42 citations · 45 across the 8 of their papers we have counts for
9 papers
Give the Drafter a Second Chance to Guess: Parallel Reconsideration for Speculative Decoding
Hoang Anh Duy Le, Huibing Dong, Alireza Khadem +6
Speculative decoding accelerates large language model inference with a small drafter, whose per-guess acceptance rate determines the achievable speedup. Feature-level drafters shar…
Characterizing Adaptive Mesh Refinement on Heterogeneous Platforms with Parthenon-VIBE
Akash Poptani, Alireza Khadem, Scott Mahlke +3
Hero-class HPC simulations rely on Adaptive Mesh Refinement (AMR) to reduce compute and memory demands while maintaining accuracy. This work analyzes the performance of Parthenon,…
A Customized Memory-aware Architecture for Biological Sequence Alignment
Nasrin Akbari, Mehdi Modarressi, Alireza Khadem
Sequence alignment is a fundamental process in computational biology which identifies regions of similarity in biological sequences. With the exponential growth in the volume of da…
DX100: A Programmable Data Access Accelerator for Indirection
Alireza Khadem, Kamalavasan Kamalakkannan, Zhenyan Zhu +8
Indirect memory accesses frequently appear in applications where memory bandwidth is a critical bottleneck. Prior indirect memory access proposals, such as indirect prefetchers, ru…
PIM Is All You Need: A CXL-Enabled GPU-Free System for Large Language Model Inference
Yufeng Gu, Alireza Khadem, Sumanth Umesh +5
Large Language Model (LLM) inference uses an autoregressive manner to generate one token at a time, which exhibits notably lower operational intensity compared to earlier Machine L…
Multi-Dimensional Vector ISA Extension for Mobile In-Cache Computing
Alireza Khadem, Daichi Fujiki, Hilbert Chen +4
In-cache computing technology transforms existing caches into long-vector compute units and offers low-cost alternatives to building expensive vector engines for mobile CPUs. Unfor…
Vector-Processing for Mobile Devices: Benchmark and Analysis
Alireza Khadem, Daichi Fujiki, Nishil Talati +2
Vector processing has become commonplace in today's CPU microarchitectures. Vector instructions improve performance and energy which is crucial for resource-constraint mobile devic…
CoDR: Computation and Data Reuse Aware CNN Accelerator
Alireza Khadem, Haojie Ye, Trevor Mudge
Computation and Data Reuse is critical for the resource-limited Convolutional Neural Network (CNN) accelerators. This paper presents Universal Computation Reuse to exploit weight s…
Design Challenges of Neural Network Acceleration Using Stochastic Computing
Alireza Khadem
The enormous and ever-increasing complexity of state-of-the-art neural networks (NNs) has impeded the deployment of deep learning on resource-limited devices such as the Internet o…