ResNet-50 Does Not Predict Inference Throughput For MegaPixel Neural Network Models

By Geoff Tate - 05 Nov, 2020 - Comments: 1

Customers are considering applications for AI inference and want to evaluate multiple inference accelerators. As we discussed last month, TOPS do NOT correlate with inference throughput and you should use real neural network models to benchmark accelerators. So is ResNet-50 a good benchmark for evaluating relative performance of inference accelerators? If your application is going to p... » read more

Memory Subsystems In Edge Inferencing Chips

By Ed Sperling - 23 Sep, 2019 - Comments: 0

Geoff Tate, CEO of Flex Logix, talks about key issues in a memory subsystem in an inferencing chip, how factors like heat can affect performance, and where these kinds of chips will be used. » read more

AI Inference Memory System Tradeoffs

By Geoff Tate - 01 Aug, 2019 - Comments: 1

When companies describe their AI inference chip they typically give TOPS but don’t talk about their memory system, which is equally important. What is TOPS? It means Trillions or Tera Operations per Second. It is primarily a measure of the maximum achievable throughput but not a measure of actual throughput. Most operations are MACs (multiply/accumulates), so TOPS = (number of MAC units) x... » read more

Do Large Batches Always Improve Neural Network Throughput?

By Geoff Tate - 06 Jun, 2019 - Comments: 1

Common benchmarks like ResNet-50 generally have much higher throughput with large batch sizes than with batch size =1. For example, the Nvidia Tesla T4 has 4x the throughput at batch=32 than when it is processing in batch=1 mode. Of course, larger batch sizes have a tradeoff: latency increases which may be undesirable in real-time applications. Why do larger batches increase throughput... » read more

Inference Acceleration: Follow The Memory

By Geoff Tate - 07 Mar, 2019 - Comments: 0

Much has been written about the computational complexity of inference acceleration: very large matrix multiplies for fully-connected layers and huge numbers of 3x3 convolutions across megapixel images, both of which require many thousands of MACs (multiplier-accumulators) to achieve high throughput for models like ResNet-50 and YOLOv3. The other side of the coin is managing the movement of d... » read more

Knowledge Centers
Entities, people and technologies explored

Shift Left Is The Tip Of The Iceberg

A transformative change is underway for semiconductor design and EDA. New languages, models, and abstractions will need to be created.

by Brian Bailey

Partitioning In The Chiplet Era

Understanding how chiplets interact under different workloads is critical to ensuring signal integrity and optimal performance in heterogeneous designs.

by Ann Mutschler

3.5D: The Great Compromise

Pros and cons of a middle-ground chiplet assembly that combines 2.5D and 3D-IC.

by Ed Sperling

AI’s Role In Chip Design Widens, Drawing In New Startups

Focus is on letting engineers do much more with the same or fewer resources — and less drudgery.

by Karen Heyman

Key Technologies To Extend EUV To 14 Angstroms

Alongside high-NA EUV will be better-performing photoresists, reduced roughness using passivation and etch, and lateral etching to reduce tip-to-tip dimensions.

by Laura Peters

What Comes After HBM For Chiplets

The standard for high-bandwidth memory limits design freedom at many levels, but that is required for interoperability. What freedoms can be taken from other functions to make chiplets possible?

by Brian Bailey

Memory Fundamentals For Engineers

eBook: Nearly everything you need to know about memory, including detailed explanations of the different types of memory; how and where these are used today; what's changing, which memories are successful and which ones might be in the future; and the limitations of each memory type.

by The SE Staff

Why Small Fab And Assembly Houses Are Thriving

Megafabs churning out the most advanced processors are not the only game in town.

by Bryon Moyer

tag: ResNet-50

ResNet-50 Does Not Predict Inference Throughput For MegaPixel Neural Network Models

Memory Subsystems In Edge Inferencing Chips

AI Inference Memory System Tradeoffs

Do Large Batches Always Improve Neural Network Throughput?

Inference Acceleration: Follow The Memory