Tushar Sharma
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Tushar Sharma has a total of 64 co-authors. Here are the top co-authors by number of shared publications:
- Saurabhsingh Rajput
Dalhousie University, Halifax, Canada
8 shared publications - Mootez Saad
Dalhousie University, Halifax, Canada
5 shared publications - Federica Sarro
Computer Science, University College London, London, UNITED KINGDOM OF GREAT BRITAIN AND NORTHERN IRELAND
4 shared publications - Maria Kechagia
University College London, London, UK
4 shared publications - Mootez Saad
Dalhousie University Halifax Canada
3 shared publications - Boqi Chen
University of Ottawa, Ottawa, Canada
2 shared publications
Latest Publications
- Open Access
- 5 July 2026
- Published by Association for Computing Machinery (ACM)
- p. 247-251
CodeGreen: Towards Improving Precision and Portability in Software Energy Measurement
- 5 July 2026
- Published by Association for Computing Machinery (ACM)
- p. 247-251
Open AccessCodeGreen: Towards Improving Precision and Portability in Software Energy Measurement
- Open Access
- 5 July 2026
- Published by Association for Computing Machinery (ACM)
- p. 1376-1380
Energy Flow Graph: Modeling Software Energy Consumption
The growing energy demands of computational systems necessitate a fundamental shift from performance-centric design to one that treats energy consumption as one of the primary design considerations. Current approaches treat energy consumption as an aggregate, deterministic property, overlooking the path-dependent nature of computation, where different execution paths through the same software consume dramatically different energy. We introduce the Energy Flow Graph (EFG), a formal model that represents computational processes as state-transition systems with energy costs for both states and transitions. EFG enables various applications in software engineering, including static analysis of energy-optimal execution paths and a multiplicative cascade model that predicts combined optimization effects without exhaustive testing. Our early experiments demonstrate EFG's versatility across domains: in software programs validated through 3.5 million executions, 15.6% of solutions exhibit high path-dependent variance (CV > 0.1), while structural optimization reveals up to 705× energy reduction. In AI pipelines, the cascade model predicts optimization combinations within 5.1% error, enabling selection from 4.2 million possibilities using only 22 measurements. The EFG transforms energy optimization from trial-and-error to systematic analysis, providing a foundation for green software engineering across computational domains.
- 5 July 2026
- Published by Association for Computing Machinery (ACM)
- p. 1376-1380
Open AccessEnergy Flow Graph: Modeling Software Energy Consumption
The growing energy demands of computational systems necessitate a fundamental shift from performance-centric design to one that treats energy consumption as one of the primary design considerations. Current approaches treat energy consumption as an aggregate, deterministic property, overlooking the path-dependent nature of computation, where different execution paths through the same software consume dramatically different energy. We introduce the Energy Flow Graph (EFG), a formal model that represents computational processes as state-transition systems with energy costs for both states and transitions. EFG enables various applications in software engineering, including static analysis of energy-optimal execution paths and a multiplicative cascade model that predicts combined optimization effects without exhaustive testing. Our early experiments demonstrate EFG's versatility across domains: in software programs validated through 3.5 million executions, 15.6% of solutions exhibit high path-dependent variance (CV > 0.1), while structural optimization reveals up to 705× energy reduction. In AI pipelines, the cascade model predicts optimization combinations within 5.1% error, enabling selection from 4.2 million possibilities using only 22 measurements. The EFG transforms energy optimization from trial-and-error to systematic analysis, providing a foundation for green software engineering across computational domains.
- Open Access
- 12 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 3053-3065
Why Attention Fails: A Taxonomy of Faults in Attention-Based Neural Networks
- 12 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 3053-3065
Open AccessWhy Attention Fails: A Taxonomy of Faults in Attention-Based Neural Networks
- Open Access
- 12 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 453-465
FlipFlop: A Static Analysis-based Energy Optimization Framework for GPU Kernels
Artificial Intelligence (AI) applications, such as Large Language Models, are primarily driven and executed by Graphics Processing Units (GPUs). These GPU programs (kernels) consume substantial amounts of energy, yet software developers often lack the hardware expertise and ad hoc knowledge required to optimize for power efficiency. We propose FlipFlop, a framework using static code analysis to predict energy consumption and recommend Pareto-optimal thread block configurations considering both power consumption and execution time. Our framework requires no runtime execution and analyzes PTX code, a low-level instruction set for CUDA-enabled GPUs. It is validated across a diverse set of GPUs and kernels, including multi-head attention, convolution, and matrix multiplication. FlipFlop achieves 83% accuracy in identifying locally optimal energy-efficient configurations, while also minimizing developer effort by reducing the optimization search space by 93.4%. For multi-head attention kernels, it yields up to 79% energy savings and 106% throughput gains relative to NVIDIA’s occupancy heuristic. By integrating static analysis with real-time monitoring and providing explainable optimization guidance, FlipFlop empowers developers to create sustainable, high-performance GPU software which minimizes environmental and computational costs.
- 12 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 453-465
Open AccessFlipFlop: A Static Analysis-based Energy Optimization Framework for GPU Kernels
Artificial Intelligence (AI) applications, such as Large Language Models, are primarily driven and executed by Graphics Processing Units (GPUs). These GPU programs (kernels) consume substantial amounts of energy, yet software developers often lack the hardware expertise and ad hoc knowledge required to optimize for power efficiency. We propose FlipFlop, a framework using static code analysis to predict energy consumption and recommend Pareto-optimal thread block configurations considering both power consumption and execution time. Our framework requires no runtime execution and analyzes PTX code, a low-level instruction set for CUDA-enabled GPUs. It is validated across a diverse set of GPUs and kernels, including multi-head attention, convolution, and matrix multiplication. FlipFlop achieves 83% accuracy in identifying locally optimal energy-efficient configurations, while also minimizing developer effort by reducing the optimization search space by 93.4%. For multi-head attention kernels, it yields up to 79% energy savings and 106% throughput gains relative to NVIDIA’s occupancy heuristic. By integrating static analysis with real-time monitoring and providing explainable optimization guidance, FlipFlop empowers developers to create sustainable, high-performance GPU software which minimizes environmental and computational costs.
- Open Access
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- p. 1
Mind the Skills Gap: Reimagining the Software Architect’s Role in the LLM Era
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- p. 1
Open AccessMind the Skills Gap: Reimagining the Software Architect’s Role in the LLM Era
- 19 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Software
- Vol. 43 (2) , 32-41
Tu(r)ning AI Green: Exploring Energy Efficiency Cascading With Orthogonal Optimizations
AI’s exponential growth intensifies computational demands and energy challenges. While practitioners employ various optimization techniques, that we refer as “knobs” in this paper, to tune model efficiency, these are typically afterthoughts and applied in isolation without understanding their combined effects on energy efficiency. The goal of this exploratory empirical study is to emphasize on treating energy efficiency as the firstclass design consideration by demonstrating how strategic knob selection across five AI pipeline stages (data, model, training, system, inference) creates cascading efficiency gains. We evaluate 30 experimental variants on ModernBERT, an encoderonly architecture, examining individual techniques and their orthogonal combinations. Results shows that model pruning provides the highest single-knob energy savings (up to 84.6%), while orthogonal combinations reduce energy consumption by up to 94.6% while preserving 95.95% of baseline F1 score. This work provides actionable frameworks for informed green AI that balance efficiency, performance, and environmental responsibility in AI systems.
- 19 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Software
- Vol. 43 (2) , 32-41
Tu(r)ning AI Green: Exploring Energy Efficiency Cascading With Orthogonal Optimizations
AI’s exponential growth intensifies computational demands and energy challenges. While practitioners employ various optimization techniques, that we refer as “knobs” in this paper, to tune model efficiency, these are typically afterthoughts and applied in isolation without understanding their combined effects on energy efficiency. The goal of this exploratory empirical study is to emphasize on treating energy efficiency as the firstclass design consideration by demonstrating how strategic knob selection across five AI pipeline stages (data, model, training, system, inference) creates cascading efficiency gains. We evaluate 30 experimental variants on ModernBERT, an encoderonly architecture, examining individual techniques and their orthogonal combinations. Results shows that model pruning provides the highest single-knob energy savings (up to 84.6%), while orthogonal combinations reduce energy consumption by up to 94.6% while preserving 95.95% of baseline F1 score. This work provides actionable frameworks for informed green AI that balance efficiency, performance, and environmental responsibility in AI systems.
- Open Access
UISearch: Graph-Based Embeddings for Multimodal Enterprise UI Screenshots Retrieval
Enterprise software companies maintain thousands of user interface screens across products and versions, creating critical challenges for design consistency, pattern discovery, and compliance check. Existing approaches rely on visual similarity or text semantics, lacking explicit modeling of structural properties fundamental to user interface (UI) composition. We present a novel graph-based representation that converts UI screenshots into attributed graphs encoding hierarchical relationships and spatial arrangements, potentially generalizable to document layouts, architectural diagrams, and other structured visual domains. A contrastive graph autoencoder learns embeddings preserving multi-level similarity across visual, structural, and semantic properties. The comprehensive analysis demonstrates that our structural embeddings achieve better discriminative power than state-of-the-art Vision Encoders, representing a fundamental advance in the expressiveness of the UI representation. We implement this representation in UISearch, a multi-modal search framework that combines structural embeddings with semantic search through a composable query language. On 20,396 financial software UIs, UISearch achieves 0.92 Top-5 accuracy with 47.5ms median latency (P95: 124ms), scaling to 20,000+ screens. The hybrid indexing architecture enables complex queries and supports fine-grained UI distinction impossible with vision-only approaches.
Open AccessUISearch: Graph-Based Embeddings for Multimodal Enterprise UI Screenshots Retrieval
Enterprise software companies maintain thousands of user interface screens across products and versions, creating critical challenges for design consistency, pattern discovery, and compliance check. Existing approaches rely on visual similarity or text semantics, lacking explicit modeling of structural properties fundamental to user interface (UI) composition. We present a novel graph-based representation that converts UI screenshots into attributed graphs encoding hierarchical relationships and spatial arrangements, potentially generalizable to document layouts, architectural diagrams, and other structured visual domains. A contrastive graph autoencoder learns embeddings preserving multi-level similarity across visual, structural, and semantic properties. The comprehensive analysis demonstrates that our structural embeddings achieve better discriminative power than state-of-the-art Vision Encoders, representing a fundamental advance in the expressiveness of the UI representation. We implement this representation in UISearch, a multi-modal search framework that combines structural embeddings with semantic search through a composable query language. On 20,396 financial software UIs, UISearch achieves 0.92 Top-5 accuracy with 47.5ms median latency (P95: 124ms), scaling to 20,000+ screens. The hybrid indexing architecture enables complex queries and supports fine-grained UI distinction impossible with vision-only approaches.
- Open Access
Smart but Costly? Benchmarking LLMs on Functional Accuracy and Energy Efficiency
The rapid advancement of AI technologies and their accelerated adoption in software development necessitates a systematic evaluation of their environmental impact alongside functional correctness. While prior studies have examined sustainability in large language models, existing approaches lack systematic frameworks for evaluating accuracy-energy trade-offs in Code Language Models (CLMs). In this paper, we present a framework, BRACE, to benchmark CLMs on a unified scale of energy efficiency and functional correctness (referred to as accuracy). We benchmark 22 state-of-the-art models on code generation and summarization tasks, proposing two rating methods: Concentric Incremental Rating Circles (CIRC) and Observation to Expectation Rating (OTER). CIRC provides deterministic Euclidean-based rankings with static trade-offs that are robust to outliers, and OTER offers trend-aware evaluation with dynamic trade-offs that capture the complex correlation between energy and accuracy, each offering a distinct perspective and addressing the problem in a unique way. These rating methods enable us to rate LLMs on a 1-5 scale reflecting their combined capabilities in terms of energy efficiency and functional correctness. Our analysis reveals models generally perform better in the code summarization tasks as they are not enforced to generate a grammar-based and syntactically correct output. Also, we find that models' size does not have a significant impact on their ratings, indicating that if models utilize their parameters efficiently, they can be ranked higher on these scales. The proposed BRACE framework empowers practitioners to make evidence-based model selections that balance sustainability with task requirements, guiding rating choice -- CIRC for deterministic comparisons or OTER for trend-aware evaluation -- based on deployment priorities.
Open AccessSmart but Costly? Benchmarking LLMs on Functional Accuracy and Energy Efficiency
The rapid advancement of AI technologies and their accelerated adoption in software development necessitates a systematic evaluation of their environmental impact alongside functional correctness. While prior studies have examined sustainability in large language models, existing approaches lack systematic frameworks for evaluating accuracy-energy trade-offs in Code Language Models (CLMs). In this paper, we present a framework, BRACE, to benchmark CLMs on a unified scale of energy efficiency and functional correctness (referred to as accuracy). We benchmark 22 state-of-the-art models on code generation and summarization tasks, proposing two rating methods: Concentric Incremental Rating Circles (CIRC) and Observation to Expectation Rating (OTER). CIRC provides deterministic Euclidean-based rankings with static trade-offs that are robust to outliers, and OTER offers trend-aware evaluation with dynamic trade-offs that capture the complex correlation between energy and accuracy, each offering a distinct perspective and addressing the problem in a unique way. These rating methods enable us to rate LLMs on a 1-5 scale reflecting their combined capabilities in terms of energy efficiency and functional correctness. Our analysis reveals models generally perform better in the code summarization tasks as they are not enforced to generate a grammar-based and syntactically correct output. Also, we find that models' size does not have a significant impact on their ratings, indicating that if models utilize their parameters efficiently, they can be ranked higher on these scales. The proposed BRACE framework empowers practitioners to make evidence-based model selections that balance sustainability with task requirements, guiding rating choice -- CIRC for deterministic comparisons or OTER for trend-aware evaluation -- based on deployment priorities.
- Open Access
- 2 November 2025
- Published by Association for Computing Machinery (ACM)
- p. 72-75
Community-Centered Spatial Intelligence for Climate Adaptation at Nova Scotia's Eastern Shore
- 2 November 2025
- Published by Association for Computing Machinery (ACM)
- p. 72-75
Open AccessCommunity-Centered Spatial Intelligence for Climate Adaptation at Nova Scotia's Eastern Shore
- Open Access
On the Effect of Token Merging on Pre-trained Models for Code
Tokenization is a fundamental component of language models for code. It involves breaking down the input into units that are later passed to the language model stack to learn high-dimensional representations used in various contexts, from classification to generation. However, the output of these tokenizers is often longer than that traditionally used in compilers and interpreters. This could result in undesirable effects, such as increased computational overhead. In this work, we investigate the effect of merging the hidden representations of subtokens that belong to the same semantic unit, such as subtokens that form a single identifier. We propose two strategies: one based on averaging the representations and another that leverages a learning-based approach. Both methods can be seamlessly integrated with existing language models for code. We conduct experiments using six language models for code: CodeBERT, GraphCodeBERT, UniXCoder, CdoeT5, CodeT5+ (220M), and CodeT5+ (770M), across three software engineering tasks: vulnerability detection, code classification, and code translation. Results show that these strategies can reduce the number of floating-point operations by $1\%$ to $19\%$. Regarding downstream performance, the most significant degradation was observed in the vulnerability detection task, where the F1 score decreased by $1.82$ points compared to the baseline. In contrast, for code translation, we observed an improvement of $2.47$ points in CodeBLEU. This work contributes to the broader effort of improving language models for code across multiple dimensions, including both computational efficiency and downstream performance.Open AccessOn the Effect of Token Merging on Pre-trained Models for Code
Tokenization is a fundamental component of language models for code. It involves breaking down the input into units that are later passed to the language model stack to learn high-dimensional representations used in various contexts, from classification to generation. However, the output of these tokenizers is often longer than that traditionally used in compilers and interpreters. This could result in undesirable effects, such as increased computational overhead. In this work, we investigate the effect of merging the hidden representations of subtokens that belong to the same semantic unit, such as subtokens that form a single identifier. We propose two strategies: one based on averaging the representations and another that leverages a learning-based approach. Both methods can be seamlessly integrated with existing language models for code. We conduct experiments using six language models for code: CodeBERT, GraphCodeBERT, UniXCoder, CdoeT5, CodeT5+ (220M), and CodeT5+ (770M), across three software engineering tasks: vulnerability detection, code classification, and code translation. Results show that these strategies can reduce the number of floating-point operations by $1\%$ to $19\%$. Regarding downstream performance, the most significant degradation was observed in the vulnerability detection task, where the F1 score decreased by $1.82$ points compared to the baseline. In contrast, for code translation, we observed an improvement of $2.47$ points in CodeBLEU. This work contributes to the broader effort of improving language models for code across multiple dimensions, including both computational efficiency and downstream performance.