RaiderChip’s AI accelerator showcased at AMD’s booth in FPGA Conference Europe

Participants had the opportunity to chat in real-time with the Microsoft Phi-2 LLM, fully accelerated inside RaiderChip’s Gen AI Hardware IP core on an offline AMD Versal FPGA.

Spain, July 4th, 2024

FPGA Conference Europe 2024 was held in Munich from July 2nd to 4th. The event, which annually brings together the most significant companies in the sector, featured the participation of the Raiderchip team this year.

RaiderChip CTO explaining Generative AI acceleration to FPGA Conference participants

Invited by AMD to showcase at its booth the first licensable IP core for Generative AI hardware acceleration, it offered attendees the opportunity to interact with the recently released Phi-2 LLM Chat model, accelerated by RaiderChip’s GenAI v1 hardware on an AMD Versal FPGA.

“With this demo, we aim to demonstrate the possibility of accelerating Generative AI on low-cost and fully standalone FPGA devices, completely offline and independently. This can serve as the foundation for numerous products capable of intelligent autonomous functioning. Generative AI is not only an incredible chatbot, that is just the tip of the iceberg; indeed it has such capability because it is the best predictor available today, so it can also be used to predict events of any kind, from imminent failures to radar ground truths, beyond what was possible with current technology. Plus, its reasoning allows it to perform as an autonomous agent, always ready and available to act on sensor measurements and take actions that prevent catastrophic failures in isolated systems, to give just one industry example,” stated Victor Lopez, CTO of RaiderChip.

RaiderChip team at the FPGA Conference showcasing Generative AI demo at AMD's booth

Indeed, Raiderchip’s IP core, the first solution available and licensable for any FPGA, shows that these devices have everything necessary to be part of this new industrial revolution, serving as the ideal foundation for a wide range of products aimed at markets ranging from industrial automation to consumer, automotive, and defense.

The overwhelmingly positive reception from conference participants undoubtedly reflects the growing relevance of Generative Artificial Intelligence over the past year and how it has gradually permeated every sector.

RaiderChip GenAI v1 demo running the Phi-2 LLM model on a Versal FPGA on AMD's booth

RaiderChip GenAI v1 demo running the Phi-2 LLM model on a Versal FPGA at AMD's booth

More information at https://raiderchip.ai/technology/hardware-ai-accelerators