中文 English

Amdahl Limits On AI


Software and hardware both place limits on how fast an application can run, but finding and eliminating the limitations is becoming more important in this age of multicore heterogeneous processing. The problem is certainly not new. Gene Amdahl (1922-2015) recognized the issue and published a paper about it in 1967. It provided the theoretical speedup for a defined task that could be expected... » read more

1.6 Tb/s Ethernet Challenges


Moving data at blazing fast speeds sounds good in theory, but it raises a number of design challenges. John Swanson, senior product marketing manager for high-performance computing digital IP at Synopsys, talks about the impact of next-generation Ethernet on switches, the types of data that need to be considered, the causes of data growth, and the size and structure of data centers, both in the... » read more

Difficult Memory Choices In AI Systems


The number of memory choices and architectures is exploding, driven by the rapid evolution in AI and machine learning chips being designed for a wide range of very different end markets and systems. Models for some of these systems can range in size from 10 billion to 100 billion parameters, and they can vary greatly from one chip or application to the next. Neural network training and infer... » read more

Compute-In Memory Accelerators Up-End Network Design Tradeoffs


An explosion in the amount of data, coupled with the negative impact on performance and power for moving that data, is rekindling interest around in-memory processing as an alternative to moving data back and forth between the memory and the processor. Compute-in-memory (CIM) arrays based on either conventional memory elements like DRAM and NAND flash, as well as emerging non-volatile memori... » read more

The Challenges Of Building Inferencing Chips


Putting a trained algorithm to work in the field is creating a frenzy of activity across the chip world, spurring designs that range from purpose-built specialty processors and accelerators to more generalized extensions of existing and silicon-proven technologies. What's clear so far is that no single chip architecture has been deemed the go-to solution for inferencing. Machine learning is ... » read more

The Cost Of Programmability


Nothing comes for free, and that is certainly true for the programmable elements in an SoC. But without them we are left with very specific devices that can only be used for one fixed application and cannot be updated. Few complex devices are created that do not have many layers of programmability, but the sizing of those capabilities is becoming more important than in the past. There are... » read more

Software In Inference Accelerators


Geoff Tate, CEO of Flex Logix, talks about the importance of hardware-software co-design for inference accelerators, how that affects performance and power, and what new approaches chipmakers are taking to bring AI chips to market. » read more

Making Sense Of ML Metrics


Steve Roddy, vice president of products for Arm’s Machine Learning Group, talks with Semiconductor Engineering about what different metrics actually mean, and why they can vary by individual applications and use cases. » read more

Using Multiple Inferencing Chips In Neural Networks


Geoff Tate, CEO of Flex Logix, talks about what happens when you add multiple chips in a neural network, what a neural network model looks like, and what happens when it’s designed correctly vs. incorrectly. » read more

Inferencing Efficiency


Geoff Tate, CEO of Flex Logix, talks with Semiconductor Engineering about how to measure efficiency in inferencing chips, how to achieve the most throughput for the lowest cost, and what the benchmarks really show. » read more

← Older posts