Data Movement Is the Energy Bottleneck of Today’s SoCs


In today’s AI-focused semiconductor landscape, raw compute performance alone no longer defines the effectiveness of a system-on-chip (SoC). The efficiency of data movement across the chip has become just as important. Whether designed for data centers or edge AI devices, SoCs must now prioritize data transport as a core architectural consideration. Moving data efficiently across the silicon f... » read more

What’s Changing In SerDes


SerDes is all about pushing data through the smallest number of physical channels. But when it comes to AI, more data needs to be moved, and it has to be moved more quickly. Todd Bermensolo, product marketing manager at Alphawave Semi, talks about the impact of faster data movement on the transmitter (more power) and on the receiver (gain and advanced equalization), how to ensure signal inte... » read more

Lines Blurring Between Supercomputing And HPC


Supercomputers and high-performance computers are becoming increasingly difficult to differentiate due to the proliferation of AI, which is driving huge performance increases in commercial and scientific applications and raising similar challenges for both. While the goals of supercomputing and high-performance computing (HPC) have always been similar — blazing fast processing — the mark... » read more

Scaling Performance In AI Systems


Improving performance in AI designs involves the usual tradeoffs in power and performance, but achieving a good balance is becoming much more challenging. There is more data to process, new heterogeneous architectures to contend with, and much higher utilization rates. Andy Nightingale, vice president of product management and marketing at Arteris, talks about where the bottlenecks are, how to ... » read more

Workload-Specific Data Movements Across AI Workloads in Multi-Chiplet AI Accelerators


A new technical paper titled "Communication Characterization of AI Workloads for Large-scale Multi-chiplet Accelerators" was published by researchers at Universitat Politecnica de Catalunya. Abstract "Next-generation artificial intelligence (AI) workloads are posing challenges of scalability and robustness in terms of execution time due to their intrinsic evolving data-intensive characteris... » read more

Higher Density, More Data Create New Bottlenecks In AI Chips


Data movement is becoming a bigger problem at advanced nodes and in advanced packaging due to denser circuitry, more physical effects that can affect the integrity of signals or the devices themselves, and a significant increase in data from AI and machine learning. Just shrinking features in a design is no longer sufficient, given the scaling mismatch between SRAM-based L1 cache and digital... » read more

The Journey To Exascale Computing And Beyond


High performance computing witnessed one of its most ambitious leaps forward with the development of the US supercomputer “Frontier.” As Scott Atchley from Oak Ridge National Laboratory discussed at Supercomputing 23 (SC23) in Denver last month, the Frontier had the ambitious goal of achieving performance levels 1000 times higher than the petascale systems that preceded it, while also stayi... » read more

Designing for Data Flow


Movement and management of data inside and outside of chips is becoming a central theme for a growing number of electronic systems, and a huge challenge for all of them. Entirely new architectures and techniques are being developed to reduce the movement of data and to accomplish more per compute cycle, and to speed the transfer of data between various components on a chip and between chips ... » read more

Dealing With Heat In Near-Memory Compute Architectures


The explosion in data forcing chipmakers to get much more granular about where logic and memory are placed on a die, how data is partitioned and prioritized to utilize those resources, and what the thermal impact will be if they are moved closer together on a die or in a package. For more than a decade, the industry has faced a basic problem — moving data can be more resource-intensive tha... » read more

Algorithm HW Framework That Minimizes Accuracy Degradation, Data Movement, And Energy Consumption Of DNN Accelerators (Georgia Tech)


This new research paper titled "An Algorithm-Hardware Co-design Framework to Overcome Imperfections of Mixed-signal DNN Accelerators" was published by researchers at Georgia Tech. According to the paper's abstract, "In recent years, processing in memory (PIM) based mixed-signal designs have been proposed as energy- and area-efficient solutions with ultra high throughput to accelerate DNN com... » read more

← Older posts