DéjàVu: KV-cache Streaming for Fast, Fault-tolerant Generative LLM Serving
Foteini Strati, Sara Mcallister, Amar Phanishayee, Jakub Tarnawski, Ana Klimovic
ICML 2024 | July 2024
Publications
-
-
-
Integrated Hardware Architecture and Device Placement Search
Irene Wang, Jakub Tarnawski, Amar Phanishayee, Divya Mahajan
2024 International Conference on Machine Learning | July 2024
-
MGit: A Model Versioning and Management System
Wei Hao, Daniel Mendoza, Rafael da Silva, Deepak Narayanan, Amar Phanishayee, Asaf Cidon, Junfeng Yang
2024 International Conference on Machine Learning | July 2024
-
Workload-Aware Hardware Accelerator Mining for Distributed Deep Learning Training
Muhammad Adnan, Amar Phanishayee,, Janardhan (Jana) Kulkarni, Prashant J. Nair, Divya Mahajan
arXiv:2404.14632 | April 2024
Publié par Microsoft
-
-
-
Blox: A Modular Toolkit for Deep Learning Schedulers
Saurabh Agarwal, Amar Phanishayee, Shivaram Venkataraman
European Conference on Computer Systems (ACM EuroSys 2024 - spring accept). | December 2023
arXiv:2312.12621 | December 2023
-
Packrat: Automatic Reconfiguration for Latency Minimization in CPU-based DNN Serving
Ankit Bhardwaj, Amar Phanishayee, Deepak Narayanan, Mihail Tarta, Ryan Stutsman
arXiv:2311.18174 | November 2023
Publié par arXiv | Nov 2023
-
MGit: A Model Versioning and Management System
Wei Hao, Daniel Mendoza, Rafael da Silva, Deepak Narayanan, Amar Phanishayee
arXiv:2307.07507 | July 2023
Publié par arXiv
-
-
-
A Study on the Intersection of GPU Utilization and CNN Inference
Jack Kosaian, Amar Phanishayee
arXiv:2212.07936 | December 2022
Publié par arXiv
-
Harmony: Overcoming the hurdles of GPU memory capacity to train massive DNN models on commodity servers
Youjie Li, Amar Phanishayee, Derek Murray, Jakub Tarnawski, Nam Sung Kim
VLDB 2022 | September 2022
-
Looking Beyond GPUs for DNN Scheduling on Multi-Tenant Clusters
Jayashree Mohan, Amar Phanishayee, Janardhan (Jana) Kulkarni, Vijay Chidambaram
USENIX Symposium on Operating Systems Design and Implementation (OSDI 2022) | July 2022
-
-
-
Piper: Multidimensional Planner for DNN Parallelization
Jakub Tarnawski, Deepak Narayanan, Amar Phanishayee
NeurIPS 2021 | December 2021
-
Efficient Large-Scale Language Model Training on GPU Clusters Using Megatron-LM
Deepak Narayanan, Mohammad Shoeybi, Jared Casper, Patrick LeGresley, Mostofa Patwary, Vijay Anand Korthikanti, Dmitri Vainbrand, Prethvi Kashinkunti, Julie Bernauer, Bryan Catanzaro, Amar Phanishayee, Matei Zaharia
International Conference for High Performance Computing, Networking, Storage and Analysis (SC 2021) | November 2021
Best Student Paper
-
Synergy: Resource Sensitive DNN Scheduling in Multi-Tenant Clusters
Jayashree Mohan, Amar Phanishayee, Janardhan (Jana) Kulkarni, Vijay Chidambaram
October 2021
-
Boosting the Throughput and Accelerator Utilization of Specialized CNN Inference Beyond Increasing Batch Size
Jack Kosaian, Amar Phanishayee, Matthai Philipose, Debadeepta Dey, Rashmi Vinayak
2021 International Conference on Machine Learning (ICML 2021) | July 2021
-
Memory-Efficient Pipeline-Parallel DNN Training
Deepak Narayanan, Amar Phanishayee, Kaiyu Shi, Xie Chen, Matei Zaharia
International Conference on Machine Learning (ICML 2021) | July 2021
-
Doing more with less: Training large DNN models on commodity servers for the masses
Youjie Li, Amar Phanishayee, Derek Murray, Nam Sung Kim
HotOS Workshop | May 2021
-
CheckFreq: Frequent, Fine-Grained DNN Checkpointing
Jayashree Mohan, Amar Phanishayee, Vijay Chidambaram
USENIX FAST 2021 | February 2021
-
Analyzing and Mitigating Data Stalls in DNN Training
Jayashree Mohan, Amar Phanishayee, Ashish Raniwala, Vijay Chidambaram
VLDB 2021 | January 2021
-
-
-
Efficient Algorithms for Device Placement of DNN Graph Operators
Jakub Tarnawski, Amar Phanishayee, Nikhil R. Devanur, Divya Mahajan, Fanny Nina Paravecino
NeurIPS 2020 | December 2020
-
Heterogeneity-Aware Cluster Scheduling Policies for Deep Learning Workloads
Deepak Narayanan, Keshav Santhanam, Fiodar Kazhamiaka, Amar Phanishayee, Matei Zaharia
USENIX Symposium on Operating Systems Design and Implementation (OSDI 2020) | November 2020
-
Daydream: Accurately Estimating the Efficacy of Performance Optimizations for DNN Training
Hongyu Zhu, Amar Phanishayee, Gennady Pekhimenko
USENIX ATC 2020 | July 2020
-
The Non-IID Data Quagmire of Decentralized Machine Learning
Kevin Hsieh, Amar Phanishayee, Onur Mutlu, Phillip Gibbons
International Conference on Machine Learning (ICML 2020) | July 2020
-
Memory-Efficient Pipeline-Parallel DNN Training
Deepak Narayanan, Amar Phanishayee, Kaiyu Shi, Xie Chen, Matei Zaharia
arXiv:2006.09503 | June 2020
Publié par arXiv
-
Blink: Fast and Generic Collectives for Distributed ML
Guanhua Wang, Shivaram Venkataraman, Amar Phanishayee, Jorgen Thelin, Nikhil Devanur, Ion Stoica
Conference on Machine Learning and Systems (MLSys 2020) | March 2020
-
THEMIS: Fair and Efficient GPU Cluster Scheduling
Kshiteej Mahajan, Arjun Balasubramanian, Arjun Singhvi, Shivaram Venkataraman, Aditya Akella, Amar Phanishayee, Shuchi Chawla
USENIX NSDI 2020 | February 2020
-
-
-
PipeDream: Generalized Pipeline Parallelism for DNN Training
Deepak Narayanan, Aaron Harlap, Amar Phanishayee, Vivek Seshadri, Nikhil Devanur, Greg Granger, Phil Gibbons, Matei Zaharia
ACM Symposium on Operating Systems Principles (SOSP 2019) | October 2019
-
The Case for Unifying Data Loading in Machine Learning Clusters
Aarati Kakaraparthy, Abhay Venkatesh, Amar Phanishayee, Shivaram Venkataraman
USENIX HotCloud | July 2019
-
Analysis of Large-Scale Multi-Tenant {GPU} Clusters for {DNN} Training Workloads
Myeongjae Jeon, Shivaram Venkataraman, Amar Phanishayee, Junjie Qian, Wencong Xiao, Fan Yang
2019 USENIX Annual Technical Conference | May 2019
-
-
-
Accelerating Deep Learning Workloads Through Efficient Multi-Model Execution
Deepak Narayanan, Keshav Santhanam, Amar Phanishayee, Matei Zaharia
NeurIPS Workshop on Systems for Machine Learning | December 2018
-
Parameter Hub: a Rack-Scale Parameter Server for Distributed Deep Neural Network Training
Liang Luo, Jacob Nelson, Luis Ceze, Amar Phanishayee, Arvind Krishnamurthy
SOCC 2018 | October 2018
-
TBD: Benchmarking and Analyzing Deep Neural Network Training
Hongyu Zhu, Mohamed Akrout, Bojian Zheng, Andrew Pelegris, Amar Phanishayee, Bianca Schroeder, Gennady Pekhimenko
International Symposium on Workload Characterization (IISWC 2018) | August 2018
-
Gist: Efficient Data Encoding for Deep Neural Network Training
Animesh Jain, Amar Phanishayee, Jason Mars, Lingjia Tang, Gennady Pekhimenko
International Symposium on Computer Architecture (ISCA 2018) | June 2018
-
PipeDream: Fast and Efficient Pipeline Parallel DNN Training
Aaron Harlap, Deepak Narayanan, Amar Phanishayee, Vivek Seshadri, Nikhil Devanur, Greg Ganger, Phil Gibbons
June 2018
arXiv preprint
-
Multi-tenant GPU Clusters for Deep Learning Workloads: Analysis and Implications (TR)
Myeongjae Jeon, Shivaram Venkataraman, Amar Phanishayee, Junjie Qian, Wencong Xiao, Fan Yang
MSR-TR-2018-13 | May 2018
Publié par Microsoft
-
-
-
Integrated Hardware Architecture and Device Placement Search
Irene Wang, Jakub Tarnawski, Amar Phanishayee, Divya Mahajan
2024 International Conference on Machine Learning | July 2024
-
DéjàVu: KV-cache Streaming for Fast, Fault-tolerant Generative LLM Serving
Foteini Strati, Sara Mcallister, Amar Phanishayee, Jakub Tarnawski, Ana Klimovic
ICML 2024 | July 2024
-
MGit: A Model Versioning and Management System
Wei Hao, Daniel Mendoza, Rafael da Silva, Deepak Narayanan, Amar Phanishayee, Asaf Cidon, Junfeng Yang
2024 International Conference on Machine Learning | July 2024
-
Workload-Aware Hardware Accelerator Mining for Distributed Deep Learning Training
Muhammad Adnan, Amar Phanishayee,, Janardhan (Jana) Kulkarni, Prashant J. Nair, Divya Mahajan
arXiv:2404.14632 | April 2024
Publié par Microsoft
-
Blox: A Modular Toolkit for Deep Learning Schedulers
Saurabh Agarwal, Amar Phanishayee, Shivaram Venkataraman
European Conference on Computer Systems (ACM EuroSys 2024 - spring accept). | December 2023
arXiv:2312.12621 | December 2023
-
Packrat: Automatic Reconfiguration for Latency Minimization in CPU-based DNN Serving
Ankit Bhardwaj, Amar Phanishayee, Deepak Narayanan, Mihail Tarta, Ryan Stutsman
arXiv:2311.18174 | November 2023
Publié par arXiv | Nov 2023
-
MGit: A Model Versioning and Management System
Wei Hao, Daniel Mendoza, Rafael da Silva, Deepak Narayanan, Amar Phanishayee
arXiv:2307.07507 | July 2023
Publié par arXiv
-
A Study on the Intersection of GPU Utilization and CNN Inference
Jack Kosaian, Amar Phanishayee
arXiv:2212.07936 | December 2022
Publié par arXiv
-
Harmony: Overcoming the hurdles of GPU memory capacity to train massive DNN models on commodity servers
Youjie Li, Amar Phanishayee, Derek Murray, Jakub Tarnawski, Nam Sung Kim
VLDB 2022 | September 2022
-
Looking Beyond GPUs for DNN Scheduling on Multi-Tenant Clusters
Jayashree Mohan, Amar Phanishayee, Janardhan (Jana) Kulkarni, Vijay Chidambaram
USENIX Symposium on Operating Systems Design and Implementation (OSDI 2022) | July 2022
-
Piper: Multidimensional Planner for DNN Parallelization
Jakub Tarnawski, Deepak Narayanan, Amar Phanishayee
NeurIPS 2021 | December 2021
-
Efficient Large-Scale Language Model Training on GPU Clusters Using Megatron-LM
Deepak Narayanan, Mohammad Shoeybi, Jared Casper, Patrick LeGresley, Mostofa Patwary, Vijay Anand Korthikanti, Dmitri Vainbrand, Prethvi Kashinkunti, Julie Bernauer, Bryan Catanzaro, Amar Phanishayee, Matei Zaharia
International Conference for High Performance Computing, Networking, Storage and Analysis (SC 2021) | November 2021
Best Student Paper
-
Synergy: Resource Sensitive DNN Scheduling in Multi-Tenant Clusters
Jayashree Mohan, Amar Phanishayee, Janardhan (Jana) Kulkarni, Vijay Chidambaram
October 2021
-
Boosting the Throughput and Accelerator Utilization of Specialized CNN Inference Beyond Increasing Batch Size
Jack Kosaian, Amar Phanishayee, Matthai Philipose, Debadeepta Dey, Rashmi Vinayak
2021 International Conference on Machine Learning (ICML 2021) | July 2021
-
Memory-Efficient Pipeline-Parallel DNN Training
Deepak Narayanan, Amar Phanishayee, Kaiyu Shi, Xie Chen, Matei Zaharia
International Conference on Machine Learning (ICML 2021) | July 2021
-
Doing more with less: Training large DNN models on commodity servers for the masses
Youjie Li, Amar Phanishayee, Derek Murray, Nam Sung Kim
HotOS Workshop | May 2021
-
CheckFreq: Frequent, Fine-Grained DNN Checkpointing
Jayashree Mohan, Amar Phanishayee, Vijay Chidambaram
USENIX FAST 2021 | February 2021
-
Analyzing and Mitigating Data Stalls in DNN Training
Jayashree Mohan, Amar Phanishayee, Ashish Raniwala, Vijay Chidambaram
VLDB 2021 | January 2021
-
Efficient Algorithms for Device Placement of DNN Graph Operators
Jakub Tarnawski, Amar Phanishayee, Nikhil R. Devanur, Divya Mahajan, Fanny Nina Paravecino
NeurIPS 2020 | December 2020
-
Heterogeneity-Aware Cluster Scheduling Policies for Deep Learning Workloads
Deepak Narayanan, Keshav Santhanam, Fiodar Kazhamiaka, Amar Phanishayee, Matei Zaharia
USENIX Symposium on Operating Systems Design and Implementation (OSDI 2020) | November 2020
-
Daydream: Accurately Estimating the Efficacy of Performance Optimizations for DNN Training
Hongyu Zhu, Amar Phanishayee, Gennady Pekhimenko
USENIX ATC 2020 | July 2020
-
The Non-IID Data Quagmire of Decentralized Machine Learning
Kevin Hsieh, Amar Phanishayee, Onur Mutlu, Phillip Gibbons
International Conference on Machine Learning (ICML 2020) | July 2020
-
Memory-Efficient Pipeline-Parallel DNN Training
Deepak Narayanan, Amar Phanishayee, Kaiyu Shi, Xie Chen, Matei Zaharia
arXiv:2006.09503 | June 2020
Publié par arXiv
-
Blink: Fast and Generic Collectives for Distributed ML
Guanhua Wang, Shivaram Venkataraman, Amar Phanishayee, Jorgen Thelin, Nikhil Devanur, Ion Stoica
Conference on Machine Learning and Systems (MLSys 2020) | March 2020
-
THEMIS: Fair and Efficient GPU Cluster Scheduling
Kshiteej Mahajan, Arjun Balasubramanian, Arjun Singhvi, Shivaram Venkataraman, Aditya Akella, Amar Phanishayee, Shuchi Chawla
USENIX NSDI 2020 | February 2020
-
PipeDream: Generalized Pipeline Parallelism for DNN Training
Deepak Narayanan, Aaron Harlap, Amar Phanishayee, Vivek Seshadri, Nikhil Devanur, Greg Granger, Phil Gibbons, Matei Zaharia
ACM Symposium on Operating Systems Principles (SOSP 2019) | October 2019
-
The Case for Unifying Data Loading in Machine Learning Clusters
Aarati Kakaraparthy, Abhay Venkatesh, Amar Phanishayee, Shivaram Venkataraman
USENIX HotCloud | July 2019
-
Analysis of Large-Scale Multi-Tenant {GPU} Clusters for {DNN} Training Workloads
Myeongjae Jeon, Shivaram Venkataraman, Amar Phanishayee, Junjie Qian, Wencong Xiao, Fan Yang
2019 USENIX Annual Technical Conference | May 2019
-
Accelerating Deep Learning Workloads Through Efficient Multi-Model Execution
Deepak Narayanan, Keshav Santhanam, Amar Phanishayee, Matei Zaharia
NeurIPS Workshop on Systems for Machine Learning | December 2018
-
Parameter Hub: a Rack-Scale Parameter Server for Distributed Deep Neural Network Training
Liang Luo, Jacob Nelson, Luis Ceze, Amar Phanishayee, Arvind Krishnamurthy
SOCC 2018 | October 2018
-
TBD: Benchmarking and Analyzing Deep Neural Network Training
Hongyu Zhu, Mohamed Akrout, Bojian Zheng, Andrew Pelegris, Amar Phanishayee, Bianca Schroeder, Gennady Pekhimenko
International Symposium on Workload Characterization (IISWC 2018) | August 2018
-
Gist: Efficient Data Encoding for Deep Neural Network Training
Animesh Jain, Amar Phanishayee, Jason Mars, Lingjia Tang, Gennady Pekhimenko
International Symposium on Computer Architecture (ISCA 2018) | June 2018
-
PipeDream: Fast and Efficient Pipeline Parallel DNN Training
Aaron Harlap, Deepak Narayanan, Amar Phanishayee, Vivek Seshadri, Nikhil Devanur, Greg Ganger, Phil Gibbons
June 2018
arXiv preprint
-
Multi-tenant GPU Clusters for Deep Learning Workloads: Analysis and Implications (TR)
Myeongjae Jeon, Shivaram Venkataraman, Amar Phanishayee, Junjie Qian, Wencong Xiao, Fan Yang
MSR-TR-2018-13 | May 2018
Publié par Microsoft
-
-
-
Integrated Hardware Architecture and Device Placement Search
Irene Wang, Jakub Tarnawski, Amar Phanishayee, Divya Mahajan
2024 International Conference on Machine Learning | July 2024
-
DéjàVu: KV-cache Streaming for Fast, Fault-tolerant Generative LLM Serving
Foteini Strati, Sara Mcallister, Amar Phanishayee, Jakub Tarnawski, Ana Klimovic
ICML 2024 | July 2024
-
MGit: A Model Versioning and Management System
Wei Hao, Daniel Mendoza, Rafael da Silva, Deepak Narayanan, Amar Phanishayee, Asaf Cidon, Junfeng Yang
2024 International Conference on Machine Learning | July 2024
-
Blox: A Modular Toolkit for Deep Learning Schedulers
Saurabh Agarwal, Amar Phanishayee, Shivaram Venkataraman
European Conference on Computer Systems (ACM EuroSys 2024 - spring accept). | December 2023
arXiv:2312.12621 | December 2023
-
Packrat: Automatic Reconfiguration for Latency Minimization in CPU-based DNN Serving
Ankit Bhardwaj, Amar Phanishayee, Deepak Narayanan, Mihail Tarta, Ryan Stutsman
arXiv:2311.18174 | November 2023
Publié par arXiv | Nov 2023
-
MGit: A Model Versioning and Management System
Wei Hao, Daniel Mendoza, Rafael da Silva, Deepak Narayanan, Amar Phanishayee
arXiv:2307.07507 | July 2023
Publié par arXiv
-
A Study on the Intersection of GPU Utilization and CNN Inference
Jack Kosaian, Amar Phanishayee
arXiv:2212.07936 | December 2022
Publié par arXiv
-
Harmony: Overcoming the hurdles of GPU memory capacity to train massive DNN models on commodity servers
Youjie Li, Amar Phanishayee, Derek Murray, Jakub Tarnawski, Nam Sung Kim
VLDB 2022 | September 2022
-
Looking Beyond GPUs for DNN Scheduling on Multi-Tenant Clusters
Jayashree Mohan, Amar Phanishayee, Janardhan (Jana) Kulkarni, Vijay Chidambaram
USENIX Symposium on Operating Systems Design and Implementation (OSDI 2022) | July 2022
-
Piper: Multidimensional Planner for DNN Parallelization
Jakub Tarnawski, Deepak Narayanan, Amar Phanishayee
NeurIPS 2021 | December 2021
-
Efficient Large-Scale Language Model Training on GPU Clusters Using Megatron-LM
Deepak Narayanan, Mohammad Shoeybi, Jared Casper, Patrick LeGresley, Mostofa Patwary, Vijay Anand Korthikanti, Dmitri Vainbrand, Prethvi Kashinkunti, Julie Bernauer, Bryan Catanzaro, Amar Phanishayee, Matei Zaharia
International Conference for High Performance Computing, Networking, Storage and Analysis (SC 2021) | November 2021
Best Student Paper
-
Synergy: Resource Sensitive DNN Scheduling in Multi-Tenant Clusters
Jayashree Mohan, Amar Phanishayee, Janardhan (Jana) Kulkarni, Vijay Chidambaram
October 2021
-
Boosting the Throughput and Accelerator Utilization of Specialized CNN Inference Beyond Increasing Batch Size
Jack Kosaian, Amar Phanishayee, Matthai Philipose, Debadeepta Dey, Rashmi Vinayak
2021 International Conference on Machine Learning (ICML 2021) | July 2021
-
Memory-Efficient Pipeline-Parallel DNN Training
Deepak Narayanan, Amar Phanishayee, Kaiyu Shi, Xie Chen, Matei Zaharia
International Conference on Machine Learning (ICML 2021) | July 2021
-
Doing more with less: Training large DNN models on commodity servers for the masses
Youjie Li, Amar Phanishayee, Derek Murray, Nam Sung Kim
HotOS Workshop | May 2021
-
CheckFreq: Frequent, Fine-Grained DNN Checkpointing
Jayashree Mohan, Amar Phanishayee, Vijay Chidambaram
USENIX FAST 2021 | February 2021
-
Analyzing and Mitigating Data Stalls in DNN Training
Jayashree Mohan, Amar Phanishayee, Ashish Raniwala, Vijay Chidambaram
VLDB 2021 | January 2021
-
Efficient Algorithms for Device Placement of DNN Graph Operators
Jakub Tarnawski, Amar Phanishayee, Nikhil R. Devanur, Divya Mahajan, Fanny Nina Paravecino
NeurIPS 2020 | December 2020
-
Heterogeneity-Aware Cluster Scheduling Policies for Deep Learning Workloads
Deepak Narayanan, Keshav Santhanam, Fiodar Kazhamiaka, Amar Phanishayee, Matei Zaharia
USENIX Symposium on Operating Systems Design and Implementation (OSDI 2020) | November 2020
-
Daydream: Accurately Estimating the Efficacy of Performance Optimizations for DNN Training
Hongyu Zhu, Amar Phanishayee, Gennady Pekhimenko
USENIX ATC 2020 | July 2020
-
The Non-IID Data Quagmire of Decentralized Machine Learning
Kevin Hsieh, Amar Phanishayee, Onur Mutlu, Phillip Gibbons
International Conference on Machine Learning (ICML 2020) | July 2020
-
Memory-Efficient Pipeline-Parallel DNN Training
Deepak Narayanan, Amar Phanishayee, Kaiyu Shi, Xie Chen, Matei Zaharia
arXiv:2006.09503 | June 2020
Publié par arXiv
-
Blink: Fast and Generic Collectives for Distributed ML
Guanhua Wang, Shivaram Venkataraman, Amar Phanishayee, Jorgen Thelin, Nikhil Devanur, Ion Stoica
Conference on Machine Learning and Systems (MLSys 2020) | March 2020
-
THEMIS: Fair and Efficient GPU Cluster Scheduling
Kshiteej Mahajan, Arjun Balasubramanian, Arjun Singhvi, Shivaram Venkataraman, Aditya Akella, Amar Phanishayee, Shuchi Chawla
USENIX NSDI 2020 | February 2020
-
PipeDream: Generalized Pipeline Parallelism for DNN Training
Deepak Narayanan, Aaron Harlap, Amar Phanishayee, Vivek Seshadri, Nikhil Devanur, Greg Granger, Phil Gibbons, Matei Zaharia
ACM Symposium on Operating Systems Principles (SOSP 2019) | October 2019
-
The Case for Unifying Data Loading in Machine Learning Clusters
Aarati Kakaraparthy, Abhay Venkatesh, Amar Phanishayee, Shivaram Venkataraman
USENIX HotCloud | July 2019
-
Accelerating Deep Learning Workloads Through Efficient Multi-Model Execution
Deepak Narayanan, Keshav Santhanam, Amar Phanishayee, Matei Zaharia
NeurIPS Workshop on Systems for Machine Learning | December 2018
-
Parameter Hub: a Rack-Scale Parameter Server for Distributed Deep Neural Network Training
Liang Luo, Jacob Nelson, Luis Ceze, Amar Phanishayee, Arvind Krishnamurthy
SOCC 2018 | October 2018
-
TBD: Benchmarking and Analyzing Deep Neural Network Training
Hongyu Zhu, Mohamed Akrout, Bojian Zheng, Andrew Pelegris, Amar Phanishayee, Bianca Schroeder, Gennady Pekhimenko
International Symposium on Workload Characterization (IISWC 2018) | August 2018
-
Gist: Efficient Data Encoding for Deep Neural Network Training
Animesh Jain, Amar Phanishayee, Jason Mars, Lingjia Tang, Gennady Pekhimenko
International Symposium on Computer Architecture (ISCA 2018) | June 2018
-
PipeDream: Fast and Efficient Pipeline Parallel DNN Training
Aaron Harlap, Deepak Narayanan, Amar Phanishayee, Vivek Seshadri, Nikhil Devanur, Greg Ganger, Phil Gibbons
June 2018
arXiv preprint
-
-
-
Efficient Algorithms for Device Placement of DNN Graph Operators
Jakub Tarnawski, Amar Phanishayee, Nikhil R. Devanur, Divya Mahajan, Fanny Nina Paravecino
NeurIPS 2020 | December 2020
-
-
-
Workload-Aware Hardware Accelerator Mining for Distributed Deep Learning Training
Muhammad Adnan, Amar Phanishayee,, Janardhan (Jana) Kulkarni, Prashant J. Nair, Divya Mahajan
arXiv:2404.14632 | April 2024
Publié par Microsoft
-
Packrat: Automatic Reconfiguration for Latency Minimization in CPU-based DNN Serving
Ankit Bhardwaj, Amar Phanishayee, Deepak Narayanan, Mihail Tarta, Ryan Stutsman
arXiv:2311.18174 | November 2023
Publié par arXiv | Nov 2023
-
MGit: A Model Versioning and Management System
Wei Hao, Daniel Mendoza, Rafael da Silva, Deepak Narayanan, Amar Phanishayee
arXiv:2307.07507 | July 2023
Publié par arXiv
-
A Study on the Intersection of GPU Utilization and CNN Inference
Jack Kosaian, Amar Phanishayee
arXiv:2212.07936 | December 2022
Publié par arXiv
-
Memory-Efficient Pipeline-Parallel DNN Training
Deepak Narayanan, Amar Phanishayee, Kaiyu Shi, Xie Chen, Matei Zaharia
arXiv:2006.09503 | June 2020
Publié par arXiv
-
Multi-tenant GPU Clusters for Deep Learning Workloads: Analysis and Implications (TR)
Myeongjae Jeon, Shivaram Venkataraman, Amar Phanishayee, Junjie Qian, Wencong Xiao, Fan Yang
MSR-TR-2018-13 | May 2018
Publié par Microsoft
-
-
-
Integrated Hardware Architecture and Device Placement Search
Irene Wang, Jakub Tarnawski, Amar Phanishayee, Divya Mahajan
2024 International Conference on Machine Learning | July 2024
-
DéjàVu: KV-cache Streaming for Fast, Fault-tolerant Generative LLM Serving
Foteini Strati, Sara Mcallister, Amar Phanishayee, Jakub Tarnawski, Ana Klimovic
ICML 2024 | July 2024
-
MGit: A Model Versioning and Management System
Wei Hao, Daniel Mendoza, Rafael da Silva, Deepak Narayanan, Amar Phanishayee, Asaf Cidon, Junfeng Yang
2024 International Conference on Machine Learning | July 2024
-
Blox: A Modular Toolkit for Deep Learning Schedulers
Saurabh Agarwal, Amar Phanishayee, Shivaram Venkataraman
European Conference on Computer Systems (ACM EuroSys 2024 - spring accept). | December 2023
arXiv:2312.12621 | December 2023
-
Harmony: Overcoming the hurdles of GPU memory capacity to train massive DNN models on commodity servers
Youjie Li, Amar Phanishayee, Derek Murray, Jakub Tarnawski, Nam Sung Kim
VLDB 2022 | September 2022
-
Looking Beyond GPUs for DNN Scheduling on Multi-Tenant Clusters
Jayashree Mohan, Amar Phanishayee, Janardhan (Jana) Kulkarni, Vijay Chidambaram
USENIX Symposium on Operating Systems Design and Implementation (OSDI 2022) | July 2022
-
Piper: Multidimensional Planner for DNN Parallelization
Jakub Tarnawski, Deepak Narayanan, Amar Phanishayee
NeurIPS 2021 | December 2021
-
Efficient Large-Scale Language Model Training on GPU Clusters Using Megatron-LM
Deepak Narayanan, Mohammad Shoeybi, Jared Casper, Patrick LeGresley, Mostofa Patwary, Vijay Anand Korthikanti, Dmitri Vainbrand, Prethvi Kashinkunti, Julie Bernauer, Bryan Catanzaro, Amar Phanishayee, Matei Zaharia
International Conference for High Performance Computing, Networking, Storage and Analysis (SC 2021) | November 2021
Best Student Paper
-
Boosting the Throughput and Accelerator Utilization of Specialized CNN Inference Beyond Increasing Batch Size
Jack Kosaian, Amar Phanishayee, Matthai Philipose, Debadeepta Dey, Rashmi Vinayak
2021 International Conference on Machine Learning (ICML 2021) | July 2021
-
Memory-Efficient Pipeline-Parallel DNN Training
Deepak Narayanan, Amar Phanishayee, Kaiyu Shi, Xie Chen, Matei Zaharia
International Conference on Machine Learning (ICML 2021) | July 2021
-
Doing more with less: Training large DNN models on commodity servers for the masses
Youjie Li, Amar Phanishayee, Derek Murray, Nam Sung Kim
HotOS Workshop | May 2021
-
CheckFreq: Frequent, Fine-Grained DNN Checkpointing
Jayashree Mohan, Amar Phanishayee, Vijay Chidambaram
USENIX FAST 2021 | February 2021
-
Analyzing and Mitigating Data Stalls in DNN Training
Jayashree Mohan, Amar Phanishayee, Ashish Raniwala, Vijay Chidambaram
VLDB 2021 | January 2021
-
Efficient Algorithms for Device Placement of DNN Graph Operators
Jakub Tarnawski, Amar Phanishayee, Nikhil R. Devanur, Divya Mahajan, Fanny Nina Paravecino
NeurIPS 2020 | December 2020
-
Heterogeneity-Aware Cluster Scheduling Policies for Deep Learning Workloads
Deepak Narayanan, Keshav Santhanam, Fiodar Kazhamiaka, Amar Phanishayee, Matei Zaharia
USENIX Symposium on Operating Systems Design and Implementation (OSDI 2020) | November 2020
-
Daydream: Accurately Estimating the Efficacy of Performance Optimizations for DNN Training
Hongyu Zhu, Amar Phanishayee, Gennady Pekhimenko
USENIX ATC 2020 | July 2020
-
The Non-IID Data Quagmire of Decentralized Machine Learning
Kevin Hsieh, Amar Phanishayee, Onur Mutlu, Phillip Gibbons
International Conference on Machine Learning (ICML 2020) | July 2020
-
Blink: Fast and Generic Collectives for Distributed ML
Guanhua Wang, Shivaram Venkataraman, Amar Phanishayee, Jorgen Thelin, Nikhil Devanur, Ion Stoica
Conference on Machine Learning and Systems (MLSys 2020) | March 2020
-
THEMIS: Fair and Efficient GPU Cluster Scheduling
Kshiteej Mahajan, Arjun Balasubramanian, Arjun Singhvi, Shivaram Venkataraman, Aditya Akella, Amar Phanishayee, Shuchi Chawla
USENIX NSDI 2020 | February 2020
-
PipeDream: Generalized Pipeline Parallelism for DNN Training
Deepak Narayanan, Aaron Harlap, Amar Phanishayee, Vivek Seshadri, Nikhil Devanur, Greg Granger, Phil Gibbons, Matei Zaharia
ACM Symposium on Operating Systems Principles (SOSP 2019) | October 2019
-
The Case for Unifying Data Loading in Machine Learning Clusters
Aarati Kakaraparthy, Abhay Venkatesh, Amar Phanishayee, Shivaram Venkataraman
USENIX HotCloud | July 2019
-
Analysis of Large-Scale Multi-Tenant {GPU} Clusters for {DNN} Training Workloads
Myeongjae Jeon, Shivaram Venkataraman, Amar Phanishayee, Junjie Qian, Wencong Xiao, Fan Yang
2019 USENIX Annual Technical Conference | May 2019
-
Accelerating Deep Learning Workloads Through Efficient Multi-Model Execution
Deepak Narayanan, Keshav Santhanam, Amar Phanishayee, Matei Zaharia
NeurIPS Workshop on Systems for Machine Learning | December 2018
-
Parameter Hub: a Rack-Scale Parameter Server for Distributed Deep Neural Network Training
Liang Luo, Jacob Nelson, Luis Ceze, Amar Phanishayee, Arvind Krishnamurthy
SOCC 2018 | October 2018
-
TBD: Benchmarking and Analyzing Deep Neural Network Training
Hongyu Zhu, Mohamed Akrout, Bojian Zheng, Andrew Pelegris, Amar Phanishayee, Bianca Schroeder, Gennady Pekhimenko
International Symposium on Workload Characterization (IISWC 2018) | August 2018
-
Gist: Efficient Data Encoding for Deep Neural Network Training
Animesh Jain, Amar Phanishayee, Jason Mars, Lingjia Tang, Gennady Pekhimenko
International Symposium on Computer Architecture (ISCA 2018) | June 2018
-
-
-
Synergy: Resource Sensitive DNN Scheduling in Multi-Tenant Clusters
Jayashree Mohan, Amar Phanishayee, Janardhan (Jana) Kulkarni, Vijay Chidambaram
October 2021
-
-
-
PipeDream: Fast and Efficient Pipeline Parallel DNN Training
Aaron Harlap, Deepak Narayanan, Amar Phanishayee, Vivek Seshadri, Nikhil Devanur, Greg Ganger, Phil Gibbons
June 2018
arXiv preprint
-