Kernel Fusion: An Effective Method for Better Power Efficiency on Multithreaded GPU

By Guibin Wang, Yisong Lin, Wei Yi. 2010

164 citations recorded by the source corpus.

Subfields: Computer Networks and Communications, Hardware and Architecture, ML Systems & Hardware

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Networks and Communications

1. Chang Zhou

In Hardware and Architecture

1. Chang Zhou
2. Yuxiong He
3. Zi-Mei Yao

Citations in the corpus

• ZeroQuant: Efficient and Affordable Post-Training Quantization for Large-Scale Transformers (2022, methodology): “To eliminate these cost, we use kernel fusion [67] to fuse quantization operation for activation with its previous element-wise and/or reduction operations such as bias-add, GeLU, and LayerNorm into a single operator, as illustrated by the green box in Figure 2.”
• Deep Interest Evolution Network for Click-Through Rate Prediction (2018, methodology): “In order to keep low latency and high throughput, we deploy several important techniques to improve serving performance: i) element parallel GRU & kernel fusion (Wang, Lin, and Yi 2010): we fuse as many independent kernels as possible.”