MapD Core Database

The world's fastest in-memory, distributed GPU database powers the world's most immersive data exploration experience. Now open source.
The Database Designed for GPUs

MapD Core is an in-memory, column store, SQL relational database that was designed from the ground up to run on GPUs. The parallel processing power of GPU hardware enables the MapD Core database to query billions of rows in milliseconds using standard SQL.

Lightning Fast

The MapD Core database is orders of magnitude faster than even the fastest CPU solutions in third-party benchmark tests. Get insights quicker from your data with the parallel power of GPUs.

See Details
Hide Details
Data source: 10x copy of flights dataset (1.2B rows) at
Query 1 `select carrier_name, avg(arrdelay) from flights group by carrier_name`
Query 2 `select origin_name, dest_name, avg(arrdelay) from flights group by origin_name, dest_name`
Query 3 `select date_trunc(month,dep_timestamp) as ym, avg(arrdelay) as del from flights group by ym`
Query 4 `select dest_name, extract(month from dep_timestamp) as m, extract(year from dep_timestamp) as y, avg(arrdelay) as del from flights group by dest_name,y,m`
Query 5 `select count(*) from flights where origin_name='Lambert-St Louis International' and dest_name = 'Lincoln Municipal'`
System configurations
MapD: 1 machine (8 core, 384 GB RAM, 2 x 2TB SSD, 8 Nvidia K40)
In-memory DB 1: 10 machines (16 core, 64 GB RAM, EBS storage, m4.4xlarge)
In-memory DB 2: 3 machines (32 core, 244 GB RAM, 2 x 320GB SSD, r3.8xlarge)
Hadoop OLAP: 10 machines (16 core, 64 GB RAM, EBS storage, m4.4xlarge)
Benefits of GPU Power
Freedom from Indexes
The parallel power of MapD Core database means that indexes are not required to achieve the best query performance. Queries are just effortlessly fast, no index creation or maintenance required.
No Need to Downsample
There is no need to downsample data to reduce load on the system. GPU-powered MapD Core has the performance to process billions of rows and terabytes of in-memory data even on a single node.
Intelligent Scale
A single MapD server using GPUs can deliver the same work as dozens, if not hundreds, of CPU-only servers. For many workloads a single server is all that is required, if more capacity is needed a cluster of servers can be used.
Speaks SQL Natively

MapD Core is a native SQL database. It runs SQL, the data language your organization already knows, but at speeds hundreds of times faster than CPU-based solutions.

MapD Core database can filter, group, aggregate and join billions of rows of data in milliseconds, allowing for interactive ad-hoc exploration of the massive datasets. And it fits easily into existing data ecosystems, with support for a wide range of connectivity options to databases and data sources including JDBC, ODBC, Apache Thrift, Kafka, and Sqoop. And ODBC and JDBC support also provides SQL connectivity to third party tools including Tableau, Birst, MicroStrategy, and Qlik, again using standard SQL.

Enterprise Scale & Durability
Distributed Scale-out
The MapD Core database includes native distributed scale-out capability which allows database size and query performance to scale as servers are added. Thanks to the parallel processing power of GPUs, MapD Core customers can query and visualize larger datasets with much smaller cluster sizes than CPU-only solutions.
High Availability
The MapD Core database includes high availability functionality that gives customers enterprise-grade durability and redundancy by replicating data across multiple servers. Inserts are distributed seamlessly across the cluster while read queries are load balanced between servers for additional throughput.
MapD Core Deep Technology

Fast Compiled SQL

MapD Core database 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. MapD compiled queries deliver performance up to orders of magnitude faster.

Advanced Memory Management

MapD Core transparently manages the super-fast VRAM of the GPUs along with the main systems memory to achieve peak performance. Data is intelligently moved between main memory and GPU VRAM as needed. The system can also leverage modern fast SSDs for persistent storage. This advanced memory management enables the MapD Core database to deliver exceptional performance across different hardware configurations and database sizes.

Hybrid execution

MapD Core fully exploits the performance of both GPUs and CPUs. The database can simultaneously execute compiled queries across multiple GPUs and CPUs. And it uses the same infrastructure to parallelize computation across all CPUs as it does all GPUs—delivering leading performance even on CPU only systems.

MapD Core Rendering Engine

The MapD Core database includes a powerful rendering engine which renders the results of SQL queries as PNG compressed image files, all in-situ on the server GPU. This gives users the ability to visualize billions of records in pixel-perfect visualizations without the need to send multi-gigabyte result sets from server to client.

Schedule a Demonstration

Find out where speed might tip the competitive scales in your favor by getting a demonstration from our team of specialists.