VSORA Announces All-in-One Single Chip Architecture to Meet Generative AI Processing Challenges
PARIS –– July 19, 2023 –– VSORA, an innovative startup offering silicon intellectual property (IP) solutions and silicon chips, today unveiled Jotunn™, an all-in-one single-chip family to provide a massive jump in performance for Generative AI inferencing processing challenges.
“As ChatGPT and other generative AI continue reaching a fevered pitch, the semiconductor industry needs to step up with supporting hardware,” remarks Khaled Maalej, VSORA founder and CEO. “That’s what we intend to do with our single-chip architecture to trounce the memory wall and accelerate generative AI implementations.”
“The VSORA ChatGPT-on-chip architecture can unleash the unbounded potential of Generative AI such as ChatGPT,” states Nicolas von Bülow, managing partner at Clipperton Finance, a VSORA financial adviser. “It will spearhead a paradigm shift in the semiconductor industry.”
Jotunn can accelerate the proliferation of low-cost, low-power generative AI implementations whose monthly visit numbers have increased geometrically to close to two billion. Current leading processors idle almost 97% of the time waiting for processing data when implementing GPT-3 or GPT-4.
ChatGPT, based on GPT-3.5, can be handled by Jotunn4 entirely on-chip, dropping power consumption by more than an order of magnitude versus its competitors. Jotunn4 improves the cost effectiveness of a GPT-4 solution by more than 150 times compared to its closest competitor.
The Jotunn4 solution provides 1.6 petaflops with efficiencies of more than 50% for GPT-3.5 and GPT-4. Typically, efficiencies currently range around 2-4%. Further, Jotunn4 has 192 GB of on-chip memory and consumes less than 100W in a 45 x 45 mm package.
Initially designed as a low-power, low-cost and high-performance chip architecture for L3-L5 autonomous driving vehicles, VSORA extended its unique and patented architecture to accelerate GenAI applications.
Algorithm implementation is quick and easy, utilizing the VSORA high-level language and graph compiler development flow. The future-proof VSORA architecture is algorithm agnostic, allowing existing as well as new algorithms to be run on any VSORA chip.
Availability and Pricing
Jotunn4 is expected to ship in 2024.
About VSORA
VSORA provides high-performance silicon solutions for autonomous driving (AD) and intellectual property (IP) solutions for chipmakers designing the latest generations of artificial intelligence (AI), general high-end signal processing used in advanced driver-assistance systems (ADAS) and digital communications systems including 5G. Its powerful multicore architecture eliminates the need for digital signal processing (DSP) and/or AI co-processors and hardware accelerators to provide a level of flexibility achievable only with software programming. Founded in 2015 by a team of DSP and AI scientists and engineers with a long history of successes, VSORA has offices in Meudon-La-Forêt, France, San Diego, Calif., and Taiwan.
Related Semiconductor IP
- Dataflow AI Processor IP
- Powerful AI processor
- AI Processor Accelerator
- High-performance AI dataflow processor with scalable vector compute capabilities
- AI DSA Processor - 9-Stage Pipeline, Dual-issue
Related News
- If not ARM, which architecture next for IoT
- Synopsys' New Model for Infineon's Next Generation TriCore Architecture Accelerates Early Automotive Software Development and Test
- Flex Logix Appoints Tony Kozaczuk as Director of Solutions Architecture
- Cadence Announces General Availability of Tensilica Xtensa LX7 Processor Architecture, Increasing Floating-Point Scalability with 2 to 64 FLOPS/Cycle
Latest News
- Alchip Reports ASIC-Leading 2nm Developments
- AI Demand Drives 4Q25 Global Top 10 Foundries Revenue Up 2.6% QoQ; Samsung Gains Share and Tower Moves Up in Rankings
- GlobalFoundries Announces Availability of AutoPro150 eMRAM Technology on Enhanced FDX Platform for Advanced Automotive Applications
- Axiomise Launches nocProve for NoC Verification
- CAST Debuts TSN-EP-10G IP for High-Performance, Time-Sensitive Networking Ethernet Designs