Skip to content

Supermicro + Kinetica

Supermicro’s servers offer superior energy-efficient performance for compute-intensive data analytics, machine learning, and engineering applications.

Supermicro and Kinetica have partnered to provide a GPU-optimized range of servers that can run Kinetica. Those working in diverse sectors such as finance, logistics, telecommunications, energy, and retail can dramatically accelerate application performance with minimal investment in development with this supercomputing solution, which is comprised of Supermicro’s servers with NVIDIA GPUs and Kinetica’s GPU-accelerated analytics database.

Supermicro has servers that support Kinetica’s GPU database and NVIDIA GPUs across a wide range of high-performance servers. Supermicro’s computing platforms achieve higher parallel processing capability with NVIDIA’s newest Tesla Volta GPU workhorse, delivering the highest quality with extreme optimization for the most computationally-intensive applications.

Supermicro’s SuperComputing servers offer several key advantages:

  •        Up to 960 TFLOPs FP16 performance in a single SuperServer
  •        Unparalleled system NVIDIA Tesla density (up to 4 Tesla V100 in 1U)
  •        Highest scalability with NVIDIA GPUDirect RDMA capable servers
  •        Industry-leading Titanium Level (96%+) power efficiency
  •        Optimized GPU peer-to-peer with single-root architecture
  •        Performance up to 1/10th the cost and 1/20th the power and consumption of traditional CPU-only servers
  •        The most comprehensive V100 SXM portfolio in the market

Supermicro offers a wide range of X11 GPU servers that allow customers to accelerate their applications and innovations to address the most complex, real-world problems:

X11 GPUs for Deep Learning

Supermicro offers best–in-class technology designed for super-compute level performance for artificial intelligence and deep learning applications. It’s an ideal system for running Kinetica with deep learning frameworks like Caffe, Torch, and TensorFlow.