SiPearl

SiPearl

Electrónica/Telecomunicaciones, Ingeniería Especializada, Oficina de diseño e ingeniería

Maisons-Laffitte, Bologna, Castelldefels, Duisbourg, Grenoble, Sophia-Antipolis

Organization and methodologies

Fabless, SiPearl focuses on the design of high-performance, low-power microprocessors, initially dedicated to supercomputing and artificial intelligence inference.

After working closely with our customers to define their needs, we design our products and prepare them for production with our back-end partners, before handing them over to our manufacturer.

To design our microprocessors, we use Arm Neoverse technologies. We work with leading IP providers for high-speed interface, memory interface, and other market leading technologies.

The Siemens EDA Veloce Strato emulation platform we use is one of the most advanced in Europe.

Currently assigned to TSMC, our production will be on advanced technology sub 6 nm (less than 1/10,000 of the diameter of a human hair).

Projects and tech challenges

Rhea1

Rhea1

Rhea1

SiPearl is the only European company building a worldclass microprocessor, targeting firs supercomputing, AI-inference, later datacenter workloads.

To meet all the requirements for supercomputing and AI inference workloads with best-in-classe energy-efficiency, Rhea1, our first generation microprocessor, will integrate:

  • 80 arm® Neoverse V1 cores with 2 Scalable Vector Extension (SVE) units of 256 bits per core,
  • built-in High Bandwidth Memory with 4 stacks of HBM,
  • 4 DDR5 interfaces.

First samples of Rhea1 in 2025.

Supercomputing & Ai-inference

Supercomputing & AI inference

SiPearl has announced a design win for the first exascale supercomputer in Europe, Jupiter. The European supercomputers are used in modelling weather and climate, articifial intelligence, medical research (drug discovery and virus prevention for exemple), energy management (renewable energy in particular) and scientific research.

Rhea1 microprocessor is perfectly suited to traditional supercomputing workloads - its initial target market - and also to AI inference workloads. Thanks to 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.

Supercomputing & Ai-inference

Recruitment process

✔️Discovery interview with one of our Talent Acquisition Specialists (45 min).

✔️ AssessFirst test - no need to worry, there is no wrong or right answer; our goal is to see beyond your resume (45 min).

✔️ 2 technical interviews (1 h each).

✔️ Interview with one HRBP in order to answer to your last questions (30 to 45 min).