Parthasarathy Ranganathan
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Parthasarathy Ranganathan has a total of 131 co-authors. Here are the top co-authors by number of shared publications:
- Urs Hölzle
Google LLC (United States)
14 shared publications - Luiz André Barroso
Google LLC (United States)
12 shared publications - Chris Kennelly
Google, New York, NY, USA
5 shared publications - David Culler
Google
3 shared publications - Martin Dixon
Google
3 shared publications - Milad Hashemi
Google Deepmind, Mountain View, USA
3 shared publications
Latest Publications
- Open Access
- 7 October 2026
- Published by Association for Computing Machinery (ACM)
- p. 4199-4209
Customizing an LLM for Enterprise Software Engineering
- 7 October 2026
- Published by Association for Computing Machinery (ACM)
- p. 4199-4209
Open AccessCustomizing an LLM for Enterprise Software Engineering
- Open Access
- 28 September 2026
- Published by Association for Computing Machinery (ACM)
- p. 827-842
QProf: Fleetwide Transitive Cost Profiling of Warehouse-Scale Services
- 28 September 2026
- Published by Association for Computing Machinery (ACM)
- p. 827-842
Open AccessQProf: Fleetwide Transitive Cost Profiling of Warehouse-Scale Services
- Open Access
- 25 September 2026
- Published by Association for Computing Machinery (ACM) in Communications of the ACM
- Vol. 69 (10) , 46-56
Instruction Set Migration at Warehouse Scale
Migrating codebases from one instruction set architecture (ISA) to another is a major engineering challenge. A recent example is the adoption of Arm (in addition to x86) across major cloud hyperscalers. Yet, this problem has seen limited attention from the academic community. Most work assumes migration is straightforward in the presence of source code, or that the main research challenges lie in binary translation used in its absence. We find that neither assumption holds today. Modern ISA migrations can build on a robust open-source ecosystem, making it possible to recompile all relevant software from scratch. However, while this avoids the need for binary translation, modern software ecosystems introduce a new, multifaceted set of challenges. By analyzing a large-scale migration from x86 to Arm at Google spanning nearly 40,000 code commits, we derive a taxonomy of tasks involved in modern ISA migration. We show how Google automated many of these steps and demonstrate how AI can play a major role in performing them. Finally, we identify tasks that remain difficult and highlight research challenges warranting further attention.
- 25 September 2026
- Published by Association for Computing Machinery (ACM) in Communications of the ACM
- Vol. 69 (10) , 46-56
Open AccessInstruction Set Migration at Warehouse Scale
Migrating codebases from one instruction set architecture (ISA) to another is a major engineering challenge. A recent example is the adoption of Arm (in addition to x86) across major cloud hyperscalers. Yet, this problem has seen limited attention from the academic community. Most work assumes migration is straightforward in the presence of source code, or that the main research challenges lie in binary translation used in its absence. We find that neither assumption holds today. Modern ISA migrations can build on a robust open-source ecosystem, making it possible to recompile all relevant software from scratch. However, while this avoids the need for binary translation, modern software ecosystems introduce a new, multifaceted set of challenges. By analyzing a large-scale migration from x86 to Arm at Google spanning nearly 40,000 code commits, we derive a taxonomy of tasks involved in modern ISA migration. We show how Google automated many of these steps and demonstrate how AI can play a major role in performing them. Finally, we identify tasks that remain difficult and highlight research challenges warranting further attention.
- Open Access
- 1 September 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Micro
- Vol. 46 (5) , 54-61
Necro-reaper: Pruning Away Dead Memory Traffic in Warehouse-Scale Computers
Memory bandwidth is emerging as a critical bottleneck in warehouse-scale computing (WSC). This work reveals that a significant portion of memory traffic in WSC is surprisingly unnecessary, consisting of needless writebacks of deallocated data and fetches of uninitialized data. This issue is particularly acute in WSC, where short-lived heap allocations bigger than a cache line are prevalent. To address this problem, this work proposes a pragmatic approach tailored to WSC. Leveraging the existing WSC ecosystem of vertical integration, profile-guided compilation flows, and customized memory allocators, this work presents NECRO-REAPER, a novel software/hardware co-design that avoids dead memory traffic without requiring the hardware tracking of prior work. Evaluations across diverse WSC workloads demonstrate that NECRO-REAPER reduces memory traffic by 26% and improves IPC by 6%. More broadly, this work drives a paradigm shift in memory efficiency, moving the design goal from simply optimizing raw bandwidth to maximizing semantic goodput.
- 1 September 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Micro
- Vol. 46 (5) , 54-61
Open AccessNecro-reaper: Pruning Away Dead Memory Traffic in Warehouse-Scale Computers
Memory bandwidth is emerging as a critical bottleneck in warehouse-scale computing (WSC). This work reveals that a significant portion of memory traffic in WSC is surprisingly unnecessary, consisting of needless writebacks of deallocated data and fetches of uninitialized data. This issue is particularly acute in WSC, where short-lived heap allocations bigger than a cache line are prevalent. To address this problem, this work proposes a pragmatic approach tailored to WSC. Leveraging the existing WSC ecosystem of vertical integration, profile-guided compilation flows, and customized memory allocators, this work presents NECRO-REAPER, a novel software/hardware co-design that avoids dead memory traffic without requiring the hardware tracking of prior work. Evaluations across diverse WSC workloads demonstrate that NECRO-REAPER reduces memory traffic by 26% and improves IPC by 6%. More broadly, this work drives a paradigm shift in memory efficiency, moving the design goal from simply optimizing raw bandwidth to maximizing semantic goodput.
- Open Access
- 1 September 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Micro
- Vol. 46 (5) , 29-36
Fast and Accurate CPU Performance Modeling With Compositional Analytical–Machine Learning Fusion
Cycle-level simulators like gem5 are widely used in microarchitecture design, but prohibitively slow for large-scale design space exploration. We present Concorde, a new methodology for learning fast and accurate performance models of microarchitectures. Unlike existing simulators and learning approaches that emulate each instruction, Concorde uses simple analytical models to compute per-component performance bounds, producing compact distributions that an ML model maps to program behavior. Concorde is more than five orders of magnitude faster than Google’s cycle-level simulator, with ∼2% average Cycles-Per-Instruction (CPI) prediction error across SPEC2017, open-source, and Google’s proprietary benchmarks. Concorde is already shaping large-scale architecture exploration at Google, contributing to internal CPU design projects. This enables analyses previously infeasible: in about an hour, we conducted a first-of-its-kind fine-grained performance attribution to microarchitectural components, requiring ∼150M CPI evaluations. The same compositional analytical-ML recipe extends beyond CPUs, suggesting a broader approach to building fast, interpretable performance models.
- 1 September 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Micro
- Vol. 46 (5) , 29-36
Open AccessFast and Accurate CPU Performance Modeling With Compositional Analytical–Machine Learning Fusion
Cycle-level simulators like gem5 are widely used in microarchitecture design, but prohibitively slow for large-scale design space exploration. We present Concorde, a new methodology for learning fast and accurate performance models of microarchitectures. Unlike existing simulators and learning approaches that emulate each instruction, Concorde uses simple analytical models to compute per-component performance bounds, producing compact distributions that an ML model maps to program behavior. Concorde is more than five orders of magnitude faster than Google’s cycle-level simulator, with ∼2% average Cycles-Per-Instruction (CPI) prediction error across SPEC2017, open-source, and Google’s proprietary benchmarks. Concorde is already shaping large-scale architecture exploration at Google, contributing to internal CPU design projects. This enables analyses previously infeasible: in about an hour, we conducted a first-of-its-kind fine-grained performance attribution to microarchitectural components, requiring ∼150M CPI evaluations. The same compositional analytical-ML recipe extends beyond CPUs, suggesting a broader approach to building fast, interpretable performance models.
- Open Access
- 22 March 2026
- Published by Association for Computing Machinery (ACM)
- p. 2
Unicorns, Centaurs, and Cyborgs: Co-design Powering the Intelligence Era
- 22 March 2026
- Published by Association for Computing Machinery (ACM)
- p. 2
Open AccessUnicorns, Centaurs, and Cyborgs: Co-design Powering the Intelligence Era
- Top Cited PapersOpen Access
- 1 January 2026
- Published by Springer Nature
The Data Center as a Computer
- 1 January 2026
- Published by Springer Nature
Top Cited PapersOpen AccessThe Data Center as a Computer
- Open Access
- 28 December 2025
- Published by Springer Nature
- p. 55-77
WSC Software: Distributed Systems Infrastructure
This chapter provides a comprehensive overview of the software ecosystem that underpins the operation of Warehouse-Scale Computers (WSCs). It covers the layered structure of the software stack, focusing on distributed systems infrastructure including foundational services and cluster-level infrastructure for resource management, unstructured and structured storage systems like Colossus and Spanner, and distributed networking extending beyond data centers. The discussion emphasizes the importance of infrastructure management, encompassing resource provisioning, configuration management, observability, and tracing tools. We also highlight the significance of practices like DevOps and Site Reliability Engineering (SRE) in managing WSCs effectively. Finally, we discuss server-level software, including the operating system, firmware, containerization, and virtualization, with a focus on security aspects like the Silicon Root of Trust.
- 28 December 2025
- Published by Springer Nature
- p. 55-77
Open AccessWSC Software: Distributed Systems Infrastructure
This chapter provides a comprehensive overview of the software ecosystem that underpins the operation of Warehouse-Scale Computers (WSCs). It covers the layered structure of the software stack, focusing on distributed systems infrastructure including foundational services and cluster-level infrastructure for resource management, unstructured and structured storage systems like Colossus and Spanner, and distributed networking extending beyond data centers. The discussion emphasizes the importance of infrastructure management, encompassing resource provisioning, configuration management, observability, and tracing tools. We also highlight the significance of practices like DevOps and Site Reliability Engineering (SRE) in managing WSCs effectively. Finally, we discuss server-level software, including the operating system, firmware, containerization, and virtualization, with a focus on security aspects like the Silicon Root of Trust.
- Open Access
- 28 December 2025
- Published by Springer Nature
- p. 9-18
Overview of WSC Architecture
This chapter provides an architectural overview of warehouse-scale computers (WSCs). It starts with a discussion of how WSCs are geo-located and explains their physical infrastructure, including buildings, power, and cooling. It then provides an overview of the computing inside WSCs, including servers, storage, and networking. It concludes with a discussion of key WSC tradeoffs around energy efficiency and resilience.
- 28 December 2025
- Published by Springer Nature
- p. 9-18
Open AccessOverview of WSC Architecture
This chapter provides an architectural overview of warehouse-scale computers (WSCs). It starts with a discussion of how WSCs are geo-located and explains their physical infrastructure, including buildings, power, and cooling. It then provides an overview of the computing inside WSCs, including servers, storage, and networking. It concludes with a discussion of key WSC tradeoffs around energy efficiency and resilience.
- Open Access
- 28 December 2025
- Published by Springer Nature
- p. 213-267
Energy and Sustainability
This chapter starts with a high-level characterization of data center energy before moving to a discussion of power, energy, and sustainability in warehouse-scale computers (WSCs). We discuss the economic and environmental importance of energy efficiency, and define metrics like PUE (Power Usage Effectiveness) and SPUE (Server PUE) that help assess and improve data center and server infrastructure efficiency. We also explore the concept of energy-proportional computing and techniques for enhancing the energy efficiency of computing resources. We then discuss data center power provisioning strategies, including oversubscription and energy storage, and conclude with a discussion of environmental sustainability and a high-level characterization of data center energy usage of the internet and cryptocurrency mining.
- 28 December 2025
- Published by Springer Nature
- p. 213-267
Open AccessEnergy and Sustainability
This chapter starts with a high-level characterization of data center energy before moving to a discussion of power, energy, and sustainability in warehouse-scale computers (WSCs). We discuss the economic and environmental importance of energy efficiency, and define metrics like PUE (Power Usage Effectiveness) and SPUE (Server PUE) that help assess and improve data center and server infrastructure efficiency. We also explore the concept of energy-proportional computing and techniques for enhancing the energy efficiency of computing resources. We then discuss data center power provisioning strategies, including oversubscription and energy storage, and conclude with a discussion of environmental sustainability and a high-level characterization of data center energy usage of the internet and cryptocurrency mining.