All IPs > Processor > AI Processor
The AI Processor category within our semiconductor IP catalog is dedicated to state-of-the-art technologies that empower artificial intelligence applications across various industries. AI processors are specialized computing engines designed to accelerate machine learning tasks and perform complex algorithms efficiently. This category includes a diverse collection of semiconductor IPs that are built to enhance both performance and power efficiency in AI-driven devices.
AI processors play a critical role in the emerging world of AI and machine learning, where fast processing of vast datasets is crucial. These processors can be found in a range of applications from consumer electronics like smartphones and smart home devices to advanced robotics and autonomous vehicles. By facilitating rapid computations necessary for AI tasks such as neural network training and inference, these IP cores enable smarter, more responsive, and capable systems.
In this category, developers and designers will find semiconductor IPs that provide various levels of processing power and architectural designs to suit different AI applications, including neural processing units (NPUs), tensor processing units (TPUs), and other AI accelerators. The availability of such highly specialized IPs ensures that developers can integrate AI functionalities into their products swiftly and efficiently, reducing development time and costs.
As AI technology continues to evolve, the demand for robust and scalable AI processors increases. Our semiconductor IP offerings in this category are designed to meet the challenges of rapidly advancing AI technologies, ensuring that products are future-ready and equipped to handle the complexities of tomorrow’s intelligence-driven tasks. Explore this category to find cutting-edge solutions that drive innovation in artificial intelligence systems today.
Akida Neural Processor IP is a groundbreaking component offering a self-contained AI processing solution capable of locally executing AI/ML workloads without reliance on external systems. This IP's configurability allows it to be tailored to various applications, emphasizing space-efficient and power-conscious designs. Supporting both convolutional and fully-connected layers, along with multiple quantization formats, it addresses the data movement challenge inherent in AI, significantly curtailing power usage while maintaining high throughput rates. Akida is designed for deployment scalability, supporting as little as two nodes up to extensive networks where complex models can thrive.
The second generation of BrainChip's Akida platform expands upon its predecessor with enhanced features for even greater performance, efficiency, and accuracy in AI applications. This platform leverages advanced 8-bit quantization and advanced neural network support, including temporal event-based neural nets and vision transformers. These advancements allow for significant reductions in model size and computational requirements, making the Akida 2nd Generation a formidable component for edge AI solutions. The platform effectively supports complex neural models necessary for a wide range of applications, from advanced vision tasks to real-time data processing, all while minimizing cloud interaction to protect data privacy.
MetaTF is BrainChip's machine learning framework for developing systems on the Akida neural processor. Designed to aid in creating, training, and testing neural networks, MetaTF integrates seamlessly with TensorFlow models. Its key feature is the ability to convert CNN models to Spiking Neural Networks (SNN), facilitating low-latency, low-power operations suited for edge environments. By utilizing Python scripting and tools, MetaTF simplifies model conversion and optimization, delivering automatic CNN to SNN transitions without needing to learn new frameworks. MetaTF also includes various development tools, encompassing runtime simulation and robust testing environments.
The KL730 AI SoC is a state-of-the-art chip incorporating Kneron's third-generation reconfigurable NPU architecture, delivering unmatched computational power with capabilities reaching up to 8 TOPS. This chip's architecture is optimized for the latest CNN network models and performs exceptionally well in transformer-based applications, reducing DDR bandwidth requirements substantially. Furthermore, it supports advanced video processing functions, capable of handling 4K 60FPS outputs with superior image handling features like noise reduction and wide dynamic range support. Applications can range from intelligent security systems to autonomous vehicles and commercial robotics.
Axelera AI has crafted a PCIe AI acceleration card, powered by their high-efficiency quad-core Metis AIPU, to tackle complex AI vision tasks. This card provides an extraordinary 214 TOPS, enabling it to process the most demanding AI workloads. Enhanced by the Voyager SDK's streamlined integration capabilities, this card promises quick deployment while maintaining superior accuracy and power efficiency. It is tailored for applications that require high throughput and minimal power consumption, making it ideal for edge computing.
The NMP-750 is a high-performance accelerator designed for edge computing, particularly suited for automotive, AR/VR, and telecommunications sectors. It boasts an impressive capacity of up to 16 TOPS and 16 MB local memory, powered by a RISC-V or Arm Cortex-R/A 32-bit CPU. The three AXI4 interfaces ensure seamless data transfer and processing. This advanced accelerator supports multifaceted applications such as mobility control, building automation, and multi-camera processing. It's designed to cope with the rigorous demands of modern digital and autonomous systems, offering substantial processing power and efficiency for intensive computational tasks. The NMP-750's ability to integrate into smart systems and manage spectral efficiency makes it crucial for communications and smart infrastructure management. It helps streamline operations, maintain effective energy management, and facilitate sophisticated AI-driven automation, ensuring that even the most complex data flows are handled efficiently.
The Tianqiao-70 is engineered for ultra-low power consumption while maintaining robust computational capabilities. This commercial-grade 64-bit RISC-V CPU core presents an ideal choice for scenarios demanding minimal power usage without conceding performance. It is primarily designed for emerging mobile applications and devices, providing both economic and environmental benefits. Its architecture prioritizes low energy profiles, making it perfect for a wide range of applications, including mobile computing, desktop devices, and intelligent IoT systems. The Tianqiao-70 fits well into environments where conserving battery life is a priority, ensuring that devices remain operational for extended periods without needing frequent charging. The core maintains a focus on energy efficiency, yet it supports comprehensive computing functions to address the needs of modern, power-sensitive applications. This makes it a flexible component in constructing a diverse array of SoC solutions and meeting specific market demands for sustainability and performance.
The Metis M.2 AI accelerator module from Axelera AI is a cutting-edge solution for embedded AI applications. Designed for high-performance AI inference, this card boasts a single quad-core Metis AIPU that delivers industry-leading performance. With dedicated 1 GB DRAM memory, it operates efficiently within compact form factors like the NGFF M.2 socket. This capability unlocks tremendous potential for a range of AI-driven vision applications, offering seamless integration and heightened processing power.
Focused on the advancement of autonomous mobility, KPIT's ADAS and Autonomous Driving solutions aim to address the multifaceted challenges that come with higher levels of vehicle autonomy. Safety remains the top priority, necessitating comprehensive testing and robust security protocols to ensure consumer trust. Current development practices often miss crucial corner cases by concentrating largely on standard conditions. KPIT tackles these issues through a holistic, multi-layered approach. Their solutions integrate state-of-the-art AI-driven decision-making systems that extend beyond basic perception, enhancing system reliability and intelligence. They've established robust simulation environments to ensure feature development covers all conceivable driving scenarios, contributing to the broader adoption of Level 3 and up autonomous systems. The company also offers extensive validation frameworks combining various testing methodologies to continually refine and prove their systems. This ensures each autonomous feature is thoroughly vetted before deployment, firmly positioning KPIT as a trusted partner for automakers aiming to bring safe, reliable, and highly autonomous vehicles to market.
RaiderChip's GenAI v1 is a pioneering hardware-based generative AI accelerator, designed to perform local inference at the Edge. This technology integrates optimally with on-premises servers and embedded devices, offering substantial benefits in privacy, performance, and energy efficiency over traditional hybrid AI solutions. The design of the GenAI v1 NPU streamlines the process of executing large language models by embedding them directly onto the hardware, eliminating the need for external components like CPUs or internet connections. With its ability to support complex models such as the Llama 3.2 with 4-bit quantization on LPDDR4 memory, the GenAI v1 achieves unprecedented efficiency in AI token processing, coupled with energy savings and reduced latency. What sets GenAI v1 apart is its scalability and cost-effectiveness, significantly outperforming competitive solutions such as Intel Gaudi 2, Nvidia's cloud GPUs, and Google's cloud TPUs in terms of memory efficiency. This solution maximizes the number of tokens generated per unit of memory bandwidth, thus addressing one of the primary limitations in generative AI workflow. Furthermore, the adept memory usage of GenAI v1 reduces the dependency on costly memory types like HBM, opening the door to more affordable alternatives without diminishing processing capabilities. With a target-agnostic approach, RaiderChip ensures the GenAI v1 can be adapted to various FPGAs and ASICs, offering configuration flexibility that allows users to balance performance with hardware costs. Its compatibility with a wide range of transformers-based models, including proprietary modifications, ensures GenAI v1's robust placement across sectors requiring high-speed processing, like finance, medical diagnostics, and autonomous systems. RaiderChip's innovation with GenAI v1 focuses on supporting both vanilla and quantized AI models, ensuring high computation speeds necessary for real-time applications without compromising accuracy. This capability underpins their strategic vision of enabling versatile and sustainable AI solutions across industries. By prioritizing integration ease and operational independence, RaiderChip provides a tangible edge in applying generative AI effectively and widely.
Designed for entry-level server-class applications, the SCR9 is a 64-bit RISC-V processor core that comes equipped with cutting-edge features, such as an out-of-order superscalar pipeline, making it apt for processing-intensive environments. It supports both single and double-precision floating-point operations adhering to IEEE standards, which ensure precise computation results. This processor core is tailored for high-performance computing needs, with a focus on AI and ML, as well as conventional data processing tasks. It integrates an advanced interrupt system featuring APLIC configurations, enabling responsive operations even under heavy workloads. SCR9 supports up to 16 cores in a multi-cluster arrangement, each utilizing coherent multi-level caches to maintain rapid data processing and management. The comprehensive development package for SCR9 includes ready-to-deploy toolchains and simulators that expedite software development, particularly within Linux environments. The core is well-suited for deployment in entry-level server markets and data-intensive applications, with robust support for virtualization and heterogeneous architectures.
Ventana's Veyron V2 CPU represents the pinnacle of high-performance AI and data center-class RISC-V processors. Engineered to deliver world-class performance, it supports extensive data center workloads, offering superior computational power and efficiency. The V2 model is particularly focused on accelerating AI and ML tasks, ensuring compute-intensive applications run seamlessly. Its design makes it an ideal choice for hyperscale, cloud, and edge computing solutions where performance is non-negotiable. This CPU is instrumental for companies aiming to scale with the latest in server-class technology.
BrainChip's Akida IP is an innovative neuromorphic processor that emulates the human brain's functionalities to analyze essential sensor inputs at the acquisition point. By maintaining AI/ML processes on-chip, Akida IP minimizes cloud dependency, reducing latency and enhancing data privacy. The scalable architecture supports up to 256 nodes interconnected over a mesh network, each node equipped with configurable Neural Network Layer Engines (NPEs). This event-based processor leverages data sparsity to decrease operational requirements significantly, which in turn improves performance and energy efficiency. With robust customization and the ability to perform on-chip learning, Akida IP adeptly supports a wide range of edge AI applications while maintaining a small silicon footprint.
As the SoC that placed Kneron on the map, the KL520 AI SoC continues to enable sophisticated edge AI processing. It integrates dual ARM Cortex M4 CPUs, ideally serving as an AI co-processor for products like smart home systems and electronic devices. It supports an array of 3D sensor technologies including structured light and time-of-flight cameras, which broadens its application in devices striving for autonomous functionalities. Particularly noteworthy is its ability to maximize power savings, making it feasible to power some devices on low-voltage battery setups for extended operational periods. This combination of size and power efficiency has seen the chip integrated into numerous consumer product lines.
The NMP-350 is a cutting-edge endpoint accelerator designed to optimize power usage and reduce costs. It is ideal for markets like automotive, AIoT/sensors, and smart appliances. Its applications span from driver authentication and predictive maintenance to health monitoring. With a capacity of up to 1 TOPS and 1 MB of local memory, it incorporates a RISC-V/Arm Cortex-M 32-bit CPU and supports three AXI4 interfaces. This makes the NMP-350 a versatile component for various industrial applications, ensuring efficient performance and integration. Developed as a low-power solution, the NMP-350 is pivotal for applications requiring efficient processing power without inflating energy consumption. It is crucial for mobile and battery-operated devices where every watt conserved adds to the operational longevity of the product. This product aligns with modern demands for eco-friendly and cost-effective technologies, supporting enhanced performance in compact electronic devices. Technical specifications further define its role in the industry, exemplifying how it brings robust and scalable solutions to its users. Its adaptability across different applications, coupled with its cost-efficiency, makes it an indispensable tool for developers working on next-gen AI solutions. The NMP-350 is instrumental for developers looking to seamlessly incorporate AI capabilities into their designs without compromising on economy or efficiency.
The Yitian 710 Processor is an advanced Arm-based server chip developed by T-Head, designed to meet the extensive demands of modern data centers and enterprise applications. This processor boasts 128 high-performance Armv9 CPU cores, each coupled with robust caches, ensuring superior processing speeds and efficiency. With a 2.5D packaging technology, the Yitian 710 integrates multiple dies into a single unit, facilitating enhanced computational capability and energy efficiency. One of the key features of the Yitian 710 is its memory subsystem, which supports up to 8 channels of DDR5 memory, achieving a peak bandwidth of 281 GB/s. This configuration guarantees rapid data access and processing, crucial for high-throughput computing environments. Additionally, the processor is equipped with 96 PCIe 5.0 lanes, offering a dual-direction bandwidth of 768 GB/s, enabling seamless connectivity with peripheral devices and boosting system performance overall. The Yitian 710 Processor is meticulously crafted for applications in cloud services, big data analytics, and AI inference, providing organizations with a robust platform for their computing needs. By combining high core count, extensive memory support, and advanced I/O capabilities, the Yitian 710 stands as a cornerstone for deploying powerful, scalable, and energy-efficient data processing solutions.
This ultra-compact and high-speed H.264 core is engineered for FPGA platforms, boasting industry-leading size and performance. Capable of providing 1080p60 H.264 Baseline support, it accommodates various customization needs, including different pixel depths and resolutions. The core is particularly noted for its minimal latency of less than 1ms at 1080p30, a significant advantage over competitors. Its flexibility allows integration with a range of FPGA systems, ensuring efficient compression without compromising on speed or size. In one versatile package, users have access to a comprehensive set of encoding features including variable and fixed bit-rate options. The core facilitates simultaneous processing of multiple video streams, adapting to various compression ratios and frame types (I and P frames). Its support for advanced video input formats and compliance with ITAR guidelines make it a robust choice for both military and civilian applications. Moreover, the availability of low-cost evaluation licenses invites experimentation and custom adaptation, promoting broad application and ease of integration in diverse projects. These cores are especially optimized for low power consumption, drawing minimal resources in contrast to other market offerings due to their efficient FPGA design architecture. They include a suite of enhanced features such as an AXI wrapper for simple system integration and significantly reduced Block RAM requirements. Embedded systems benefit from its synchronous design and wide support for auxiliary functions like simultaneous stream encoding, making it a versatile addition to complex signal processing environments.
The Chimera GPNPU by Quadric redefines AI computing on devices by combining processor flexibility with NPU efficiency. Tailored for on-device AI, it tackles significant machine learning inference challenges faced by SoC developers. This licensable processor scales massively offering performance from 1 to 864 TOPs. One of its standout features is the ability to execute matrix, vector, and scalar code in a single pipeline, essentially merging the functionalities of NPUs, DSPs, and CPUs into a single core. Developers can easily incorporate new ML networks such as vision transformers and large language models without the typical overhead of partitioning tasks across multiple processors. The Chimera GPNPU is entirely code-driven, empowering developers to optimize their models throughout a device's lifecycle. Its architecture allows for future-proof flexibility, handling newer AI workloads as they emerge without necessitating hardware changes. In terms of memory efficiency, the Chimera architecture is notable for its compiler-driven DMA management and support for multiple levels of data storage. Its rich instruction set optimizes both 8-bit integer operations and complex DSP tasks, providing full support for C++ coded projects. Furthermore, the Chimera GPNPU integrates AXI Interfaces for efficient memory handling and configurable L2 memory to minimize off-chip access, crucial for maintaining low power dissipation.
The KL630 AI SoC represents Kneron's sophisticated approach to AI processing, boasting an architecture that accommodates Int4 precision and transformers, making it incredibly adept in delivering performance efficiency alongside energy conservation. This chip shines in contexts demanding high computational intensity such as city surveillance and autonomous operation. It sports an ARM Cortex A5 CPU and a specialized NPU with 1 eTOPS computational power at Int4 precision. Suitable for running diverse AI applications, the KL630 is optimized for seamless operation in edge AI devices, providing comprehensive support for industry-standard AI frameworks and displaying superior image processing capabilities.
Polar ID offers an advanced solution for secure facial recognition in smartphones. This system harnesses the revolutionary capabilities of meta-optics to capture a unique polarization signature from human faces, adding a distinct layer of security against sophisticated spoofing methods like 3D masks. With its compact design, Polar ID replaces the need for bulky optical modules and costly time-of-flight sensors, making it a cost-effective alternative for facial authentication. The Polar ID system operates efficiently under diverse lighting conditions, ensuring reliable performance both in bright sunlight and in total darkness. This adaptability is complemented by the system’s high-resolution capability, surpassing that of traditional facial recognition technologies, allowing it to function seamlessly even when users are wearing face coverings, such as glasses or masks. By incorporating this high level of precision and security, Polar ID provides an unprecedented user experience in biometric solutions. As an integrated solution, Polar ID leverages state-of-the-art polarization imaging, combined with near-infrared technology operating at 940nm, which provides robust and secure face unlock functionality for an increasing range of mobile devices. This innovation delivers enhanced digital security and convenience, significantly reducing complexity and integration costs for manufacturers, while setting a new standard for biometric authentication in smartphones and beyond.
The xcore.ai platform from XMOS is engineered to revolutionize the scope of intelligent IoT by offering a powerful yet cost-efficient solution that combines high-performance AI processing with flexible I/O and DSP capabilities. At its heart, xcore.ai boasts a multi-threaded architecture with 16 logical cores divided across two processor tiles, each equipped with substantial SRAM and a vector processing unit. This setup ensures seamless execution of integer and floating-point operations while facilitating high-speed communication between multiple xcore.ai systems, allowing for scalable deployments in varied applications. One of the standout features of xcore.ai is its software-defined I/O, enabling deterministic processing and precise timing accuracy, which is crucial for time-sensitive applications. It integrates embedded PHYs for various interfaces such as MIPI, USB, and LPDDR, enhancing its adaptability in meeting custom application needs. The device's clock frequency can be adjusted to optimize power consumption, affirming its cost-effectiveness for IoT solutions demanding high efficiency. The platform's DSP and AI performances are equally impressive. The 32-bit floating-point pipeline can deliver up to 1600 MFLOPS with additional block floating point capabilities, accommodating complex arithmetic computations and FFT operations essential for audio and vision processing. Its AI performance reaches peaks of 51.2 GMACC/s for 8-bit operations, maintaining substantial throughput even under intensive AI workloads, making xcore.ai an ideal candidate for AI-enhanced IoT device creation.
The EW6181 is a cutting-edge multi-GNSS silicon solution offering the lowest power consumption and high sensitivity for exemplary accuracy across a myriad of navigation applications. This GNSS chip is adept at processing signals from numerous satellite systems including GPS L1, Glonass, BeiDou, Galileo, and several augmentation systems like SBAS. The integrated chip comprises an RF frontend, a digital baseband processor, and an ARM microcontroller dedicated to operating the firmware, allowing for flexible integration across devices needing efficient power usage. Designed with a built-in DC-DC converter and LDOs, the EW6181 silicon streamlines its bill of materials, making it perfect for battery-powered devices, providing extended operational life without compromising on performance. By incorporating patent-protected algorithms, the EW6181 achieves a remarkably compact footprint while delivering superior performance characteristics. Especially suited for dynamic applications such as action cameras and wearables, its antenna diversity capabilities ensure exceptional connectivity and positioning fidelity. Moreover, by enabling cloud functionality, the EW6181 pushes boundaries in power efficiency and accuracy, catering to connected environments where greater precision is paramount.
The NaviSoC by ChipCraft is a highly integrated GNSS system-on-chip (SoC) designed to bring navigation technologies to a single die. Combining a GNSS receiver with an application processor, the NaviSoC delivers unmatched precision in a dependable, scalable, and cost-effective package. Designed for minimal energy consumption, it caters to cutting-edge applications in location-based services (LBS), the Internet of Things (IoT), and autonomous systems like UAVs and drones. This innovative product facilitates a wide range of customizations, adaptable to varied market needs. Whether the application involves precise lane-level navigation or asset tracking and management, the NaviSoC meets and exceeds market expectations by offering enhanced security and reliability, essential for synchronization and smart agricultural processes. Its compact design, which maintains high efficiency and flexibility, ensures that clients can tailor their systems to exact specifications without compromise. NaviSoC stands as a testament to ChipCraft's pioneering approach to GNSS technologies.
The NMP-550 is tailored for enhanced performance efficiency, serving sectors like automotive, mobile, AR/VR, drones, and robotics. It supports applications such as driver monitoring, image/video analytics, and security surveillance. With a capacity of up to 6 TOPS and 6 MB local memory, this accelerator leverages either a RISC-V or Arm Cortex-M/A 32-bit CPU. Its three AXI4 interface support ensures robust interconnections and data flow. This performance boost makes the NMP-550 exceptionally suited for devices requiring high-frequency AI computations. Typical use cases include industrial surveillance and smart robotics, where precise and fast data analysis is critical. The NMP-550 offers a blend of high computational power and energy efficiency, facilitating complex AI tasks like video super-resolution and fleet management. Its architecture supports modern digital ecosystems, paving the way for new digital experiences through reliable and efficient data processing capabilities. By addressing the needs of modern AI workloads, the NMP-550 stands as a significant upgrade for those needing robust processing power in compact form factors.
The SCR7 is a 64-bit RISC-V application core crafted to meet high-performance demands of applications requiring powerful data processing. Featuring a sophisticated dual-issue pipeline with out-of-order execution, it enhances computational efficiency across varied tasks. The core is equipped with a robust floating-point unit and supports extensive RISC-V ISA extensions for advanced computing capabilities. SCR7's memory system includes L1 to L3 caches, with options for expansive up to 16MB L3 caching, ensuring data availability and integrity in demanding environments. Its multicore architecture supports up to eight cores, facilitating intensive computational tasks across industries such as AI and machine learning. Ideal for high-performance computing and big data applications, the SCR7 leverages its advanced interrupt systems and intelligent memory management for seamless operation. Comprehensive development resources, from simulators to SDKs, augment its integration across Linux-based systems, accelerating project development timelines.
The RISC-V Core-hub Generators from InCore are tailored for developers who need advanced control over their core architectures. This innovative tool enables users to configure core-hubs precisely at the instruction set and microarchitecture levels, allowing for optimized design and functionality. The platform supports diverse industry applications by facilitating the seamless creation of scalable and customizable RISC-V cores. With the RISC-V Core-hub Generators, InCore empowers users to craft their own processor solutions from the ground up. This flexibility is pivotal for businesses looking to capitalize on the burgeoning RISC-V ecosystem, providing a pathway to innovation with reduced risk and cost. Incorporating feedback from leading industry partners, these generators are designed to lower verification costs while accelerating time-to-market for new designs. Users benefit from InCore's robust support infrastructure and a commitment to simplifying complex chip design processes. This product is particularly beneficial for organizations aiming to integrate RISC-V technology efficiently into their existing systems, ensuring compatibility and enhancing functionality through intelligent automation and state-of-the-art tools.
The Avispado core is a 64-bit in-order RISC-V processor that provides an excellent balance of performance and power efficiency. With a focus on energy-conscious designs, Avispado facilitates the development of machine learning applications and is prime for environments with limited silicon resources. It leverages Semidynamics' innovative Gazzillion Misses™ technology to address challenges with sparse tensor weights, enhancing energy efficiency and operational performance for AI tasks. Structured to support multiprocessor configurations, Avispado is integral in systems requiring cache coherence and high memory throughput. It is particularly suitable for setups aimed at recommendation systems due to its ability to manage numerous outstanding memory requests, thanks to its advanced memory interface architectures. Integration with Semidynamics' Vector Unit enriches its offering, allowing dense computations and providing optimal performance in handling vector tasks. The ability to engage with Linux-ready environments and support for RISC-V Vector Specification 1.0 ensures that Avispado integrates seamlessly into existing frameworks, fostering innovative applications in fields like data centers and beyond.
The Veyron V1 CPU is designed to meet the demanding needs of data center workloads. Optimized for robust performance and efficiency, it handles a variety of tasks with precision. Utilizing RISC-V open architecture, the Veyron V1 is easily integrated into custom high-performance solutions. It aims to support the next-generation data center architectures, promising seamless scalability for various applications. The CPU is crafted to compete effectively against ARM and x86 data center CPUs, providing the same class-leading performance with added flexibility for bespoke integrations.
The Jotunn8 is engineered to redefine performance standards for AI datacenter inference, supporting prominent large language models. Standing as a fully programmable and algorithm-agnostic tool, it supports any algorithm, any host processor, and can execute generative AI like GPT-4 or Llama3 with unparalleled efficiency. The system excels in delivering cost-effective solutions, offering high throughput up to 3.2 petaflops (dense) without relying on CUDA, thus simplifying scalability and deployment. Optimized for cloud and on-premise configurations, Jotunn8 ensures maximum utility by integrating 16 cores and a high-level programming interface. Its innovative architecture addresses conventional processing bottlenecks, allowing constant data availability at each processing unit. With the potential to operate large and complex models at reduced query costs, this accelerator maintains performance while consuming less power, making it the preferred choice for advanced AI tasks. The Jotunn8's hardware extends beyond AI-specific applications to general processing (GP) functionalities, showcasing its agility. By automatically selecting the most suitable processing paths layer-by-layer, it optimizes both latency and power consumption. This provides its users with a flexible platform that supports the deployment of vast AI models under efficient resource utilization strategies. This product's configuration includes power peak consumption of 180W and an impressive 192 GB on-chip memory, accommodating sophisticated AI workloads with ease. It aligns closely with theoretical limits for implementation efficiency, accentuating VSORA's commitment to high-performance computational capabilities.
The Dynamic Neural Accelerator (DNA) II offers a groundbreaking approach to enhancing edge AI performance. This neural network architecture core stands out due to its runtime reconfigurable architecture that allows for efficient interconnections between compute components. DNA II supports both convolutional and transformer network applications, accommodating an extensive array of edge AI functions. By leveraging scalable performance, it makes itself a valuable asset in the development of systems-on-chip (SoC) solutions. DNA II is spearheaded by EdgeCortix's patented data path architecture, focusing on technical optimization to maximize available computing resources. This architecture uniquely allows DNA II to maintain low power consumption while flexibly adapting to various task demands across diverse AI models. Its higher utilization rates and faster processing set it apart from traditional IP core solutions, addressing industry demands for more efficient and effective AI processing. In concert with the MERA software stack, DNA II optimally sequences computation tasks and resource distribution, further refining efficiency and effectiveness in processing neural networks. This integration of hardware and software not only aids in reducing on-chip memory bandwidth usage but also enhances the parallel processing ability of the system, catering to the intricate needs of modern AI computing environments.
Topaz FPGAs are designed for high-volume production applications where cost efficiency, compact form factor, and energy efficiency are paramount. These FPGAs integrate a set of commonly used features and protocols, such as MIPI, Ethernet, and PCIe Gen3, making them ideal for use in machine vision, robotics, and consumer electronics. With logic densities ranging from 52,160 to 326,080 logic elements, Topaz FPGAs provide versatile support for complex applications while keeping power consumption low.\n\nThe advanced Quantum™ compute fabric in Topaz allows for effective packing of logic in XLR cells, which enhances the scope for innovation and design flexibility. These FPGAs excel in applications requiring substantial computational resources without a hefty power draw, ensuring broad adaptability across various use cases. Topaz's integration capabilities allow for straightforward system expansion, enabling seamless scaling of operations from R&D phases to full production.\n\nThe Topaz FPGA family is engineered to cater to extended product life cycles, which is crucial for industries like automotive and industrial automation where long-term system stability is essential. With multiple package options, including small QFP packages for reduced BoM costs, Topaz FPGAs provide an economically attractive option while ensuring support for high-speed data applications. Efinix's commitment to maintaining a stable product supply until at least 2045 assures partners of sustained innovation and reliability.
The GenAI v1-Q from RaiderChip brings forth a specialized focus on quantized AI operations, reducing memory requirements significantly while maintaining impressive precision and speed. This innovative accelerator is engineered to execute large language models in real-time, utilizing advanced quantization techniques such as Q4_K and Q5_K, thereby enhancing AI inference efficiency especially in memory-constrained environments. By offering a 276% boost in processing speed alongside a 75% reduction in memory footprint, GenAI v1-Q empowers developers to integrate advanced AI capabilities into smaller, less powerful devices without sacrificing operational quality. This makes it particularly advantageous for applications demanding swift response times and low latency, including real-time translation, autonomous navigation, and responsive customer interactions. The GenAI v1-Q diverges from conventional AI solutions by functioning independently, free from external network or cloud auxiliaries. Its design harmonizes superior computational performance with scalability, allowing seamless adaptation across variegated hardware platforms including FPGAs and ASIC implementations. This flexibility is crucial for tailoring performance parameters like model scale, inference velocity, and power consumption to meet exacting user specifications effectively. RaiderChip's GenAI v1-Q addresses crucial AI industry needs with its ability to manage multiple transformer-based models and confidential data securely on-premises. This opens doors for its application in sensitive areas such as defense, healthcare, and financial services, where confidentiality and rapid processing are paramount. With GenAI v1-Q, RaiderChip underscores its commitment to advancing AI solutions that are both environmentally sustainable and economically viable.
NeuroMosAIc Studio is a comprehensive software platform designed to maximize AI processor utilization through intuitive model conversion, mapping, simulation, and profiling. This advanced software suite supports Edge AI models by optimizing them for specific application needs. It offers precision analysis, network compression, and quantization tools to streamline the process of deploying AI models across diverse hardware setups. The platform is notably adept at integrating multiple AI functions and facilitating edge training processes. With tools like the NMP Compiler and Simulator, it allows developers to optimize functions at different stages, from quantization to training. The Studio's versatility is crucial for developers seeking to enhance AI solutions through customized model adjustments and optimization, ensuring high performance across AI systems. NeuroMosAIc Studio is particularly valuable for its edge training support and comprehensive optimization capabilities, paving the way for efficient AI deployment in various sectors. It offers a robust toolkit for AI model developers aiming to extract the maximum performance from hardware in dynamic environments.
The Tyr Superchip is engineered to tackle the most daunting computational challenges in edge AI, autonomous driving, and decentralized AIoT applications. It merges AI and DSP functionalities into a single, unified processing unit capable of real-time data management and processing. This all-encompassing chip solution handles vast amounts of sensor data necessary for complete autonomous driving and supports rapid AI computing at the edge. One of the key challenges it addresses is providing massive compute power combined with low-latency outputs, achieving what traditional architectures cannot in terms of energy efficiency and speed. Tyr chips are surrounded by robust safety protocols, being ISO26262 and ASIL-D ready, making them ideally suited for the critical standards required in automotive systems. Designed with high programmability, the Tyr Superchip accommodates the fast-evolving needs of AI algorithms and supports modern software-defined vehicles. Its low power consumption, under 50W for higher-end tasks, paired with a small silicon footprint, ensures it meets eco-friendly demands while staying cost-effective. VSORA’s Superchip is a testament to their innovative prowess, promising unmatched efficiency in processing real-time data streams. By providing both power and processing agility, it effectively supports the future of mobility and AI-driven automation, reinforcing VSORA’s position as a forward-thinking leader in semiconductor technology.
The Ultra-Low-Power 64-Bit RISC-V Core by Micro Magic, Inc. is engineered to operate efficiently with minimal power consumption, making it a standout solution for high-performance applications. This processor core is capable of running at an impressive 5GHz, yet it only consumes 10mW at 1GHz, illustrating its capability to deliver exceptional performance while keeping power usage to a minimum. Ideal for scenarios where energy efficiency is crucial, it leverages advanced design techniques to reduce voltage alongside high-speed processing. Maximizing power efficiency without compromising speed, this RISC-V core is suited for a wide array of applications ranging from IoT devices to complex computing systems. Its design allows it to maintain performance even at lower power inputs, a critical feature in sectors that prioritize energy savings and sustainability. The core's architecture supports full configurability, catering to diverse design needs across different technological fields. In addition to its energy-efficient design, the core offers robust computational capabilities, making it a competitive choice for companies looking to implement high-speed, low-power processing solutions in their product lines. The flexibility and power of this core accentuate Micro Magic's commitment to delivering top-tier semiconductor solutions that meet the evolving demands of modern technology.
The Hanguang 800 AI Accelerator by T-Head is an advanced semiconductor technology designed to accelerate AI computations and machine learning tasks. This accelerator is specifically optimized for high-performance inference, offering substantial improvements in processing times for deep learning applications. Its architecture is developed to leverage parallel computing capabilities, making it highly suitable for tasks that require fast and efficient data handling. This AI accelerator supports a broad spectrum of machine learning frameworks, ensuring compatibility with various AI algorithms. It is equipped with specialized processing units and a high-throughput memory interface, allowing it to handle large datasets with minimal latency. The Hanguang 800 is particularly effective in environments where rapid inferencing and real-time data processing are essential, such as in smart cities and autonomous driving. With its robust design and multi-faceted processing abilities, the Hanguang 800 Accelerator empowers industries to enhance their AI and machine learning deployments. Its capability to deliver swift computation and inference results ensures it is a valuable asset for companies looking to stay at the forefront of technological advancement in AI applications.
aiWare stands out as a premier hardware IP for high-performance neural processing, tailored for complex automotive AI applications. By offering exceptional efficiency and scalability, aiWare empowers automotive systems to harness the full power of neural networks across a wide variety of functions, from Advanced Driver Assistance Systems (ADAS) to fully autonomous driving platforms. It boasts an innovative architecture optimized for both performance and energy efficiency, making it capable of handling the rigorous demands of next-generation AI workloads. The aiWare hardware features an NPU designed to achieve up to 256 Effective Tera Operations Per Second (TOPS), delivering high performance at significantly lower power. This is made possible through a thoughtfully engineered dataflow and memory architecture that minimizes the need for external memory bandwidth, thus enhancing processing speed and reducing energy consumption. The design ensures that aiWare can operate efficiently across a broad range of conditions, maintaining its edge in both small and large-scale applications. A key advantage of aiWare is its compatibility with aiMotive's aiDrive software, facilitating seamless integration and optimizing neural network configurations for automotive production environments. aiWare's development emphasizes strong support for AI algorithms, ensuring robust performance in diverse applications, from edge processing in sensor nodes to high central computational capacity. This makes aiWare a critical component in deploying advanced, scalable automotive AI solutions, designed specifically to meet the safety and performance standards required in modern vehicles.
ZIA Stereo Vision by Digital Media Professionals Inc. revolutionizes three-dimensional image processing by delivering exceptional accuracy and performance. This stereo vision technology is particularly designed for use in autonomous systems and advanced robotics, where precise spatial understanding is crucial. It incorporates deep learning algorithms to provide robust 3D mapping and object recognition capabilities. The IP facilitates extensive depth perception and analyzed spatial data for applications in areas like automated surveillance and navigation. Its ability to create detailed 3D maps of environments assists machines in interpreting and interacting with their surroundings effectively. By applying sophisticated AI algorithms, it enhances the ability of devices to make intelligent decisions based on rich visual data inputs. Integration into existing systems is simplified due to its compatibility with a variety of platforms and configurations. By enabling seamless deployment in sectors demanding high reliability and accuracy, ZIA Stereo Vision stands as a core component in the ongoing evolution towards more autonomous and smart digital environments.
ISPido represents a fully configurable RTL Image Signal Processing Pipeline, adhering to the AMBA AXI4 standards and tailored through the AXI4-LITE protocol for seamless integration with systems such as RISC-V. This advanced pipeline supports a variety of image processing functions like defective pixel correction, color filter interpolation using the Malvar-Cutler algorithm, and auto-white balance, among others. Designed to handle resolutions up to 7680x7680, ISPido provides compatibility for both 4K and 8K video systems, with support for 8, 10, or 12-bit depth inputs. Each module within this pipeline can be fine-tuned to fit specific requirements, making it a versatile choice for adapting to various imaging needs. The architecture's compatibility with flexible standards ensures robust performance and adaptability in diverse applications, from consumer electronics to professional-grade imaging solutions. Through its compact design, ISPido optimizes area and energy efficiency, providing high-quality image processing while keeping hardware demands low. This makes it suitable for battery-operated devices where power efficiency is crucial, without sacrificing the processing power needed for high-resolution outputs.
CodaCache is the last-level cache solution from Arteris, designed to solve significant system-on-chip design challenges, including performance bottlenecks, data access latency, and power efficiency constraints. By leveraging high-performance caching techniques, CodaCache effectively optimizes data flow and power consumption across complex SoC architectures, ensuring accelerated memory access times and improved overall system efficiency. This cache solution is highly configurable, enabling developers to fine-tune features such as cache associativity and partitioning, which is critical for maximizing performance in specific application scenarios. Moreover, CodaCache benefits from seamless integration with the Arteris NoC environment, facilitating streamlined data traffic management across integrated systems. The product supports real-time processing needs by enabling a scalable cache that addresses challenges in timing closure and system integration. Performance monitoring and hardware-supported coherency management features empower engineers with tools for enhanced control and monitoring, ensuring the cache operates at peak efficiency. CodaCache’s functional safety and resilience options further its use in critical applications where high reliability is mandatory.
The SAKURA-II AI accelerator is designed specifically to address the challenges of energy efficiency and processing demands in edge AI applications. This powerhouse delivers top-tier performance while maintaining a compact and low-power silicon architecture. The key advantage of SAKURA-II is its capability to handle vision and Generative AI applications with unmatched efficiency, thanks to the integration of the Dynamic Neural Accelerator (DNA) core. This core exhibits run-time reconfigurability that supports multiple neural network models simultaneously, adapting in real-time without compromising on speed or accuracy. Focusing on the demanding needs of modern AI applications, the SAKURA-II easily manages models with billions of parameters, such as Llama 2 and Stable Diffusion, all within a mere power envelope of 8W. It supports a large memory bandwidth and DRAM capacity, ensuring smooth handling of complex workloads. Furthermore, its multiple form factors, including modules and cards, allow for versatile system integration and rapid development, significantly shortening the time-to-market for AI solutions. EdgeCortix has engineered the SAKURA-II to offer superior DRAM bandwidth, allowing for up to 4x the DRAM bandwidth of other accelerators, crucial for low-latency operations and nimbly executing large-scale AI workflows such as Language and Vision Models. Its architecture promises higher AI compute utilization than traditional solutions, thus delivering significant energy efficiency advantages.
Emphasizing energy efficiency and processing power, the KL530 AI SoC is equipped with a newly developed NPU architecture, making it one of the first chips to adopt Int4 precision commercially. It offers remarkable computing capacity with lower energy consumption compared to its predecessors, making it ideal for IoT and AIoT scenarios. Embedded with an ARM Cortex M4 CPU, this chip enhances comprehensive image processing performance and multimedia codec efficiency. Its ISP capabilities leverage AI-based enhancements for superior image quality while maintaining low power usage during operation, thereby extending its competitiveness in fields such as robotics and smart appliances.
The KL720 AI SoC stands out for its excellent performance-to-power ratio, designed specifically for real-world applications where such efficiency is critical. Delivering nearly 0.9 TOPS per Watt, this chip underlines significant advancement in Kneron's edge AI capabilities. The KL720 is adept for high-performance devices like cutting-edge IP cameras, smart TVs, and AI-driven consumer electronics. Its architecture, based on the ARM Cortex M4 CPU, facilitates high-quality image and video processing, from 4K imaging to natural language processing, thereby advancing capabilities in devices needing rigorous computational work without draining power excessively.
The Spiking Neural Processor T1 is an innovative ultra-low power microcontroller designed for always-on sensing applications, bringing intelligence directly to the sensor edge. This processor utilizes the processing power of spiking neural networks, combined with a nimble RISC-V processor core, to form a singular chip solution. Its design supports next-generation AI and signal processing capabilities, all while operating within a very narrow power envelope, crucial for battery-powered and latency-sensitive devices. This microcontroller's architecture supports advanced on-chip signal processing capabilities that include both Spiking Neural Networks (SNNs) and Deep Neural Networks (DNNs). These processing capabilities enable rapid pattern recognition and data processing similar to how the human brain functions. Notably, it operates efficiently under sub-milliwatt power consumption and offers fast response times, making it an ideal choice for devices such as wearables and other portable electronics that require continuous operation without significant energy draw. The T1 is also equipped with diverse interface options, such as QSPI, I2C, UART, JTAG, GPIO, and a front-end ADC, contained within a compact 2.16mm x 3mm, 35-pin WLCSP package. The device boosts applications by enabling them to execute with incredible efficiency and minimal power, allowing for direct connection and interaction with multiple sensor types, including audio and image sensors, radar, and inertial units for comprehensive data analysis and interaction.
The C100 from Chipchain is a highly integrated, low-power consumption single-chip solution tailored for IoT applications. Featuring an advanced 32-bit RISC-V CPU capable of operating at speeds up to 1.5GHz, it houses embedded RAM and ROM for efficient processing and computational tasks. This chip's core strength lies in its multifunctional nature, integrating Wi-Fi, various transmission interfaces, an ADC, LDO, and temperature sensors, facilitating a streamlined and rapid application development process. The C100 chip is engineered to support a diverse set of applications, focusing heavily on expanding IoT capabilities with enhanced control and connectivity features. Beyond its technical prowess, the C100 stands out with its high-performance wireless microcontrollers, designed specifically for the burgeoning IoT market. By leveraging various embedded technologies, the C100 enables simplified, fast, and adaptive application deployment across a wide array of sectors including security, healthcare, smart home devices, and digital entertainment. The chip’s integrated features ensure it can meet the rigorous demands of modern IoT applications, characterized by high integration and reliability. Moreover, the C100 represents a leap forward in IoT product development with its extensive focus on energy efficiency, compact size, and secure operations. Providing a complete IoT solution, this chip is instrumental in advancing robust IoT ecosystems, driving innovation in smart connectivity. Its comprehensive integration provides IoT developers with a significant advantage, allowing them to develop solutions that are not only high-performing but also ensure sustainability and user safety.
Wormhole is a versatile communication system designed to enhance data flow within complex computational architectures. By employing state-of-the-art connectivity solutions, it enables efficient data exchange, critical for high-speed processing and low-latency communication. This technology is essential for maintaining optimal performance in environments demanding seamless data integration. Wormhole's ability to manage significant data loads with minimal latency makes it particularly suitable for applications requiring real-time data processing and transfer. Its integration into existing systems can enhance overall efficiency, fostering a more responsive computational environment. This makes it an invaluable asset for sectors undergoing digital transformation. The adaptability of Wormhole to various technological requirements ensures it remains relevant across diverse industry applications. This flexibility means that it can scale with ongoing technological advancements, cementing its role as a cornerstone in the evolving landscape of high-speed data communications.
The RWM6050 Baseband Modem from Blu Wireless is integral to their high bandwidth, high capacity mmWave solutions. Designed for cost-effectiveness and power efficiency, this modem forms a central component of multi-gigabit radio interfaces. It provides robust connectivity for access and backhaul markets through its notable flexibility and high performance. Partnering with mmWave RF chipsets, the RWM6050 delivers flexible channelisation modes and modulation coding capabilities, enabling it to handle extensive bandwidth requirements and achieve multi-gigabit data rates. This is supported by dual modems that include a mixed-signal front-end, enhancing its adaptability across a vast range of communications environments. Key technical features include integrated network synchronization and a programmable real-time scheduler. These features, combined with advanced beam forming support and digital front-end processing, make the RWM6050 a versatile tool in optimizing connectivity solutions. The modem's specifications ensure high efficiency in various network topologies, highlighting its role as a crucial asset in contemporary telecommunications settings.
The SiFive Intelligence X280 processor is crafted for the demands of AI and ML within edge computing environments. It integrates high-performance scalar and vector computing capabilities, making it ideal for data-heavy AI tasks like management, object detection, and speech processing. The X280 leverages the RISC-V architecture's open standards, bringing a high level of customizability and performance efficiency to AI applications. Equipped with SiFive's Matrix Engine, the X280 is capable of handling sophisticated AI workloads with its impressive maximum throughput of 16 TOPS for INT8 operations. This performance is achieved without compromising on power efficiency, maintaining a small footprint that makes it suitable for diverse deployment scenarios. The processor's scalability is a key feature, supporting vector lengths up to 512 bits to accommodate the demands of intensive machine learning operations. SiFive Intelligence X280 stands out for its role in reshaping the possibilities of AI at the edge, pushing forward the capabilities of machine learning with a comprehensive software and hardware integration. This approach ensures that the X280 can handle emerging AI challenges with ease, presenting a formidable solution for today's AI-driven applications.
The CTAccel Image Processor for Alveo U200 represents a pinnacle of image processing acceleration, catering to the massive data produced by the explosion of smartphone photography. Through the offloading of intensive image processing tasks from CPUs to FPGAs, it achieves notable gains in performance and efficiency for data centers. By using an FPGA as a heterogenous coprocessor, the CIP speeds up typical workflows—such as image encoding and decoding—up to six times, while drastically cutting latency by fourfold. Its architecture allows for expanded compute density, meaning less rack space and reduced operational costs for managing data centers. This is crucial for handling the everyday influx of image data driven by social media and cloud storage. The solution maintains full software compatibility with popular tools like ImageMagick and OpenCV, meaning migration is seamless and straightforward. Moreover, the system's remote reconfiguration capabilities enable users to optimize processing for varying scenarios swiftly, ensuring peak performance without the need for server restarts.
Efinix's Trion FPGAs provide an ideal solution for edge computing and IoT applications, where power efficiency, speed, and integration capabilities are critical. The Trion family, built on a 40 nm process node, offers a range of devices with logic elements between 4K to 120K, catering to both simple and complex application needs. Their comprehensive interface support, including MIPI, DDR, and LVDS, enhances their suitability for high-bandwidth applications in communication, consumer, and industrial sectors.\n\nTrion FPGAs are designed for high integration in space-constrained environments. The small package sizes, such as the WLCSP, make it feasible to integrate these FPGAs directly onto small-scale devices. The incorporation of hardened MIPI and DDR controllers further streamlines the ability to handle video and data-heavy tasks, which is increasingly relevant in today's data-centric tech landscape.\n\nWith robust I/O features, these FPGAs provide versatile connection options for a range of peripherals, fulfilling the demands of industries that rely on high-speed and reliable data transfer. The support for a variety of standards combined with their easy-to-use development environment fosters a more straightforward transition from design to deployment. Efinix ensures these products are capable of handling future advancements by committing to a longer product lifecycle, promising designers a secure investment.
Join the world's most advanced semiconductor IP marketplace!
It's free, and you'll get all the tools you need to discover IP, meet vendors and manage your IP workflow!