Skip to main content

Stream Computing partners with VRULL and Software Ecosystems Solutions to build a richer software ecosystem for RISC-V AI application

By June 28, 2024July 2nd, 2024No Comments

By Makeljana Shkurti, SES

VRULL, a company headquartered in Austria that provides consulting services and outsourced R&D to semiconductor companies, is entering a strategic partnership with Stream Computing to improve programmer productivity in AI/ML accelerators featuring the Stream Computing NPU IP.

A European Platform for AI/ML Innovation

This collaboration aims to establish a robust European support ecosystem for the Stream Computing NPU solution, to enhance the region’s technological capabilities, and to foster innovation.

is platform will feature localized documentation and a dedicated developer portal, ensuring European developers can access tailored resources that meet their specific needs. Additionally, the partnership will provide application examples and pre-trained models customized for European applications, facilitating faster and more efficient deployment of AI solutions.

“Creating a European ecosystem for the Stream Computing NPU will leverage local know-how, and build documentation, technical content, tutorials and examples for applications found in our region. Working with our partners and universities will help to accelerate development of AI/ML talent in Europe.”, explains Makeljana Shkurti, Managing Director of Software Ecosystems Solutions Sh.a.

Alignment with Emerging Industry Standards

The collaboration between VRULL and Stream Computing is aligned with emerging industry standards, promising seamless integration with widely adopted middleware technologies. This includes support for LLVM-MLIR, IREE, OpenXLA, TensorFlow, and PyTorch.

By standardising on these platforms, the partnership aims to enhance return on investment and protect investments through a common software environment that will support both current-generation NPUs and future advancements in RISC-V standardisation for matrix and tensor operations.

“AI/ML is a hot spot of RISC-V innovation and ongoing standards-development. Building a future-proof enablement strategy for the Stream Computing will benefit all adopters of the production-proven NPU by ensuring investment protection and forward-compatibility in an environment that is characterised by rapid innovation.”, summarises Dr. Philipp Tomsich, Chief Technologist and Founder of VRULL GmbH.

Creating New Opportunities and Reducing Duplication

By enabling the reuse of technology across different geographies, this partnership aims to reduce duplication of effort and create new opportunities for Europe. VRULL’s proximity to users and applications positions it as a critical bridge between standards development, software enablement, and platform development, ensuring that European industries can leverage cutting-edge AI/ML technologies effectively.

“We work with our European partners, VRULL and SES, to build a richer software ecosystem for our NPU.  Leveraging their expertise in open-source software and standards-development will ease adoption and enhance user’s ROI by ensuring upstream open-source support.” states Andy Mei, CEO of Stream Computing.

Building a Global Community

Beyond Europe, the partnership is set to build a worldwide community spanning Europe, the Middle East, and the Americas. VRULL and SES will collaborate to create an open ecosystem rich with technical content, tutorials, and demos. By open-sourcing existing solutions and enhancing quality engineering, documentation, and user experience, the partnership aims to capture a broad user base and drive the adoption of Stream Computing’s hardware solutions.

Engagement with RISC-V Standardization

This strategic partnership between VRULL GmbH and Stream Computing represents a milestone in the development and adoption of AI/ML solutions across Europe. By leveraging their combined expertise and resources, the two companies are poised to drive unprecedented innovation, create new opportunities, and establish Europe as a leader in the global AI/ML arena.

“Working with VRULL, we can provide a future-proof software enablement and engage with RISC-V standardisation for matrix and tensor operations.”, adds David Chen, EVP of Ecosystem  and Business Development at Stream Computing

Stay Connected With RISC-V

We send occasional news about RISC-V technical progress, news, and events.