Aleksandar Ilic
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Aleksandar Ilic has a total of 83 co-authors. Here are the top co-authors by number of shared publications:
- Leonel Sousa
INESC-ID
48 shared publications - Nuno Roma
INESC-ID, Instituto Superior Técnico, Universidade de Lisboa, Lisbon, Portugal
19 shared publications - Pedro Tomás
INESC-ID, Instituto Superior Técnico, Universidade de Lisboa, Lisbon, Portugal
12 shared publications Frederico Pratas
Synopsys, Lisbon, Portugal
6 shared publications- Joao Guerreiro
INESC-ID, Instituto Superior Técnico, Universidade de Lisboa, 1000-029 Lisbon, Portugal
6 shared publications - Sergio Santander-Jiménez
Universidad de Extremadura
5 shared publications
Latest Publications
- Open Access
- 7 August 2026
- Published by Association for Computing Machinery (ACM) in ACM Computing Surveys
- Vol. 58 (14) , 1-36
Survey of GPU Acceleration in Acoustics and Audio Signal Processing
Acoustics and audio signal processing play a substantial role in our daily lives. Multimedia applications and valuable audio tools rely on this field for technological advancements. On the other hand, their efficiency depends on the capabilities of computing systems that provide the required computing resources. Graphics Processing Units (GPUs) were originally designed for graphics rendering, but they have transformed various fields by accelerating computationally intensive tasks with their parallel processing architecture. This survey explores the potential of GPUs in accelerating acoustics and audio signal processing tasks. A comprehensive overview is provided, and it is found that audio filtering, room acoustics, and sound synthesis are the foremost research fields benefiting from this massively parallel processing. Opportunities have been identified for further research in unexplored domains and established areas that require updates due to improved algorithms and new hardware technology. This survey can be useful for anyone looking for a background on the subject and wanting to understand which tools and applications have been explored and tested, and potential future directions.
- 7 August 2026
- Published by Association for Computing Machinery (ACM) in ACM Computing Surveys
- Vol. 58 (14) , 1-36
Open AccessSurvey of GPU Acceleration in Acoustics and Audio Signal Processing
Acoustics and audio signal processing play a substantial role in our daily lives. Multimedia applications and valuable audio tools rely on this field for technological advancements. On the other hand, their efficiency depends on the capabilities of computing systems that provide the required computing resources. Graphics Processing Units (GPUs) were originally designed for graphics rendering, but they have transformed various fields by accelerating computationally intensive tasks with their parallel processing architecture. This survey explores the potential of GPUs in accelerating acoustics and audio signal processing tasks. A comprehensive overview is provided, and it is found that audio filtering, room acoustics, and sound synthesis are the foremost research fields benefiting from this massively parallel processing. Opportunities have been identified for further research in unexplored domains and established areas that require updates due to improved algorithms and new hardware technology. This survey can be useful for anyone looking for a background on the subject and wanting to understand which tools and applications have been explored and tested, and potential future directions.
- Open Access
- 13 July 2026
- Published by Association for Computing Machinery (ACM)
- p. 374-386
Endeavor: Efficient PairHMM for Detection of DNA Variants in Genome-Scale Datasets
DNA variant calling represents a key operation in bioinformatics pipelines that aims at identifying genetic variants. Given an evidenced explosion in genomic data availability, there is an urgent need for a high-performant, portable and efficient solution for variant calling, which can further improve our understanding of genomic structure and genetic basis for complex diseases. In its most common formulation, the Pair Hidden Markov Model (PairHMM) algorithm for variant calling stands as the main bottleneck in the pipeline, accounting for up to 70% of the execution time in large-scale genomic datasets. The state-of-the-art approaches for accelerating PairHMM in CPUs and GPUs do not scale to long DNA sequences and only explore very limited anti-diagonal data parallelism, which yields poor performance. In this work, Endeavor is proposed as a new parallelization strategy for PairHMM that redefines its traditional formulation to explore row-level fine-grained parallelism without loss in solution accuracy. Based on this, a novel and portable SIMD-based approach is derived for efficient and high-performance processing of short and long sequences in CPUs and GPUs, leveraging novel levels of parallelism and synchronization to achieve high throughput in sequences up to 100k basepairs for the first time. Evaluation on Intel and AMD CPUs shows that Endeavor outperforms GKL up to 2.14x in peak throughput and GATK HaplotypeCaller by at least 2x in real-world datasets, while NVIDIA and AMD GPUs achieve up to 2.05x speedups in genome-scale datasets when compared to state-of-the-art GPU-based methods.
- 13 July 2026
- Published by Association for Computing Machinery (ACM)
- p. 374-386
Open AccessEndeavor: Efficient PairHMM for Detection of DNA Variants in Genome-Scale Datasets
DNA variant calling represents a key operation in bioinformatics pipelines that aims at identifying genetic variants. Given an evidenced explosion in genomic data availability, there is an urgent need for a high-performant, portable and efficient solution for variant calling, which can further improve our understanding of genomic structure and genetic basis for complex diseases. In its most common formulation, the Pair Hidden Markov Model (PairHMM) algorithm for variant calling stands as the main bottleneck in the pipeline, accounting for up to 70% of the execution time in large-scale genomic datasets. The state-of-the-art approaches for accelerating PairHMM in CPUs and GPUs do not scale to long DNA sequences and only explore very limited anti-diagonal data parallelism, which yields poor performance. In this work, Endeavor is proposed as a new parallelization strategy for PairHMM that redefines its traditional formulation to explore row-level fine-grained parallelism without loss in solution accuracy. Based on this, a novel and portable SIMD-based approach is derived for efficient and high-performance processing of short and long sequences in CPUs and GPUs, leveraging novel levels of parallelism and synchronization to achieve high throughput in sequences up to 100k basepairs for the first time. Evaluation on Intel and AMD CPUs shows that Endeavor outperforms GKL up to 2.14x in peak throughput and GATK HaplotypeCaller by at least 2x in real-world datasets, while NVIDIA and AMD GPUs achieve up to 2.05x speedups in genome-scale datasets when compared to state-of-the-art GPU-based methods.
- Open Access
- 23 March 2026
- Published by Association for Computing Machinery (ACM)
- p. 664-666
Move the Roof: Model-driven Methodology for Designing Efficient Deep Learning Architectures
- 23 March 2026
- Published by Association for Computing Machinery (ACM)
- p. 664-666
Open AccessMove the Roof: Model-driven Methodology for Designing Efficient Deep Learning Architectures
- Open Access
- 8 September 2025
- Published by Association for Computing Machinery (ACM)
- p. 299-308
Thievory: Graph Processing with Multi-GPU Memory Stealing
- 8 September 2025
- Published by Association for Computing Machinery (ACM)
- p. 299-308
Open AccessThievory: Graph Processing with Multi-GPU Memory Stealing
- 11 June 2025
- Published by Springer Nature
- p. 487-500
Exploring the Limits of Cross-Platform Sparse Tensor Processing
- 11 June 2025
- Published by Springer Nature
- p. 487-500
Exploring the Limits of Cross-Platform Sparse Tensor Processing
- Open Access
- 8 June 2025
- Published by Association for Computing Machinery (ACM)
- p. 293-307
EPIClear: Exploiting Domain-Specific Features for Epistasis Detection Acceleration on Tensor Cores
- 8 June 2025
- Published by Association for Computing Machinery (ACM)
- p. 293-307
Open AccessEPIClear: Exploiting Domain-Specific Features for Epistasis Detection Acceleration on Tensor Cores
- 23 September 2024
- Published by Elsevier in Journal of Parallel and Distributed Computing
- Vol. 195
SpEpistasis: A sparse approach for three-way epistasis detection
- 23 September 2024
- Published by Elsevier in Journal of Parallel and Distributed Computing
- Vol. 195
SpEpistasis: A sparse approach for three-way epistasis detection
- 16 April 2024
- Published by Springer Nature
- p. 97-109
Sparse-Aware CARM: Rooflining Locality of Sparse Computations
- 16 April 2024
- Published by Springer Nature
- p. 97-109
Sparse-Aware CARM: Rooflining Locality of Sparse Computations
- 14 April 2024
- Published by Springer Nature
- p. 143-154
A Performance Modelling-Driven Approach to Hardware Resource Scaling
- 14 April 2024
- Published by Springer Nature
- p. 143-154
A Performance Modelling-Driven Approach to Hardware Resource Scaling
- Open Access
- 11 November 2023
- Published by Association for Computing Machinery (ACM)
- p. 1-13
Bringing Order to Sparsity: A Sparse Matrix Reordering Study on Multicore CPUs
Many real-world computations involve sparse data structures in the form of sparse matrices. A common strategy for optimizing sparse matrix operations is to reorder a matrix to improve data locality. However, it's not always clear whether reordering will provide benefits over the unordered matrix, as its effectiveness depends on several factors, such as structural features of the matrix, the reordering algorithm and the hardware that is used. This paper aims to establish the relationship between matrix reordering algorithms and the performance of sparse matrix operations. We thoroughly evaluate six different matrix reordering algorithms on 490 matrices across eight multicore architectures, focusing on the commonly used sparse matrix-vector multiplication (SpMV) kernel. We find that reordering based on graph partitioning provides better SpMV performance than the alternatives for a large majority of matrices, and that the resulting performance is explained through a combination of data locality and load balancing concerns.
- 11 November 2023
- Published by Association for Computing Machinery (ACM)
- p. 1-13
Open AccessBringing Order to Sparsity: A Sparse Matrix Reordering Study on Multicore CPUs
Many real-world computations involve sparse data structures in the form of sparse matrices. A common strategy for optimizing sparse matrix operations is to reorder a matrix to improve data locality. However, it's not always clear whether reordering will provide benefits over the unordered matrix, as its effectiveness depends on several factors, such as structural features of the matrix, the reordering algorithm and the hardware that is used. This paper aims to establish the relationship between matrix reordering algorithms and the performance of sparse matrix operations. We thoroughly evaluate six different matrix reordering algorithms on 490 matrices across eight multicore architectures, focusing on the commonly used sparse matrix-vector multiplication (SpMV) kernel. We find that reordering based on graph partitioning provides better SpMV performance than the alternatives for a large majority of matrices, and that the resulting performance is explained through a combination of data locality and load balancing concerns.