Convolutional Neural Network With INT4 Optimization


Xilinx provides an INT8 AI inference accelerator on Xilinx hardware platforms — Deep Learning Processor Unit (XDPU). However, in some resource-limited, high-performance and low-latency scenarios (such as the resource-power-sensitive edge side and low-latency ADAS scenario), low bit quantization of neural networks is required to achieve lower power consumption and higher performance than provi... » read more

Versal: The First Adaptive Compute Acceleration Platform (ACAP)


Recent technical challenges have forced the industry to explore options beyond the conventional “one size fits all” CPU scalar processing solution. Very large vector processing (DSP, GPU) solves some problems, but it runs into traditional scaling challenges due to inflexible, inefficient memory bandwidth usage. Traditional FPGA solutions provide programmable memory hierarchy, but the tradit... » read more

The Best Platform When It Comes To PTP Accuracy


Synchronization is pervasive in most industry sectors, including finance, telecom, industrial, consumer, and aerospace and defense. All of these markets have several applications that rely on synchronization. The first part of this white paper describes a few typical examples where synchronization enables applications that would not be possible otherwise. Synchronization can be achieved with... » read more

Xilinx Reduces Risk And Increases Efficiency For IEC61508 And ISO26262 Certified Safety Applications


This white paper introduces key dependability aspects for industrial and automotive customers who are designing and developing programmable electronic equipment for safety applications using Xilinx FPGA and SoC devices. The main focus of this white paper is to explain how to create solutions with highly integrated, high-performance certifiable systems that target IEC 61508 / ISO 26262 norms. Th... » read more

Computational Software


Electronics technology is evolving rapidly, becoming pervasive in our lives. There are more smart phones in use than there are people on earth, driver assistance is now common in automobiles, commercial airplanes have increasingly sophisticated infotainment, and wearable health monitors exist for a plethora of missions. Every device is generating and communicating massive amounts of data, inclu... » read more

The Growing Market For Specialized Artificial Intelligence IP In SoCs


Over the past decade, designers have developed silicon technologies that run advanced deep learning mathematics fast enough to explore and implement artificial intelligence (AI) applications such as object identification, voice and facial recognition, and more. Machine vision applications, which are now often more accurate than a human, are one of the key functions driving new system-on-chip (S... » read more

A Better Path From Simulink To RTL With Catapult HLS


Design teams working on ASIC or FPGA projects often start with algorithm exploration using MATLAB in order to prove out the mathematical behavior of the functional blocks at a high level of abstraction. MATLAB as a high-level programming language doesn’t support hardware architecture modeling, so many teams use the Simulink environment for performing model-based, multi-domain simulation of th... » read more

RISC-V Custom Instruction Flow Application Note


A RISC-V processor has several defined decode spaces, for example custom0, custom1 etc. into which new custom instructions can be added. OVP Fast processor models can be extended without modification to the pre-compiled and verified base processor model source code using one or more extension libraries. An extension library can be loaded as part of a virtual platform definition in addition to t... » read more

HDL Simulation Acceleration Solution For Microchip FPGA Designs


Mission-critical FPGA designs for space and radar applications continue to increase in complexity, such that they require a comprehensive and robust verification environment. There are hardware-in-the-loop solutions in the market that utilize FPGA boards, but when it comes to establishing functional coverage and debugging the custom logic, users would typically need to go back to HDL simulation... » read more

A Study Of Wiggling AA Modeling And Its Impact On Device Performance In Advanced DRAM


In this paper, a wiggling active area (fin) in an advanced 1x DRAM process was analyzed and modeled using the pattern-dependent etch simulation capabilities of the SEMulator3D semiconductor modeling software. Nonuniformity in sidewall passivation caused by hard mask pattern density loading was identified as the root cause of the wiggling profile. The calibrated model mimicked these phenomena, g... » read more

← Older posts Newer posts →