Sitemap
A list of all the posts and pages found on the site. For you robots out there, there is an XML version available for digesting as well.
Pages
Posts
portfolio
publications
Habitat: A Runtime-Based Computational Performance Predictor for Deep Neural Network Training
Published in USENIX Annual Technical Conference (ATC), 2021
A runtime-based approach to predict computational performance for DNN training across different GPU hardware.
Recommended citation: Geoffrey X. Yu, Yubo Gao, Pavel Golikov, Gennady Pekhimenko. (2021). "Habitat: A Runtime-Based Computational Performance Predictor for Deep Neural Network Training." USENIX ATC 2021. pp. 503-521.
Download Paper
GPUPool: A Holistic Approach to Fine-Grained GPU Sharing in the Cloud
Published in International Conference on Parallel Architectures and Compilation Techniques (PACT), 2022
A holistic approach to fine-grained GPU sharing that improves GPU utilization in cloud computing environments.
Recommended citation: Xiaodan Serina Tan, Pavel Golikov, Nandita Vijaykumar, Gennady Pekhimenko. (2022). "GPUPool: A Holistic Approach to Fine-Grained GPU Sharing in the Cloud." PACT 2022. pp. 317-332.
Download Paper
Fusing Adds and Shifts for Efficient Dot Products
Published in IEEE Computer Architecture Letters, 2026
FASED: A method for fusing addition and shift operations to compute dot products more efficiently in hardware.
Recommended citation: Pavel Golikov, Karthik Ganesan, Gennady Pekhimenko, Mark C. Jeffrey. (2026). "Fusing Adds and Shifts for Efficient Dot Products." IEEE Computer Architecture Letters. 25(1), pp. 33-36.
Download Paper