A Predictive, Per-Engine Energy Roofline of the Apple Neural Engine

The Apple Neural Engine (ANE) is in over a billion devices, but is characterized indirectly via CoreML. Using a CoreML-free fused-graph dispatch path, we measure the ANE, along with the Apple GPU and CPU on the same device, with per-engine energy accounting. The consumer-available roofline does not predict the ANE in practice. A critical correction is that the ANE sees two effective memory bandwidths: a slow activation path and a fast weight-stream path. Two roofline ridges follow, and the weight-stream ridge is below that of the GPU. The ANE single-operand compute peak decreases instead of plateauing. Its weight stream is byte-bound in fp16 but limited by the dequantizer’s element rate in int8. As a result, int8 compression saves energy at unchanged throughput and an int4 lookup-table stream raises throughput and efficiency at a cost of 2.55 points of ImageNet top-1. The corrected roofline identifies the lower-energy ANE paths and bounds it from below. A constructed device map reproduces across four unique ANE generations. Results show that operators should be placed by arithmetic intensity and fused for optimal performance.

Authors

Institutions

Publication Details

Journal
ACM Transactions on Architecture and Code Optimization
Published
2026-11-19
DOI
https://doi.org/10.1145/3864456
Primary Topic
Parallel Computing and Optimization Techniques
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

A Predictive, Per-Engine Energy Roofline of the Apple Neural Engine

Spencer H. Bryngelson
3 links
ACM Transactions on Architecture and Code Optimization
Parallel Computing and Optimization Techniques
article

A Predictive, Per-Engine Energy Roofline of the Apple Neural Engine

Spencer H. Bryngelson
article en

Abstract

The Apple Neural Engine (ANE) is in over a billion devices, but is characterized indirectly via CoreML. Using a CoreML-free fused-graph dispatch path, we measure the ANE, along with the Apple GPU and CPU on the same device, with per-engine energy accounting. The consumer-available roofline does not predict the ANE in practice. A critical correction is that the ANE sees two effective memory bandwidths: a slow activation path and a fast weight-stream path. Two roofline ridges follow, and the weight-stream ridge is below that of the GPU. The ANE single-operand compute peak decreases instead of plateauing. Its weight stream is byte-bound in fp16 but limited by the dequantizer’s element rate in int8. As a result, int8 compression saves energy at unchanged throughput and an int4 lookup-table stream raises throughput and efficiency at a cost of 2.55 points of ImageNet top-1. The corrected roofline identifies the lower-energy ANE paths and bounds it from below. A constructed device map reproduces across four unique ANE generations. Results show that operators should be placed by arithmetic intensity and fused for optimal performance.

ACM Transactions on Architecture and Code Optimization
The University of Texas at Austin (US)
Affordable and clean energy
Parallel Computing and Optimization Techniques
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon