Arteris, Inc. (NASDAQ: AIP) witnessed a notable 7.04% surge in its stock price, reaching $5.78 during the extended trading session on Wednesday. The preceding regular session had seen Arteris’ stock decline by 0.92%, closing at $5.40. This uptick in AIP stock value has occurred in anticipation of the forthcoming release of its financial results scheduled for the upcoming week.
Arteris (AIP) has recently entered into a strategic partnership aimed at facilitating compatibility between Arteris’ FlexNoC and Ncore network-on-chip (NoC) development environment, and Fraunhofer IESE’s DRAM subsystem design space exploration framework. This collaboration is expected to yield enhancements in performance, cost reduction, and the acceleration of advanced DRAM-centric NoC development for their shared clientele.
The timely and precise modeling of the attributes of the most current DRAM architectures holds a pivotal role in achieving power-optimized SoC architectures. Through the establishment of interoperability and integration between Arteris’ FlexNoC and Ncore with DRAMSys4.0, clients can proactively assess the impact of cutting-edge DRAM technology on NoC performance and power consumption during the initial phases of their projects. This mitigates the risk of unexpected challenges necessitating architectural revisions later in the project lifecycle.
Generative AI Systems-on-Chip (SoCs) predominantly adopt memory-centric architectural designs. The performance and adaptability of Arteris’ interconnect IP products are well-suited to handling exceptionally high-bandwidth data flows, crucial for supporting advanced memory architectures facilitated by the Fraunhofer IESE memory exploration framework. As a combined effort, AIP empowers its clients to achieve cost savings and expedite the development of their distinct and performance-optimized NoC architectures.
Arteris has also recently disclosed that NeuReality has chosen to integrate Arteris FlexNoC interconnect IP into the NR1 network-addressable inference server-on-a-chip, a move designed to deliver marked improvements in performance, cost efficiency, and power consumption for machine and deep learning computing within its AI inference products.
This integration is structured within an 8-tier NoC configuration boasting an aggregated bandwidth of 4.5TB per second, effectively meeting low-latency prerequisites for running AI applications at scale, all while maintaining cost-effectiveness. The NeuReality inference server targets a range of AI workloads, including Generative AI, Large Language Models (LLMs), and others of similar nature.
