Stas Bekman
Research Subjects
Stas Bekman has a total of 8 co-authors. Here are the top co-authors by number of shared publications:
- Aamir Shafi
The Ohio State University, Columbus, OH, USA
1 shared publication - Deepak Narayanan
NVIDIA, USA
1 shared publication - Dhabaleswar Kumar Panda
The Ohio State University, Columbus, USA
1 shared publication - Hari Subramoni
The Ohio State University, Columbus, USA
1 shared publication - Jacob Hatef
The Ohio State University, Columbus, OH, USA
1 shared publication - Junqi Yin
Oak Ridge National Laboratory, Oak Ridge, TN, USA
1 shared publication
Latest Publications
- Open Access
- 12 August 2024
- Published by Association for Computing Machinery (ACM)
- p. 84-96
The Case for Co-Designing Model Architectures with Hardware
While GPUs are responsible for training the vast majority of state-of-the-art deep learning models, the implications of their architecture are often overlooked when designing new deep learning (DL) models. As a consequence, modifying a DL model to be more amenable to the target hardware can significantly improve the runtime performance of DL training and inference. In this paper, we provide a set of guidelines for users to maximize the runtime performance of their transformer models. These guidelines have been created by carefully considering the impact of various model hyperparameters controlling model shape on the efficiency of the underlying computation kernels executed on the GPU. We find the throughput of models with “efficient” model shapes is up to 39% higher while preserving accuracy compared to models with a similar number of parameters but with unoptimized shapes.
- 12 August 2024
- Published by Association for Computing Machinery (ACM)
- p. 84-96
Open AccessThe Case for Co-Designing Model Architectures with Hardware
While GPUs are responsible for training the vast majority of state-of-the-art deep learning models, the implications of their architecture are often overlooked when designing new deep learning (DL) models. As a consequence, modifying a DL model to be more amenable to the target hardware can significantly improve the runtime performance of DL training and inference. In this paper, we provide a set of guidelines for users to maximize the runtime performance of their transformer models. These guidelines have been created by carefully considering the impact of various model hyperparameters controlling model shape on the efficiency of the underlying computation kernels executed on the GPU. We find the throughput of models with “efficient” model shapes is up to 39% higher while preserving accuracy compared to models with a similar number of parameters but with unoptimized shapes.