Neuchips will show off its revolutionary Raptor Gen AI accelerator at CES 2024

Neuchips will be unveiling its revolutionary new Raptor Gen AI accelerator and Evo PCIe accelerator card LLM products at CES 2024, according to a new press release issued by the company.

2

Neuchips new Raptor Gen AI accelerator chip (source: Neuchips)

The upcoming Neuchips Raptor Gen AI accelerator was previously known as the N3000, with the new chip solution enabling enterprises to deploy large language models (LLMs) inference at a "fraction of the cost" of existing solutions. We will get the full skinny on Neuchips' new AI processor at CES 2024 next week.

Neuchips CEO Ken Lau said in the press release: "We are thrilled to unveil our Raptor chip and Evo card to the industry at CES 2024. Neuchips' solutions represent a massive leap in price to performance for natural language processing. With Neuchips, any organisation can now access the power of LLMs for a wide range of AI applications".

Neuchips will be demoing its new Raptor and Evo AI processors at CES 2024, accelerating the Whisper and Llama AI chatbots on a Personal AI Assistant application. Neuchips says it will be using its AI accelerators to show off the power of LLM inferencing for real business needs.

Now, down to business. Neuchips' upcoming Raptor Gen AI chip has 200 tera operations per second (TOPS) per chip, with its outstanding performance for AI inferencing operations like Matrix Multiply, Vector, and embedding table lookup suits Gen AI and transformer-based AI models. Neuchips says that this new groundbreaking throughput is achieved through Neuchips' patented compression and efficiency optimizations that were tailor-made for neural networks.

Neuchips' new Raptor AI processor will be joined by the ultra-low powered Evo acceleration card, which combines PCIe Gen5 with 8 lanes and 32GB of LPDDR5 memory which has 64GB/s host I/O bandwidth, and 1.6Tbps of memory bandwidth, all for just 55W per card. Very, very nice to see, Neuchips.

Evo also features 100% scalability, which will allow customers to linearly increase performance by easily adding more chips. Neuchips points out that this modular design helps ensure customers' investment production for future AI workloads.

Neuchips is also teasing its upcoming half-height, half-length (HHHL) form factor product, Viper, which will be launching in the second half of 2024. Viper will have "even greater deployment flexibility," which the company says brings data center-class AI acceleration in a compact, small design.

We will see Neuchips' new Raptor Gen AI and Evo AI acceleration cards at CES 2024, so if you're there, the company says to drop by its booth (62700) to "enroll in a free trial program'. Neuchips will have technical sessions that will showcase the power of its new Raptor and Evo products, where they can slash deployment costs for speech-to-text applications, and more.