SiPearl: Rhea1 key features to accelerate HPC & AI inference

Attending ISC tradeshow in Hamburg (booth L22), SiPearl, the company building the high-performance low-power European microprocessor for HPC and AI inference announces the main features of Rhea1, its firstgeneration microprocessor.

Designed with high-performance energy-efficient arm® Neoverse V1 platform, Rhea1 will include in a single package:

80 arm® Neoverse V1 cores ensuring high compute performance and efficient
performance per watt. Each core includes 2 Scalable Vector Extension (SVE) of 256 bits
each, enabling fast vector computations while optimizing area and energy use;

  • Built-in High Bandwidth Memory, with 4 stacks of HBM, to provide a balanced solution
    ideal for HPC, big data and AI Inference applications, which are often memory bandwidth
  • 4 DDR5 interfaces supporting 2 DIMMs Per Channel (2DPC);
  • 104 lanes of PCIe Gen5 interface: up to 6 x16 lanes + 2 x4 lanes;
  • High-performance arm® Neoverse CMN-700 Coherent Mesh Network on Chip (NoC) to
    interconnect compute and I/O elements;
  • Support for Flat or Quadrant mode.
    Rhea1 is supported by a wide range of compilers, library and tools, from traditional programming languages such as C/C++, GO and RUST to modern AI frameworks such as TensorFlow or PyTorch.

    Rhea1 is perfectly suited to traditional HPC workloads – its initial target market – and also to AI inference workloads. Thanks to both the generous memory capacity and high bandwidth of in-package HBM, it will deliver extraordinary performance and energy-efficiency with an unrivalled byte-per-flop ratio.

    First samples in 2025.

    “Combining the performance and energy-efficiency of arm® Neoverse V1 cores with inpackage HBM and embedding SiPearl patented memory and power management schemes, Rhea1 will fulfill the mission entrusted by EuroHPC JU and the European Processor Initiative consortium: to bring dedicated high-performance microprocessor technologies back to Europe. Rhea1 will be a world class microprocessor for HPC and AI inference. In the fastgrowing generative AI market, it will be a great alternative to existing solutions for AI inference workloads at lower cost while offering higher flexibility to model changes”, concluded Philippe Notton, CEO and founder of SiPearl.