AHASD: Asynchronous Heterogeneous Architecture for LLM Adaptive Drafting Speculative Decoding on Mobile Devices
Open Access
- 26 July 2026
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 1-7
Abstract
No abstract availableFunding Information
- National Key Research and Development Program of China (2023YFB4503500)
- CAS Project for Young Scientists in Basic Research (YSBR-029)
- National Key Research and Development Program of China (62502498)
- Beijing National Laboratory for Molecular Sciences (L234078)
This publication has 14 references indexed in Scilit:
- SpecASR: Accelerating LLM-based Automatic Speech Recognition via Speculative Decoding2025
- ONNXim: A Fast, Cycle-Level Multi-Core NPU SimulatorIEEE Computer Architecture Letters, 2024
- AttAcc! Unleashing the Power of PIM for Batched Transformer-based Generative Model Inference2024
- SpecPIM: Accelerating Speculative Inference on PIM-Enabled System via Architecture-Dataflow Co-Exploration2024
- SpecInfer: Accelerating Large Language Model Serving with Tree-based Speculative Inference and Verification2024
- System Architecture and Software Stack for GDDR6-AiM2022
- Aquabolt-XL HBM2-PIM, LPDDR5-PIM With In-Memory Processing, and AXDIMM With Acceleration BufferIEEE Micro, 2022
- Hardware Architecture and Software Stack for PIM Based on Commercial DRAM Technology : Industrial Product2021
- MAESTRO: A Data-Centric Approach to Understand Reuse, Performance, and Hardware Cost of DNN MappingsIEEE Micro, 2020
- CACTI 7ACM Transactions on Architecture and Code Optimization, 2017