Is this your business? Claim it to manage your IP and profile
Avispado is a sophisticated 64-bit RISC-V core that emphasizes efficiency and adaptability within in-order execution frameworks. It's engineered to cater to energy-efficient SoC designs, making it an excellent choice for machine learning applications with its compact design and ability to seamlessly communicate with RISC-V Vector Units. By utilizing the Gazzillion Misses™ technology, the Avispado core effectively handles high sparsity in tensor weights, resulting in superior energy efficiency per operation. This core features a 2-wide in-order configuration and supports the RISC-V Vector Specification 1.0 as well as Semidynamics' Open Vector Interface. With support for large memory capacities, it includes complete MMU features and is Linux-ready, ensuring it's prepared for demanding computational tasks. The core's native CHI interface can be fine-tuned to AXI, promoting cache-coherent multiprocessing capabilities. Avispado is optimized for various demanding workloads, with optional extensions for specific needs such as bit manipulation and cryptography. The core's customizable configuration allows changes to its instruction and data cache sizes (I$ and D$ from 8KB to 32KB), ensuring it meets specific application demands while retaining operational efficiency.
The Atrevido is a robust 64-bit RISC-V core known for its customizability and performance in out-of-order execution environments. Aimed at complex machine learning and high-performance computing tasks, this core can be configured to support a 2/3/4-wide out-of-order pipeline, complemented by Semidynamics' innovative Gazzillion Misses™ technology. This architecture manages heavily sparse data efficiently, essential for applications like recommendation systems and HPC workloads. The core also supports large memory capacities with features like register renaming, SV48 MMU, and native multiprocessing capabilities, making it Linux-ready. Additionally, Atrevido is vector-ready, supporting both the RISC-V Vector Specification 1.0 and the Semidynamics Open Vector Interface. This adaptability allows for seamless integration of custom vector units or existing Semidynamics offerings, optimizing energy consumption through its dense vector instruction set. The core's native interface, adaptable from CHI to AXI, facilitates coherent multiprocessing environments, scaling from small to large setups, and supporting cache-coherent multiprocessing environments. Atrevido offers extensive customization options including branch predictors and cache sizes (I$ and D$ from 8KB to 32KB), allowing users to tailor performance to specific requirements. It is designed to capitalize on high-bandwidth memory systems and long-memory latencies, commonly encountered in advanced machine learning applications, making it a comprehensive solution for edge and datacenter deployments.
The Vector Unit offered by Semidynamics is a specialized core extension designed to enhance the computational power of RISC-V processors through vector processing capabilities. This unit is tailored to accelerate workloads that demand high data throughput and parallel processing, such as those found in machine learning and scientific computing applications. By adopting vector instructions, the Vector Unit minimizes operational energy usage, efficiently handling bulk computations by encapsulating multiple operations in single instructions. Incorporating the RISC-V Vector Specification 1.0, Semidynamics' Vector Unit provides users with flexibility in choosing either custom solutions or Semidynamics' own offerings. This versatility empowers designers to balance between performance requirements and design constraints to fit various architectural needs within different SoCs. The unit supports dense computations using both gather and scatter vector operations, crucial for applications requiring rapid data manipulation and retrieval from scattered memory addresses. This approach greatly enhances the ability to manage tensor computations efficiently, making the Vector Unit an essential component for modern computing systems focused on high efficiency and performance.
Gazzillion Misses™ is an innovative technology designed to enhance data throughput for applications that involve extensive memory accesses and sparse data sets, making it particularly effective for use in data centers and HPC environments. The technology is integrated within Semidynamics' processor cores to support numerous simultaneous memory transactions, preserving high data bandwidth with minimal core requirements. It substantially reduces the silicon area needed for specific compute tasks by balancing memory throughput with core count. This IP is valuable in advanced machine learning workloads and recommendation systems, which require the processing of large amounts of sparse data with efficiency. It is geared towards optimizing memory transactions, allowing for efficient in-memory caching and key-value processing. These capabilities make Gazzillion Misses™ a vital component in building scalable and efficient SoCs for high-performance data processing tasks. Gazzillion Misses™ further stands out by enabling systems to sustain full memory bandwidth utilization, which is critical for meeting the demands of modern computing applications. Its ability to handle hundreds of misses per RISC-V core enhances the performance of applications that deal with multi-level parallelism and sparse matrix computations, key in areas like scientific computing and large-scale data analysis.
The Tensor Unit developed by Semidynamics is aimed at optimizing the processing of tensor-based computations typically required in artificial intelligence and machine learning tasks. This unit specializes in executing operations like tensor contractions and matrix manipulations with high efficiency and speed, addressing the demands of data-intensive applications which feature significant parallel processing. Engineered to support a variety of tensor operation instructions, the Tensor Unit delivers advanced capabilities in managing computation-heavy tasks such as deep learning and neural network applications. By integrating seamlessly with RISC-V cores, it broadens the scope of computational workloads that can be addressed using RISC-V hardware solutions, ensuring compatibility with current technological advancements in data processing. The Tensor Unit offers customizability in its operational design, allowing developers to tailor functions to specific needs within different computing environments. It promotes reduced energy consumption while maintaining high computational output, a necessary balance for achieving optimal performance in both edge and data center settings.
Join the world's most advanced semiconductor IP marketplace!
It's free, and you'll get all the tools you need to evaluate IP, download trial versions and datasheets, and manage your evaluation workflow!
To evaluate IP you need to be logged into a buyer profile. Select a profile below, or create a new buyer profile for your company.
