Alert button
Picture for Amar Phanishayee

Amar Phanishayee

Alert button

Packrat: Automatic Reconfiguration for Latency Minimization in CPU-based DNN Serving

Nov 30, 2023
Ankit Bhardwaj, Amar Phanishayee, Deepak Narayanan, Mihail Tarta, Ryan Stutsman

Viaarxiv icon

A Study on the Intersection of GPU Utilization and CNN Inference

Dec 15, 2022
Jack Kosaian, Amar Phanishayee

Figure 1 for A Study on the Intersection of GPU Utilization and CNN Inference
Figure 2 for A Study on the Intersection of GPU Utilization and CNN Inference
Figure 3 for A Study on the Intersection of GPU Utilization and CNN Inference
Figure 4 for A Study on the Intersection of GPU Utilization and CNN Inference
Viaarxiv icon

Harmony: Overcoming the hurdles of GPU memory capacity to train massive DNN models on commodity servers

Feb 02, 2022
Youjie Li, Amar Phanishayee, Derek Murray, Jakub Tarnawski, Nam Sung Kim

Figure 1 for Harmony: Overcoming the hurdles of GPU memory capacity to train massive DNN models on commodity servers
Figure 2 for Harmony: Overcoming the hurdles of GPU memory capacity to train massive DNN models on commodity servers
Figure 3 for Harmony: Overcoming the hurdles of GPU memory capacity to train massive DNN models on commodity servers
Figure 4 for Harmony: Overcoming the hurdles of GPU memory capacity to train massive DNN models on commodity servers
Viaarxiv icon

Synergy: Resource Sensitive DNN Scheduling in Multi-Tenant Clusters

Oct 12, 2021
Jayashree Mohan, Amar Phanishayee, Janardhan Kulkarni, Vijay Chidambaram

Figure 1 for Synergy: Resource Sensitive DNN Scheduling in Multi-Tenant Clusters
Figure 2 for Synergy: Resource Sensitive DNN Scheduling in Multi-Tenant Clusters
Figure 3 for Synergy: Resource Sensitive DNN Scheduling in Multi-Tenant Clusters
Figure 4 for Synergy: Resource Sensitive DNN Scheduling in Multi-Tenant Clusters
Viaarxiv icon

Efficient Large-Scale Language Model Training on GPU Clusters

Apr 09, 2021
Deepak Narayanan, Mohammad Shoeybi, Jared Casper, Patrick LeGresley, Mostofa Patwary, Vijay Korthikanti, Dmitri Vainbrand, Prethvi Kashinkunti, Julie Bernauer, Bryan Catanzaro, Amar Phanishayee, Matei Zaharia

Figure 1 for Efficient Large-Scale Language Model Training on GPU Clusters
Figure 2 for Efficient Large-Scale Language Model Training on GPU Clusters
Figure 3 for Efficient Large-Scale Language Model Training on GPU Clusters
Figure 4 for Efficient Large-Scale Language Model Training on GPU Clusters
Viaarxiv icon

Analyzing and Mitigating Data Stalls in DNN Training

Jul 14, 2020
Jayashree Mohan, Amar Phanishayee, Ashish Raniwala, Vijay Chidambaram

Figure 1 for Analyzing and Mitigating Data Stalls in DNN Training
Figure 2 for Analyzing and Mitigating Data Stalls in DNN Training
Figure 3 for Analyzing and Mitigating Data Stalls in DNN Training
Figure 4 for Analyzing and Mitigating Data Stalls in DNN Training
Viaarxiv icon

Efficient Algorithms for Device Placement of DNN Graph Operators

Jun 29, 2020
Jakub Tarnawski, Amar Phanishayee, Nikhil R. Devanur, Divya Mahajan, Fanny Nina Paravecino

Figure 1 for Efficient Algorithms for Device Placement of DNN Graph Operators
Figure 2 for Efficient Algorithms for Device Placement of DNN Graph Operators
Figure 3 for Efficient Algorithms for Device Placement of DNN Graph Operators
Figure 4 for Efficient Algorithms for Device Placement of DNN Graph Operators
Viaarxiv icon

Memory-Efficient Pipeline-Parallel DNN Training

Jun 16, 2020
Deepak Narayanan, Amar Phanishayee, Kaiyu Shi, Xie Chen, Matei Zaharia

Figure 1 for Memory-Efficient Pipeline-Parallel DNN Training
Figure 2 for Memory-Efficient Pipeline-Parallel DNN Training
Figure 3 for Memory-Efficient Pipeline-Parallel DNN Training
Figure 4 for Memory-Efficient Pipeline-Parallel DNN Training
Viaarxiv icon

Daydream: Accurately Estimating the Efficacy of Optimizations for DNN Training

Jun 05, 2020
Hongyu Zhu, Amar Phanishayee, Gennady Pekhimenko

Figure 1 for Daydream: Accurately Estimating the Efficacy of Optimizations for DNN Training
Figure 2 for Daydream: Accurately Estimating the Efficacy of Optimizations for DNN Training
Figure 3 for Daydream: Accurately Estimating the Efficacy of Optimizations for DNN Training
Figure 4 for Daydream: Accurately Estimating the Efficacy of Optimizations for DNN Training
Viaarxiv icon

Blink: Fast and Generic Collectives for Distributed ML

Oct 11, 2019
Guanhua Wang, Shivaram Venkataraman, Amar Phanishayee, Jorgen Thelin, Nikhil Devanur, Ion Stoica

Figure 1 for Blink: Fast and Generic Collectives for Distributed ML
Figure 2 for Blink: Fast and Generic Collectives for Distributed ML
Figure 3 for Blink: Fast and Generic Collectives for Distributed ML
Figure 4 for Blink: Fast and Generic Collectives for Distributed ML
Viaarxiv icon