STAC-M3 Benchmark Results: KX KDB+ 4.1 on Intel, Micron and Supermicro

Type: Audited

Specs: STAC-M3 Benchmarks (Antuco and Kanaga suites)

STAC recently performed a STAC-M3™ benchmark audit on a solution featuring the KDB+ database system sharded across six Supermicro Storage SuperServer SSG-222B-NE3X24R servers. Each system contained 24 Micron 9550 MAX NVMe SSDs and two Intel Xeon 6767P 64-core CPUs. Unlike many previous tests using a shared disk array, this solution used local storage across all nodes.

STAC-M3 is the industry standard enterprise tick-analytics benchmark for database and hardware stacks that manage large time series of market data. It provides a vendor-neutral framework for evaluating the performance, efficiency, and scalability of time-series databases used in financial analytics and backtesting.

The benchmark suite includes two sets of workloads: Antuco, the baseline suite for performance under fixed conditions, and Kanaga, an optional scaling suite that tests responsiveness and throughput at larger data volumes and user loads.

The System Under Test (SUT) was:

  • STAC Pack
    • STAC-M3 Pack for kdb+, Rev I
  • Database
    • KDB 4.1 (2025.02.18) running in shard mode
    Database nodes
    • 6x Supermicro Storage SuperServer SSG-222B-NE3X24R
    • 2 x Intel® Xeon® 6767P @ 2.40GHz
    • 16 x 128GB DDR5-6400 @ 6400MT/s [2 TB total]
    • NVIDIA® ConnectX-7 400Gb IPoIB
    • OS RHEL9.6 (Plow)
  • Storage system (direct attached / local to each node)
    • 24x 12.8 TB Micron® 9550 MAX NVMe™ SSD per server
    • PCIe® Gen5
    • 144 SSDs in total, 23x6 were used for data (1.84PB)
  • Network
    • MQM9700-NS2F, NVIDIA® 64-Port NDR 400G InfiniBand Data Center Switch

Unlike many previous tests using a shared disk array, this configuration employed local NVMe storage across all nodes, enabling low-latency access and distributed parallel performance.

The following is a summary of key results:

  • This solution outperformed all publicly disclosed results on 19 of 24 Kanaga mean-time response benchmarks.
    • STAC-M3.β1.100T.YR{1,2,3,4,5}VWAB-12D-HO.TIME
    • STAC-M3.β1.1T.{2,3,4,5}YRHIBID.TIME
    • STAC-M3.β1.100T.YR{1,2,3,4,5}-12D-HO.TIME
    • STAC-M3.β1.50T.YR{1,2,3,4,5}VWAB-12D-HO.TIME
  • This solution outperformed all publicly disclosed results on 3 of 3 Antuco 50-user and 100-user benchmarks.
    • 1.36x faster than a DDN based SUT (KDB221014) on STAC-M3.β1.100T.STATS-UI.TIME
    • 2.12x faster than an Intel Optane based SUT (KDB210428) on STAC-M3.β1.100T.VWAB-12D-NO.TIME
    • 1.29x faster than an Intel Optane based SUT (KDB210428) on STAC-M3.β1.50T.STATS-UI.TIME
  • This solution outperformed all publicly disclosed results on 3 of 5 Kanaga throughput benchmarks.
    • 1.77x faster than a Pure Storage based SUT (KDB231122) on STAC-M3.β1.1T.3YRHIBID.MBPS
    • 2.70x faster than a Pure Storage based SUT (KDB231122) on STAC-M3.β1.1T.4YRHIBID.MBPS
    • 2.51x faster than a Pure Storage based SUT (KDB231122) on STAC-M3.β1.1T.5YRHIBID.MBPS

The benchmark report is available to all STAC Observer members. STAC insights Subscribers have access to extensive visualizations of all test results, micro-detailed system configuration information, and the code used in this project—allowing them to run these same benchmarks in their own labs.

Please log in to access the report. To learn about subscription options, please contact us.

Please log in to see file attachments. If you are not registered, you may register for no charge.

The STAC-M3 Benchmark suite is the industry standard for testing solutions that enable high-speed analytics on time series data, such as tick-by-tick market data (aka "tick database" stacks).

Read more