Shivaram Venkataraman
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Shivaram Venkataraman has a total of 98 co-authors. Here are the top co-authors by number of shared publications:
- Matthew D. Sinclair
University of Wisconsin–Madison, Madison, WI, USA
6 shared publications - Rutwik Jain
University of Wisconsin-Madison, Madison, WI, USA
6 shared publications - Konstantinos Kanellis
University of Wisconsin-Madison, Madison, WI, USA
5 shared publications - Johannes Freischuetz
University of Wisconsin - Madison, Madison, Wisconsin, USA
3 shared publications - Seth Ockerman
University of Wisconsin-Madison
3 shared publications - Sujay Yadalam
The University of Texas at Austin, Austin, USA
3 shared publications
Latest Publications
- Open Access
- 29 September 2026
- Published by Association for Computing Machinery (ACM)
- p. 14-20
Survival of the Luckiest: The Impacts of Noisy Evaluation in Agentic Evolution for Systems
- 29 September 2026
- Published by Association for Computing Machinery (ACM)
- p. 14-20
Open AccessSurvival of the Luckiest: The Impacts of Noisy Evaluation in Agentic Evolution for Systems
- Open Access
- 28 September 2026
- Published by Association for Computing Machinery (ACM)
- p. 1-8
Regional Policies for Memory Tiering Systems
- 28 September 2026
- Published by Association for Computing Machinery (ACM)
- p. 1-8
Open AccessRegional Policies for Memory Tiering Systems
- Open Access
- 20 July 2026
- Published by Association for Computing Machinery (ACM) in ACM SIGMETRICS Performance Evaluation Review
- Vol. 54  (1) , 117-119
Minos: Systematically Classifying Performance and Power Characteristics of GPU Workloads on HPC Clusters
As large-scale HPC compute clusters increasingly adopt accelerators such as GPUs to meet the voracious demands of modern workloads, these clusters are increasingly becoming power constrained. Given that modern applications can often temporarily exceed the power ratings of the accelerators (''power spikes''), current and future HPC systems must optimize for both power and performance together. However, this is made difficult by increasingly diverse applications, which often require bespoke optimizations to run efficiently on each cluster. Traditionally researchers overcome this problem by profiling applications on specific clusters and optimizing, but the scale, algorithmic diversity, and lack of effective tools make this challenging. To overcome these inefficiencies, we propose Minos, a systematic classification mechanism that identifies similar application characteristics via low-cost profiling for power and performance. This allows us to group similarly behaving workloads into a finite number of distinct classes and reduce the overhead of extensively profiling new workloads. For example, when predicting frequency capping behavior for a previously unseen application, Minos reduces profiling time by 89%. Moreover, across 18 popular graph analytics, HPC, HPC+ML, and ML workloads, Minos achieves a mean error of 4% for power predictions and 3% for performance predictions, significantly improving predictions over state-of-the-art approaches by 10%.
- 20 July 2026
- Published by Association for Computing Machinery (ACM) in ACM SIGMETRICS Performance Evaluation Review
- Vol. 54  (1) , 117-119
Open AccessMinos: Systematically Classifying Performance and Power Characteristics of GPU Workloads on HPC Clusters
As large-scale HPC compute clusters increasingly adopt accelerators such as GPUs to meet the voracious demands of modern workloads, these clusters are increasingly becoming power constrained. Given that modern applications can often temporarily exceed the power ratings of the accelerators (''power spikes''), current and future HPC systems must optimize for both power and performance together. However, this is made difficult by increasingly diverse applications, which often require bespoke optimizations to run efficiently on each cluster. Traditionally researchers overcome this problem by profiling applications on specific clusters and optimizing, but the scale, algorithmic diversity, and lack of effective tools make this challenging. To overcome these inefficiencies, we propose Minos, a systematic classification mechanism that identifies similar application characteristics via low-cost profiling for power and performance. This allows us to group similarly behaving workloads into a finite number of distinct classes and reduce the overhead of extensively profiling new workloads. For example, when predicting frequency capping behavior for a previously unseen application, Minos reduces profiling time by 89%. Moreover, across 18 popular graph analytics, HPC, HPC+ML, and ML workloads, Minos achieves a mean error of 4% for power predictions and 3% for performance predictions, significantly improving predictions over state-of-the-art approaches by 10%.
- Open Access
- 8 June 2026
- Published by Association for Computing Machinery (ACM)
- p. 117-119
Minos: Systematically Classifying Performance and Power Characteristics of GPU Workloads on HPC Clusters
As large-scale HPC compute clusters increasingly adopt accelerators such as GPUs to meet the voracious demands of modern workloads, these clusters are increasingly becoming power constrained. Given that modern applications can often temporarily exceed the power ratings of the accelerators (''power spikes''), current and future HPC systems must optimize for both power and performance together. However, this is made difficult by increasingly diverse applications, which often require bespoke optimizations to run efficiently on each cluster. Traditionally researchers overcome this problem by profiling applications on specific clusters and optimizing, but the scale, algorithmic diversity, and lack of effective tools make this challenging. To overcome these inefficiencies, we propose Minos, a systematic classification mechanism that identifies similar application characteristics via low-cost profiling for power and performance. This allows us to group similarly behaving workloads into a finite number of distinct classes and reduce the overhead of extensively profiling new workloads. For example, when predicting frequency capping behavior for a previously unseen application, Minos reduces profiling time by 89%. Moreover, across 18 popular graph analytics, HPC, HPC+ML, and ML workloads, Minos achieves a mean error of 4% for power predictions and 3% for performance predictions, significantly improving predictions over state-of-the-art approaches by 10%.
- 8 June 2026
- Published by Association for Computing Machinery (ACM)
- p. 117-119
Open AccessMinos: Systematically Classifying Performance and Power Characteristics of GPU Workloads on HPC Clusters
As large-scale HPC compute clusters increasingly adopt accelerators such as GPUs to meet the voracious demands of modern workloads, these clusters are increasingly becoming power constrained. Given that modern applications can often temporarily exceed the power ratings of the accelerators (''power spikes''), current and future HPC systems must optimize for both power and performance together. However, this is made difficult by increasingly diverse applications, which often require bespoke optimizations to run efficiently on each cluster. Traditionally researchers overcome this problem by profiling applications on specific clusters and optimizing, but the scale, algorithmic diversity, and lack of effective tools make this challenging. To overcome these inefficiencies, we propose Minos, a systematic classification mechanism that identifies similar application characteristics via low-cost profiling for power and performance. This allows us to group similarly behaving workloads into a finite number of distinct classes and reduce the overhead of extensively profiling new workloads. For example, when predicting frequency capping behavior for a previously unseen application, Minos reduces profiling time by 89%. Moreover, across 18 popular graph analytics, HPC, HPC+ML, and ML workloads, Minos achieves a mean error of 4% for power predictions and 3% for performance predictions, significantly improving predictions over state-of-the-art approaches by 10%.
- Open Access
- 30 May 2026
- Published by Association for Computing Machinery (ACM)
- p. 102-105
Demo of SemWeave: Semantic Common Expressions for LLM-powered Query Processing
- 30 May 2026
- Published by Association for Computing Machinery (ACM)
- p. 102-105
Open AccessDemo of SemWeave: Semantic Common Expressions for LLM-powered Query Processing
- Open Access
- 29 May 2026
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Measurement and Analysis of Computing Systems
- Vol. 10  (2) , 1-27
Minos: Systematically Classifying Performance and Power Characteristics of GPU Workloads on HPC Clusters
As large-scale HPC compute clusters increasingly adopt accelerators such as GPUs to meet the voracious demands of modern workloads, these clusters are increasingly becoming power constrained. Given that modern applications can often temporarily exceed the power ratings of the accelerators (''power spikes''), current and future HPC systems must optimize for both power and performance together. However, this is made difficult by increasingly diverse applications, which often require bespoke optimizations to run efficiently on each cluster. Traditionally researchers overcome this problem by profiling applications on specific clusters and optimizing, but the scale, algorithmic diversity, and lack of effective tools make this challenging. To overcome these inefficiencies, we propose Minos, a systematic classification mechanism that identifies similar application characteristics via low-cost profiling for power and performance. This allows us to group similarly behaving workloads into a finite number of distinct classes and reduce the overhead of extensively profiling new workloads. For example, when predicting frequency capping behavior for a previously unseen application, Minos reduces profiling time by 89%. Moreover, across 18 popular graph analytics, HPC, HPC+ML, and ML workloads, Minos achieves a mean error of 4% for power predictions and 3% for performance predictions, significantly improving predictions over state-of-the-art approaches by 10%.
- 29 May 2026
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Measurement and Analysis of Computing Systems
- Vol. 10  (2) , 1-27
Open AccessMinos: Systematically Classifying Performance and Power Characteristics of GPU Workloads on HPC Clusters
As large-scale HPC compute clusters increasingly adopt accelerators such as GPUs to meet the voracious demands of modern workloads, these clusters are increasingly becoming power constrained. Given that modern applications can often temporarily exceed the power ratings of the accelerators (''power spikes''), current and future HPC systems must optimize for both power and performance together. However, this is made difficult by increasingly diverse applications, which often require bespoke optimizations to run efficiently on each cluster. Traditionally researchers overcome this problem by profiling applications on specific clusters and optimizing, but the scale, algorithmic diversity, and lack of effective tools make this challenging. To overcome these inefficiencies, we propose Minos, a systematic classification mechanism that identifies similar application characteristics via low-cost profiling for power and performance. This allows us to group similarly behaving workloads into a finite number of distinct classes and reduce the overhead of extensively profiling new workloads. For example, when predicting frequency capping behavior for a previously unseen application, Minos reduces profiling time by 89%. Moreover, across 18 popular graph analytics, HPC, HPC+ML, and ML workloads, Minos achieves a mean error of 4% for power predictions and 3% for performance predictions, significantly improving predictions over state-of-the-art approaches by 10%.
- 21 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computers
- Vol. 75  (4) , 1378-1390
From Good to Great: Parameter Tuning in Memory Tiering Systems
To meet the memory demands of modern data-intensive applications such as graph processing and machine learning, datacenters are adopting tiered memory architectures that include fast tiers such as locally attached DRAM and slower tiers built using CXL-attached memory. While these newer tiers expand capacity, they have higher access latencies and lower bandwidth, which can lead to application performance degradation. To minimize performance loss, frequently accessed (hot) data must be placed close to the CPU in faster tiers and infrequently accessed (cold) data can be placed in farther, slower memory tiers. Existing software-managed tiering solutions employ heuristics and pre-configured thresholds to determine hot and cold data, and make placement and migration decisions. Unfortunately, due to their use of static thresholds, these systems fail to adapt to different workloads and the underlying hardware, and hence perform poorly. In this paper, we study the relationship between thresholds and parameters used by tiering systems and application performance. We leverage Bayesian Optimization to tune the parameters of three tiering systems: HeMem, Memtis, and HMSDK. We find that applications are sensitive to these parameter values, and that best parameter values vary widely across workloads, inputs, thread counts, memory ratios, and hardware characteristics. Our experiments reveal that the Bayesian optimizer can learn workload behaviors and set parameter values that result in best performance. With proper parameter tuning, performance can improve by up to 2x compared to default configurations, highlighting the importance of parameter selection for memory tiering systems.
- 21 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computers
- Vol. 75  (4) , 1378-1390
From Good to Great: Parameter Tuning in Memory Tiering Systems
To meet the memory demands of modern data-intensive applications such as graph processing and machine learning, datacenters are adopting tiered memory architectures that include fast tiers such as locally attached DRAM and slower tiers built using CXL-attached memory. While these newer tiers expand capacity, they have higher access latencies and lower bandwidth, which can lead to application performance degradation. To minimize performance loss, frequently accessed (hot) data must be placed close to the CPU in faster tiers and infrequently accessed (cold) data can be placed in farther, slower memory tiers. Existing software-managed tiering solutions employ heuristics and pre-configured thresholds to determine hot and cold data, and make placement and migration decisions. Unfortunately, due to their use of static thresholds, these systems fail to adapt to different workloads and the underlying hardware, and hence perform poorly. In this paper, we study the relationship between thresholds and parameters used by tiering systems and application performance. We leverage Bayesian Optimization to tune the parameters of three tiering systems: HeMem, Memtis, and HMSDK. We find that applications are sensitive to these parameter values, and that best parameter values vary widely across workloads, inputs, thread counts, memory ratios, and hardware characteristics. Our experiments reveal that the Bayesian optimizer can learn workload behaviors and set parameter values that result in best performance. With proper parameter tuning, performance can improve by up to 2x compared to default configurations, highlighting the importance of parameter selection for memory tiering systems.
- Open Access
AI Benchmark Democratization and Carpentry
Benchmarks are a cornerstone of modern machine learning, enabling reproducibility, comparison, and scientific progress. However, AI benchmarks are increasingly complex, requiring dynamic, AI-focused workflows. Rapid evolution in model architectures, scale, datasets, and deployment contexts makes evaluation a moving target. Large language models often memorize static benchmarks, causing a gap between benchmark results and real-world performance. Beyond traditional static benchmarks, continuous adaptive benchmarking frameworks are needed to align scientific assessment with deployment risks. This calls for skills and education in AI Benchmark Carpentry. From our experience with MLCommons, educational initiatives, and programs like the DOE's Trillion Parameter Consortium, key barriers include high resource demands, limited access to specialized hardware, lack of benchmark design expertise, and uncertainty in relating results to application domains. Current benchmarks often emphasize peak performance on top-tier hardware, offering limited guidance for diverse, real-world scenarios. Benchmarking must become dynamic, incorporating evolving models, updated data, and heterogeneous platforms while maintaining transparency, reproducibility, and interpretability. Democratization requires both technical innovation and systematic education across levels, building sustained expertise in benchmark design and use. Benchmarks should support application-relevant comparisons, enabling informed, context-sensitive decisions. Dynamic, inclusive benchmarking will ensure evaluation keeps pace with AI evolution and supports responsible, reproducible, and accessible AI deployment. Community efforts can provide a foundation for AI Benchmark Carpentry.
Open AccessAI Benchmark Democratization and Carpentry
Benchmarks are a cornerstone of modern machine learning, enabling reproducibility, comparison, and scientific progress. However, AI benchmarks are increasingly complex, requiring dynamic, AI-focused workflows. Rapid evolution in model architectures, scale, datasets, and deployment contexts makes evaluation a moving target. Large language models often memorize static benchmarks, causing a gap between benchmark results and real-world performance. Beyond traditional static benchmarks, continuous adaptive benchmarking frameworks are needed to align scientific assessment with deployment risks. This calls for skills and education in AI Benchmark Carpentry. From our experience with MLCommons, educational initiatives, and programs like the DOE's Trillion Parameter Consortium, key barriers include high resource demands, limited access to specialized hardware, lack of benchmark design expertise, and uncertainty in relating results to application domains. Current benchmarks often emphasize peak performance on top-tier hardware, offering limited guidance for diverse, real-world scenarios. Benchmarking must become dynamic, incorporating evolving models, updated data, and heterogeneous platforms while maintaining transparency, reproducibility, and interpretability. Democratization requires both technical innovation and systematic education across levels, building sustained expertise in benchmark design and use. Benchmarks should support application-relevant comparisons, enabling informed, context-sensitive decisions. Dynamic, inclusive benchmarking will ensure evaluation keeps pace with AI evolution and supports responsible, reproducible, and accessible AI deployment. Community efforts can provide a foundation for AI Benchmark Carpentry.
- Open Access
- 15 November 2025
- Published by Association for Computing Machinery (ACM)
- p. 575-581
Exploring Distributed Vector Databases Performance on HPC Platforms: A Study with Qdrant
- 15 November 2025
- Published by Association for Computing Machinery (ACM)
- p. 575-581
Open AccessExploring Distributed Vector Databases Performance on HPC Platforms: A Study with Qdrant
- Open Access
- 15 November 2025
- Published by Association for Computing Machinery (ACM)
- p. 217-236
PGT-I: Scaling Spatiotemporal GNNs with Memory-Efficient Distributed Training
- 15 November 2025
- Published by Association for Computing Machinery (ACM)
- p. 217-236
Open AccessPGT-I: Scaling Spatiotemporal GNNs with Memory-Efficient Distributed Training