A Dynamic Allocation Scheme for Adaptive Shared-Memory Mapping on Kilo-Core RV Clusters for Attention-Based Model Deployment


METADATA ONLY
Loading...

Date

2025

Publication Type

Conference Paper

ETH Bibliography

yes

Citations

Scopus:
Altmetric
METADATA ONLY

Data

Rights / License

Abstract

Attention-based models demand flexible hardware to manage diverse kernels with varying arithmetic intensities and memory access patterns. Large clusters with shared L1 memory, a common architectural pattern, struggle to fully utilize their processing elements (PEs) when scaled up due to reduced throughput in the hierarchical PE-to-L1 intra-cluster interconnect. This paper presents Dynamic Allocation Scheme (DAS), a runtime programmable address remapping hardware unit coupled with a unified memory allocator, designed to minimize data access contention of PEs onto the multi-banked L1. We evaluated DAS on an aggressively scaled-up 1024-PE RISC-V cluster with Non-Uniform Memory Access (NUMA) PE-to-L1 interconnect to demonstrate its potential for improving data locality in large parallel machine learning workloads. For a Vision Transformer (ViT)-L/16 model, each encoder layer executes in 5.67 ms, achieving a 1.94× speedup over the fixed word-level interleaved baseline with 0.81 PE utilization. Implemented in 12nm FinFET technology, DAS incurs <0.1% area overhead.

Publication status

published

Editor

Book title

2025 IEEE 36th International Conference on Application-specific Systems, Architectures and Processors (ASAP)

Journal / series

Volume

Pages / Article No.

9 - 16

Publisher

IEEE

Event

36th International Conference on Application-specific Systems, Architectures and Processors (ASAP 2025)

Edition / version

Methods

Software

Geographic location

Date collected

Date created

Subject

RISC-V; Manycore; Transformers

Organisational unit

03996 - Benini, Luca / Benini, Luca check_circle

Notes

Funding

Related publications and datasets