 2 x 1TB SSD, 8 Nvidia Pascal Titan X GPUs)
Redshift: 6 machines (36 cores, 244 GB RAM, 16TB HDD, AWS ds2.8xlarge)
Presto: 50 machines (4 cores, 15 GB RAM, 100GB SSD, GCP n1-standard-4)
Spark: 11 machines (4 cores, 15 GB RAM, 2 X 40GB storage, AWS m3.xlarge)

Three Versions for Three Purposes: Enterprise, Community and Open Source

MapD Enterprise edition includes Core Enterprise, for operational workloads in business and governmental use cases in the Intelligence, Defense and Civilian sectors. Core Enterprise includes features not available in Core Community, such as: high availability, distributed (sharding), LDAP authentication and an ODBC connector. MapD Community edition comes with Core Community for non-commercial purposes on a single node implementation. MapD Core Open Source edition can be downloaded from GitHub. Users can then build their own instance of MapD Core and contribute code back to the open source project.
Millisecond Speed
MapD Core delivers extraordinary speeds with its LLVM compilation engine, 3-tier data caching and the parallel processing power of GPUs.
Freedom from Indexes
MapD Core uses columnar storage, so every column is an index. This eliminates the need for indexing data in order to optimize query performance.
Native Rendering
MapD Core uses GPUs for in-situ rendering of large amounts of data. This makes visualization both responsive and interactive, even with billions of data points.

MapD Core Technology Available In MapD Enterprise Edition

MapD Enterprise edition includes Enterprise Core features required by operating the MapD Platform at scale for mission-critical uses in business and government. The following features are available in MapD Enterprise but not in MapD Community edition, nor in MapD Open Source.
Distributed Scale-out
Distributed scale-out capabilities allow data size and query performance to scale linearly as servers are added. MapD distributes inserts seamlessly across the cluster while read queries are load balanced between servers for improved throughput.
High Availability
High Availability (HA) functionality replicates data across multiple servers for enterprise-grade resiliency and redundancy. HA helps an organization meet its service level agreements for performance and uptime by synchronizing servers in a guaranteed way.
Server-Side Rendering
MapD Core includes a powerful rendering engine which renders the results of SQL queries as PNG compressed image files. This allows users to see billions of records in pixel-perfect visualizations without the need to send multi-gigabyte result sets from server to client.
LDAP authentication and ODBC connectivity is also available in Core Enterprise, to facilitate integration with existing enterprise analytic environments.

MapD Core Technology Available in All Versions of MapD

The following MapD Core features are available in all versions of MapD: Enterprise, Community and Open Source.

Advanced Memory Management

The MapD Core SQL engine intelligently moves data between main memory and GPU VRAM as needed. The system can also leverage modern SSDs for persistent storage. This advanced memory management system enables the MapD Core SQL engine to deliver exceptional performance across different hardware configurations and dataset sizes.

Hybrid Execution

Although architected originally for the GPU, MapD Core fully exploits the performance of both GPUs and CPUs. The SQL engine can intelligently execute compiled queries across multiple GPUs and CPUs. MapD Core uses the same infrastructure to parallelize computation across all CPUs and GPUs, so it delivers superior performance even on CPU-only systems.

Native SQL

MapD Core uses native SQL, the data language your organization already knows, but it returns query results at speeds hundreds of times faster than CPU-based solutions. It can filter, group, and aggregate billions of rows of data, allowing for interactive exploration and ad-hoc analysis of those massive datasets.

Rapid Compilation

MapD Core SQL queries are compiled with a just-in-time (JIT) LLVM based compiler, and run as NVIDIA GPU and x64 CPU machine code. In-memory CPU-based databases often use slow interpreters or source-to-source compilers, while queries compiled in MapD deliver orders-of-magnitude performance improvements.

Schedule a Demonstration

Schedule a demonstration of the MapD Platform with a use case from your industry.