Tempus Core: Area-Power Efficient Temporal-Unary Convolution Core for Low-Precision Edge DLAs

2024-12-25

AI Summary

To address stringent resource and power constraints for edge AI inference, this paper proposes Tempus Core—a temporal-unary convolutional processor compatible with the NVDLA dataflow. Methodologically, it introduces the first temporal-unary-binary (TUB) multiplier architecture, enabling seamless hardware integration of unary computing with commercial deep learning accelerators at both dataflow and interface levels. Fabricated in 45 nm CMOS, Tempus Core employs temporal-encoded unary computation, NVDLA-compatible PE array reconfiguration, and backend optimizations. Results show that a single PE reduces area and power by 59.3% and 15.3%, respectively, versus NVDLA’s CMAC unit; a 16×16 array achieves 75% area and 62% power reduction. Throughput per unit area improves by 5× for INT8 and 4× for INT4. A compact 16×4 array operating in INT4 mode occupies only 0.017 mm² and consumes 6.2 mW.

Technology Category

• Machine Learning: Hardware-aware ML
• Planning, Routing, and Scheduling: Temporal Planning
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Graph Algorithms and Modeling for the Web: Algorithms and analysis for heterogeneous, signed, attributed, multi-relational, temporal, higher-order, and annotated Web-related graphs
• Search and Retrieval-Augmented AI: Vertical and domain-specific search

Abstract

The increasing complexity of deep neural networks (DNNs) poses significant challenges for edge inference deployment due to resource and power constraints of edge devices. Recent works on unary-based matrix multiplication hardware aim to leverage data sparsity and low-precision values to enhance hardware efficiency. However, the adoption and integration of such unary hardware into commercial deep learning accelerators (DLA) remain limited due to processing element (PE) array dataflow differences. This work presents Tempus Core, a convolution core with highly scalable unary-based PE array comprising of tub (temporal-unary-binary) multipliers that seamlessly integrates with the NVDLA (NVIDIA's open-source DLA for accelerating CNNs) while maintaining dataflow compliance and boosting hardware efficiency. Analysis across various datapath granularities shows that for INT8 precision in 45nm CMOS, Tempus Core's PE cell unit (PCU) yields 59.3% and 15.3% reductions in area and power consumption, respectively, over NVDLA's CMAC unit. Considering a 16x16 PE array in Tempus Core, area and power improves by 75% and 62%, respectively, while delivering 5x and 4x iso-area throughput improvements for INT8 and INT4 precisions. Post-place and route analysis of Tempus Core's PCU shows that the 16x4 PE array for INT4 precision in 45nm CMOS requires only 0.017 mm^2 die area and consumes only 6.2mW of total power. We demonstrate that area-power efficient unary-based hardware can be seamlessly integrated into conventional DLAs, paving the path for efficient unary hardware for edge AI inference.

Problem

• Deep Neural Networks
• Edge Devices
• Efficient Convolution Processing

Innovation

• Tempus Core
• low-precision edge devices
• NVDLA collaboration