Multi-Layer Processing Boosts Inference Throughput/Watt


The focus in discussion of inference throughput is often on the computations required. For example, YOLOv3, a power real time object detection and recognition model, requires 227 BILLION MACs (multiply-accumulates) to process a single 2 Mega Pixel image! This is with the Winograd Transformation; it’s more than 300 Billion without it. And there is a lot of discussion of the large size ... » read more

GDDR6 – HBM2 Tradeoffs


Steven Woo, Rambus fellow and distinguished inventor, talks about why designers choose one memory type over another. Applications for each were clearly delineated in the past, but the lines are starting to blur. Nevertheless, tradeoffs remain around complexity, cost, performance, and power efficiency.   Related Video Latency Under Load: HBM2 vs. GDDR6 Why data traffic and bandw... » read more

Designing Networking Chips


Susheel Tadikonda, vice president of networking and storage at Synopsys, talks about what’s changed in the way networking chips are being designed to deal with a massive increase in data. One of those shifts involves software-defined networking, where the greatest complexity resides in the software. That also has a big impact on the entire design flow, from pre-silicon to post-silicon. htt... » read more

Edge Inferencing Challenges


Geoff Tate, CEO of Flex Logix, talks about balancing different variables to improve performance and reduce power at the lowest cost possible in order to do inferencing in edge devices. https://youtu.be/1BTxwew--5U » read more

Lies, Damn Lies, And TOPS/Watt


There are almost a dozen vendors promoting inferencing IP, but none of them gives even a ResNet-50 benchmark. The only information they state typically is TOPS (Tera-Operations/Second) and TOPS/Watt. These two indicators of performance and power efficiency are almost useless by themselves. So what, exactly, does X TOPS really tell you about performance for your application? When a vendor ... » read more

Impacts Of Reliability On Power And Performance


Making sure a complex system performs as planned, and providing proper access to memories, requires a series of delicate tradeoffs that often were ignored in the past. But with performance improvements increasingly tied to architectures and microarchitectures, rather than just scaling to the next node, approaches such as determinism and different kinds of caching increasingly are becoming criti... » read more

Making Sense Of DRAM


Graham Allan, senior manager for product marketing at Synopsys, examines the different types of DRAM, from GDDR to HBM, which markets they’re used in, and why there is such disparity between them. https://youtu.be/ynvcPfD2cZU     __________________________________ See more tech talk videos here. » read more

Fusion Compiler: Comprehensive RTL-to-GDSII Implementation System


The semiconductor industry is going through a renaissance period with waves of technological advancements and innovation. There has been a significant uptick in demand for silicon in recent years, driven by market sectors including automotive, artificial intelligence, cloud computing, and internet of things (IoT) that have their own unique mix of design and implementation requirements. The mobi... » read more

AI Training Chips


Kurt Shuler, vice president of marketing at Arteris IP, talks with Semiconductor Engineering about how to architect an AI training chip, how different processing elements are used to accelerate training algorithms, and how to achieve improved performance. https://youtu.be/4cnBCX-9jlk     See other tech talk videos here. » read more

Power Issues Grow For Cloud Chips


Performance levels in traditional or hyperscale data centers are being limited by power and heat caused by an increasing number of processors, memory, disk and operating systems within servers. The problem is so complex and intertwined, though, that solving it requires a series of steps that hopefully add up to a significant reduction across a system. But at 7nm and below, predicting exactly... » read more

← Older posts