Filippo Spiga
Research Interests
Filippo Spiga has a total of 78 co-authors. Here are the top co-authors by number of shared publications:
- Fabio Banchelli
Barcelona Supercomputing Center, Barcelona, Spain
2 shared publications - Filippo Mantovani
Barcelona Supercomputing Center, Barcelona, Spain
2 shared publications - Oscar Hernandez
Oak Ridge National Laboratory, Oak Ridge, USA
2 shared publications - Ross Miller
Oak Ridge National Laboratory (ORNL), Oak Ridge, TN (United States)
2 shared publications - Wael R. Elwasif
Oak Ridge National Laboratory
2 shared publications Adriano Tiribocchi
1 shared publication
Latest Publications
- 1 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 224-233
Accelerating the Particle-In-Cell Code ECsim with OpenACC
- 1 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 224-233
Accelerating the Particle-In-Cell Code ECsim with OpenACC
- Open Access
- 25 January 2026
- Published by Association for Computing Machinery (ACM)
- p. 122-132
Cross-architecture power efficiency analysis through micro-benchmarking
- 25 January 2026
- Published by Association for Computing Machinery (ACM)
- p. 122-132
Open AccessCross-architecture power efficiency analysis through micro-benchmarking
- Open Access
- 17 October 2025
- Published by IOP Publishing in Plasma Physics and Controlled Fusion
- Vol. 67 (10) , 105020
GPU porting of ASCOT5 code for Monte Carlo simulations in fusion plasmas
We present the graphics processing unit (GPU) porting of a Monte Carlo particle-following code, called ASCOT5. The code solves the distribution function of minority species in fusion plasmas. Originally developed with an MPI-OpenMP hybrid parallelism taking full advantage of single instruction, multiple data operations, the code has been ported to GPU architecture using the OpenACC programming model. Subsequently, modifications were made to implement three distinct algorithmic strategies: history-based, event-based, and event-based-packing. In the first implementation, each GPU processing unit deals with the entire history of one or more particles, while the event-based algorithm operates on the principle of executing a single low-level event type at a time for all particles still alive (with or without packing particles). Performance results on NVIDIA GPUs are presented to showcase the effectiveness and efficiency of the code adaptations for GPU execution. These results provide insights into the comparative performance of the implemented approaches on the specified hardware architecture. Portability across other architectures such as INTEL and AMD GPUs with OpenMP Offload is also presented.
- 17 October 2025
- Published by IOP Publishing in Plasma Physics and Controlled Fusion
- Vol. 67 (10) , 105020
Open AccessGPU porting of ASCOT5 code for Monte Carlo simulations in fusion plasmas
We present the graphics processing unit (GPU) porting of a Monte Carlo particle-following code, called ASCOT5. The code solves the distribution function of minority species in fusion plasmas. Originally developed with an MPI-OpenMP hybrid parallelism taking full advantage of single instruction, multiple data operations, the code has been ported to GPU architecture using the OpenACC programming model. Subsequently, modifications were made to implement three distinct algorithmic strategies: history-based, event-based, and event-based-packing. In the first implementation, each GPU processing unit deals with the entire history of one or more particles, while the event-based algorithm operates on the principle of executing a single low-level event type at a time for all particles still alive (with or without packing particles). Performance results on NVIDIA GPUs are presented to showcase the effectiveness and efficiency of the code adaptations for GPU execution. These results provide insights into the comparative performance of the implemented approaches on the specified hardware architecture. Portability across other architectures such as INTEL and AMD GPUs with OpenMP Offload is also presented.
- 25 February 2025
- Published by AIP Publishing in The Journal of Chemical Physics
- Vol. 162 (8)
Accelerated linear algebra for large scale DFT calculations of materials on CPU/GPU architectures with CRYSTAL
We discuss the implementation strategy, numerical accuracy, and computational performance of the acceleration of linear algebra operations through graphics processing units (GPUs) for the self-consistent field driver of the Crystal electronic structure package for solid state density functional theory simulations. Accelerated tasks include matrix multiplication, diagonalization, and inversion, as well as Cholesky decomposition. The scaling of the implemented strategy over multiple accelerating devices is assessed in the range of 1–8 GPUs per node and found to be remarkably regular. Tests are performed on three systems: α-quartz, a microporous zeolitic imidazolate framework (ZIF-8), and a giant mesoporous metal–organic framework (bio-MOF). Scaling with system size is investigated via supercells of increasing size of both α-quartz and ZIF-8 (up to 648 and 2208 atoms per cell, respectively). The bio-MOF model structure has 2808 atoms per cell, with 33 672 basis functions. We test the performance of the accelerated code with both generalized gradient approximation (GGA) and hybrid GGA exchange–correlation functionals. The efficiency of the new accelerated code is compared to the previous central processing unit (CPU)-only parallelization strategies based on MPI or MPI/OpenMP within either replicated or distributed memory (i.e., massively parallel) approaches. Such a comparison highlights how the new GPU-accelerated code enables calculations on large systems at a significantly reduced computational cost relative to CPU-only strategies. For instance, we find that for the bio-MOF system, the computing time of the linear algebra tasks from a single GPU is comparable to that from the reference approach in the range of 512–1024 CPU cores and 4–8 nodes.
- 25 February 2025
- Published by AIP Publishing in The Journal of Chemical Physics
- Vol. 162 (8)
Accelerated linear algebra for large scale DFT calculations of materials on CPU/GPU architectures with CRYSTAL
We discuss the implementation strategy, numerical accuracy, and computational performance of the acceleration of linear algebra operations through graphics processing units (GPUs) for the self-consistent field driver of the Crystal electronic structure package for solid state density functional theory simulations. Accelerated tasks include matrix multiplication, diagonalization, and inversion, as well as Cholesky decomposition. The scaling of the implemented strategy over multiple accelerating devices is assessed in the range of 1–8 GPUs per node and found to be remarkably regular. Tests are performed on three systems: α-quartz, a microporous zeolitic imidazolate framework (ZIF-8), and a giant mesoporous metal–organic framework (bio-MOF). Scaling with system size is investigated via supercells of increasing size of both α-quartz and ZIF-8 (up to 648 and 2208 atoms per cell, respectively). The bio-MOF model structure has 2808 atoms per cell, with 33 672 basis functions. We test the performance of the accelerated code with both generalized gradient approximation (GGA) and hybrid GGA exchange–correlation functionals. The efficiency of the new accelerated code is compared to the previous central processing unit (CPU)-only parallelization strategies based on MPI or MPI/OpenMP within either replicated or distributed memory (i.e., massively parallel) approaches. Such a comparison highlights how the new GPU-accelerated code enables calculations on large systems at a significantly reduced computational cost relative to CPU-only strategies. For instance, we find that for the bio-MOF system, the computing time of the linear algebra tasks from a single GPU is comparable to that from the reference approach in the range of 512–1024 CPU cores and 4–8 nodes.
- Open Access
- 19 February 2025
- Published by Association for Computing Machinery (ACM)
- p. 11-22
Preliminary Study on Fine-Grained Power and Energy Measurements on Grace Hopper GH200 with Open-Source Performance Tools
- 19 February 2025
- Published by Association for Computing Machinery (ACM)
- p. 11-22
Open AccessPreliminary Study on Fine-Grained Power and Energy Measurements on Grace Hopper GH200 with Open-Source Performance Tools
- Open Access
- 2 October 2024
- Published by Cambridge University Press (CUP) in Journal of Fluid Mechanics
- Vol. 996, A23
Effects of Rayleigh and Weber numbers on two-layer turbulent Rayleigh–Bénard convection
This study presents direct numerical simulation results of two-layer Rayleigh–Bénard convection, investigating the previously unexplored Rayleigh–Weber parameter space $10^6\leq Ra\leq 10^8$ and $10^2\leq We\leq 10^3$ . Global properties, such as the Nusselt and Reynolds numbers, are compared against the extended Grossmann–Lohse theory for two fluid layers, confirming a weak Weber number dependence for all global quantities and considerably larger Reynolds numbers in the lighter fluid. Statistics of the flow reveal that the interface fluctuates more intensely for larger Weber and smaller Rayleigh numbers, something also reflected in the increased temperature root mean square values next to the interface. The dynamics of the deformed two-fluid interface is further investigated using spectral analysis. Temporal and spatial spectrum distributions reveal a capillary wave range at small Weber and large Rayleigh numbers, and a secondary energy peak at smaller Rayleigh numbers. Furthermore, the maxima of the space–time spectra lie in an intermediate dispersion regime, between the theoretical predictions for capillary and gravity-capillary waves, showing that the gravitational energy of the interfacial waves is strongly altered by temperature gradients.- 2 October 2024
- Published by Cambridge University Press (CUP) in Journal of Fluid Mechanics
- Vol. 996, A23
Open AccessEffects of Rayleigh and Weber numbers on two-layer turbulent Rayleigh–Bénard convection
This study presents direct numerical simulation results of two-layer Rayleigh–Bénard convection, investigating the previously unexplored Rayleigh–Weber parameter space $10^6\leq Ra\leq 10^8$ and $10^2\leq We\leq 10^3$ . Global properties, such as the Nusselt and Reynolds numbers, are compared against the extended Grossmann–Lohse theory for two fluid layers, confirming a weak Weber number dependence for all global quantities and considerably larger Reynolds numbers in the lighter fluid. Statistics of the flow reveal that the interface fluctuates more intensely for larger Weber and smaller Rayleigh numbers, something also reflected in the increased temperature root mean square values next to the interface. The dynamics of the deformed two-fluid interface is further investigated using spectral analysis. Temporal and spatial spectrum distributions reveal a capillary wave range at small Weber and large Rayleigh numbers, and a secondary energy peak at smaller Rayleigh numbers. Furthermore, the maxima of the space–time spectra lie in an intermediate dispersion regime, between the theoretical predictions for capillary and gravity-capillary waves, showing that the gravitational energy of the interfacial waves is strongly altered by temperature gradients. - Open Access
- 11 January 2024
- Published by Association for Computing Machinery (ACM)
- p. 45-54
NVIDIA Grace Superchip Early Evaluation for HPC Applications
Arm-based system in HPC are a reality since more than a decade. However, when a new chip enters the market always implies challenges, not only at ISA level, but also with regards to the SoC integration, the memory subsystem, the board integration, the node interconnection, and finally the OS and all layers of the system software (compiler and libraries). Guided by the procurement of an NVIDIA Grace HPC cluster within the deployment of MareNostrum 5, and emulating the approach of a scientist who needs to migrate its scientific research to a new HPC system, we evaluated five complex scientific applications on engineering sample nodes of NVIDIA Grace CPU Superchip and NVIDIA Grace Hopper Superchip (CPU-only). We report intra-node and inter-node scalability and early performance results showing a speed-up between 1.3 × and 4.28 × for all codes when compared to the current generation of MareNostrum 4 powered by Intel Skylake CPUs.
- 11 January 2024
- Published by Association for Computing Machinery (ACM)
- p. 45-54
Open AccessNVIDIA Grace Superchip Early Evaluation for HPC Applications
Arm-based system in HPC are a reality since more than a decade. However, when a new chip enters the market always implies challenges, not only at ISA level, but also with regards to the SoC integration, the memory subsystem, the board integration, the node interconnection, and finally the OS and all layers of the system software (compiler and libraries). Guided by the procurement of an NVIDIA Grace HPC cluster within the deployment of MareNostrum 5, and emulating the approach of a scientist who needs to migrate its scientific research to a new HPC system, we evaluated five complex scientific applications on engineering sample nodes of NVIDIA Grace CPU Superchip and NVIDIA Grace Hopper Superchip (CPU-only). We report intra-node and inter-node scalability and early performance results showing a speed-up between 1.3 × and 4.28 × for all codes when compared to the current generation of MareNostrum 4 powered by Intel Skylake CPUs.
- Open Access
- 27 February 2023
- Published by Association for Computing Machinery (ACM)
- Vol. 2023, 35-49
Application Experiences on a GPU-Accelerated Arm-based HPC Testbed
This paper assesses and reports the experience of ten teams working to port, validate, and benchmark several High Performance Computing applications on a novel GPU-accelerated Arm testbed system. The testbed consists of eight NVIDIA Arm HPC Developer Kit systems, each one equipped with a server-class Arm CPU from Ampere Computing and two data center GPUs from NVIDIA Corp. The systems are connected together using InfiniBand interconnect. The selected applications and mini-apps are written using several programming languages and use multiple accelerator-based programming models for GPUs such as CUDA, OpenACC, and OpenMP offloading. Working on application porting requires a robust and easy-to-access programming environment, including a variety of compilers and optimized scientific libraries. The goal of this work is to evaluate platform readiness and assess the effort required from developers to deploy well-established scientific workloads on current and future generation Arm-based GPU-accelerated HPC systems. The reported case studies demonstrate that the current level of maturity and diversity of software and tools is already adequate for large-scale production deployments.
- 27 February 2023
- Published by Association for Computing Machinery (ACM)
- Vol. 2023, 35-49
Open AccessApplication Experiences on a GPU-Accelerated Arm-based HPC Testbed
This paper assesses and reports the experience of ten teams working to port, validate, and benchmark several High Performance Computing applications on a novel GPU-accelerated Arm testbed system. The testbed consists of eight NVIDIA Arm HPC Developer Kit systems, each one equipped with a server-class Arm CPU from Ampere Computing and two data center GPUs from NVIDIA Corp. The systems are connected together using InfiniBand interconnect. The selected applications and mini-apps are written using several programming languages and use multiple accelerator-based programming models for GPUs such as CUDA, OpenACC, and OpenMP offloading. Working on application porting requires a robust and easy-to-access programming environment, including a variety of compilers and optimized scientific libraries. The goal of this work is to evaluate platform readiness and assess the effort required from developers to deploy well-established scientific workloads on current and future generation Arm-based GPU-accelerated HPC systems. The reported case studies demonstrate that the current level of maturity and diversity of software and tools is already adequate for large-scale production deployments.
- 27 April 2022
- Published by Elsevier in Computer Physics Communications
- Vol. 277
LBcuda: A high-performance CUDA port of LBsoft for simulation of colloidal systems
- 27 April 2022
- Published by Elsevier in Computer Physics Communications
- Vol. 277
LBcuda: A high-performance CUDA port of LBsoft for simulation of colloidal systems
- Top Cited Papers
- 1 July 2015
- Published by Elsevier in Computer Physics Communications
- Vol. 192, 97-107
Strong scaling of general-purpose molecular dynamics simulations on GPUs
- 1 July 2015
- Published by Elsevier in Computer Physics Communications
- Vol. 192, 97-107
Top Cited PapersStrong scaling of general-purpose molecular dynamics simulations on GPUs