MW_logo

Data Analytics Engine

Next-generation OLAP analytics engine that uses the massively parallel processing power of GPUs to speed up SQL operations

Read More

Features

  • Native support for SQL

    Native support for SQL

    MegaWise supports standard SQL and PostgreSQL grammar. The following interfaces are provided for MegaWise:
    - Command-line interface
    - JDBC and ODBC

  • Multilayer data cache

    Multilayer data cache

    MegaWise completely optimized the memory and the computing layer for the best performance. MegaWise builds 3 layers of cache in each physical node, including GPU video memory, main memory, and SSD, thus reaching the highest access speed.

  • Dynamic query compilation

    Dynamic query compilation

    ZILLIZ builds a JIT compilation system based on LLVM in MegaWise. Currently, MegaWise can run with high efficiency in NVIDIA GPU, x64 CPU, POWER CPU, and ARM CPU.

  • No complex indexing, <br /> downsampling, or pre-aggregation

    No complex indexing,
    downsampling, or pre-aggregation

    MegaWise does not need indexing, downsampling, or pre-aggregation. With the support of GPUs, MegaWise can perform real-time queries for billion-scale datasets in tens or hundreds of milliseconds.

Learn more about how the GPU-accelerated analytics database speeds up your geospatial and temporal data analytics.

View Doc
Performance Comparison
MegaWise -> 1 node/48 cores/4 GPU/512G Mem/1TB SSD
Redshift -> 6 nodes/192 cores/0 GPU/1464G Mem/15.4TB SSD
Spark 2.4 -> 21 nodes/84 cores/0 GPU/315G Mem/1.7TB SSD
Presto 0.214 -> 21 nodes/84 cores/0 GPU/315G Mem/1.7TB SSD
 
354.0x
2.3x
1.5x
1

Exploit the Potential of Your Data to the Fullest

Contact Us