Is this your business? Claim it to manage your IP and profile
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 Atrevido core is a highly customizable 64-bit out-of-order RISC-V processor. It is adaptable for multiprocessor environments, supporting 2/3/4-wide configurations. Designed for intensive machine learning and HPC applications, it integrates seamlessly with vector units, capitalizing on its inherent capability to manage sparse datasets and reduce memory latency through Semidynamics' proprietary Gazzillion Misses™ technology. The core supports high-bandwidth memory systems, making it ideal for environments requiring efficient data handling and processing. Atrevido's architecture includes a range of sophisticated features such as direct support for unaligned accesses and comprehensive register renaming capabilities. Further, its compatibility with standard Linux environments and support for RISC-V vector specifications underscore its versatility. Applications in AI use its ability to decode multiple instructions per cycle, supporting both complex matrix operations and real-time processing needs. The core's capability to integrate with Semidynamics' Vector Unit offers additional computing power, enhancing its performance in tasks requiring parallel data processing. This combination makes it suitable for dense computational tasks, such as those found in recommendation systems and key-value stores, providing a robust platform that scales efficiently across various industry needs.
Semidynamics' Vector Unit is a fully customizable RISC-V processor designed to maximize data processing capabilities through parallel computing. Supporting up to 2048 bits, this Vector Unit is engineered to handle a mix of data types and sizes, from FP64 to INT8, providing substantial flexibility for diverse application needs. This unit stands out due to its customizable data path length (DLEN) and vector register length (VLEN), which can be adjusted according to the specific performance and power trade-offs required by applications. Its intricate architecture supports all RISC-V Vector Interface specifications, enabling broad compatibility and integration with existing systems. Optimized for high-performance applications such as AI and HPC, the Vector Unit's architecture efficiently manages vector arithmetic operations, significantly improving processing speed for data-intensive tasks. As an advanced feature, it supports simultaneous operations across multiple vector cores, leveraging Semidynamics' Gazzillion Misses™ technology to maintain high bandwidth and low latency in all operations.
The Tensor Unit from Semidynamics is a revolutionary component in their AI and machine learning portfolio, designed as the world's first fully-coherent RISC-V tensor unit. Optimized for ultra-fast AI computations, this unit integrates seamlessly with the company's Vector Unit, facilitating matrix multiplications which are essential in deep learning models. Its architecture ensures low power consumption while delivering high performance for AI tasks, particularly those involving complex matrix operations like fully connected and convolutional layers. The Tensor Unit takes advantage of the Gazzillion Misses™ technology to efficiently manage data flow, reducing latency and increasing computational throughput. The Tensor Unit's design accommodates universal RISC-V compatibility, eliminating the need for DMAs and ensuring direct integration within existing systems. By focusing on efficient hardware utilization, it supports AI operations across a wide range of applications, from edge devices to large data centers, making it an integral component in modern AI systems.
Gazzillion Misses™ is a unique technology developed by Semidynamics designed to overcome latency issues in memory-intensive applications. Embedded within the company's RISC-V processor cores, this technology allows multiple outstanding memory requests simultaneously, bypassing traditional pipeline stalls caused by cache misses. This enables processors to continue executing instructions rather than waiting for data retrieval, significantly reducing idle cycles. By handling up to 128 simultaneous requests, Gazzillion Misses™ significantly boosts system performance and is highly beneficial in big data, HPC, and AI environments where bandwidth is critical. This technology effectively increases memory throughput, making it ideal for applications with extensive data stream requirements, such as machine learning models and complex algorithms in data centers. Moreover, its implementation allows for more straightforward software development processes, easing complexity by providing an efficient framework to sustain high memory bandwidth. It ensures that the processor's compute capabilities are fully utilized, enhancing overall system response and efficiency. This makes Gazzillion Misses™ a keystone in modern architecture for high-performance computing solutions.
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!
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.
