Pasar al contenido principal

Resultados relacionados

Resultados totales (Incluyendo duplicados): 12
Encontrada(s) 1 página(s)

UPCommons. Portal del coneixement obert de la UPC
Publicación Contribución de congreso (ConferenceOutput)

Practically tackling memory bottlenecks of graph-processing workloads

  • Jamet, Alexandre Valentin
  • Vavouliotis, Georgios
  • Jiménez, Daniel A.
  • Álvarez Martí, Lluc|||0000-0003-0506-8867
  • Casas, Marc|||0000-0003-4564-2093
Graph-processing workloads have become widespread due to their relevance on a wide range of application domains such as network analysis, path- planning, bioinformatics, and machine learning. Graph-processing workloads have massive data footprints that exceed cache storage capacity and exhibit highly irregular memory access patterns due to data-dependent graph traversals. This irregular behaviour causes graph-processing workloads to exhibit poor data locality, undermining their performance. This paper makes two fundamental observations on the memory access patterns of graph-processing workloads: First, conventional cache hierarchies become mostly useless when dealing with graph-processing workloads, since 78.6% of the accesses that miss in the L1 Data Cache (L1D) result in misses in the L2 Cache (L2C) and in the Last Level Cache (LLC), requiring a DRAM access. Second, it is possible to predict whether a memory access will be served by DRAM or not in the context of graph-processing workloads by observing strides between accesses triggered by instructions with the same Program Counter (PC). Our key insight is that bypassing the L2C and the LLC for highly irregular accesses significantly reduces latency cost while also reducing pressure on the lower levels of the cache hierarchy. Based on these observations, this paper proposes the Large Predictor (LP), a low-cost micro-architectural predictor capable of distinguishing between regular and irregular memory accesses. We propose to serve accesses tagged as regular by LP via the standard memory hierarchy, while irregular access are served via the Side Data Cache (SDC). The SDC is a private per-core set-associative cache placed alongside the L1D specifically aimed at reducing the latency cost of highly irregular accesses while avoiding polluting the rest of the cache hierarchy with data that exhibits poor locality. SDC coupled with LP yields geometric mean speed-ups of 20.3% and 20.2% on single- and multi-core scenarios, respectively, over an architecture featuring a conventional cache hierarchy across a set of contemporary graph-processing workloads. In addition, SDC combined with LP outperforms the Transpose-based Cache Replacement (T-OPT), the state-of-the-art cache replacement policy for graph-processing applications, by 10.9% and 13.8% on single-core and multi-core contexts, respectively. Regarding the hardware budget, SDC coupled with LP requires 10KB of storage per core., This work has been partially supported by the European HiPEAC Network of Excellence, by the Spanish Ministry of Science and Innovation MCIN/AEI/10.13039/501100011033 (contracts PID2019- 107255GB-C21 and PID2019-105660RB-C22) and by the Generalitat de Catalunya (contract 2021-SGR-00763). This work is supported by the National Science Foundation through grant CCF-1912617 and generous gifts from Intel. Marc Casas has been partially supported by the Grant RYC-2017-23269 funded by MCIN/AEI/10.13039/501100011033 and by ESF Investing in your future. Els autors agraeixen el suport del Departament de Recerca i Universitats de la Generalitat de Catalunya al Grup de Recerca ”Performance understanding, analysis, and simulation/emulation of novel architectures” (Codi: 2021 SGR 00865). This research has received funding from the European High Performance Computing Joint Undertaking (JU) under Framework Partnership Agreement No 800928 (European Processor Initiative) and Specific Grant Agreement No 101036168 (EPI SGA2). The JU receives support from the European Union’s Horizon 2020 research and innovation programme and from Croatia, France, Germany, Greece, Italy, Netherlands, Portugal, Spain, Sweden, and Switzerland. The EPI-SGA2 project, PCI2022-132935 is also co-funded by MCIN/AEI /10.13039/501100011033 and by the UE NextGenerationEU/PRTR., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publicación Contribución de congreso (ConferenceOutput)

Batched DGEMMs for scientific codes running on long vector architectures

  • Banchelli Gracia, Fabio
  • Garcia Gasulla, Marta|||0000-0003-3682-9905
  • Mantovani, Filippo|||0000-0003-3559-4825
In this work, we evaluate the performance of SeisSol, a simulator of seismic wave phenomena and earthquake dynamics, on a RISCV-based system utilizing a vector processing unit. We focus on GEMM libraries and address their limited ability to leverage long vector architectures by developing a batched DGEMM library in plain C. This library achieves speedups ranging from approximately 3.5× to 32.6× compared to the reference implementation. We then integrate the batched approach into the SeisSol application, ensuring portability across different CPU architectures. Lastly, we demonstrate that our implementation is portable to an Intel CPU, resulting in improved execution times in most cases., Supported by the EuroHPC Joint Undertaking (JU): FPA N. 800928 (EPI), SGA N. 101036168 (EPI-SGA2), and GA N. 101093038 (ChEESE-2P CoE). The JU receives support from the EU Horizon 2020 research and innovation programme and from Croatia, France, Germany, Greece, Italy, Netherlands, Portugal, Spain, Sweden, Denmark and Switzerland. The EPI-SGA2 project, PCI2022-132935 is also co-funded by MCIN/AEI /10.13039/501100011033 and by the UE NextGenerationEU/PRTR. Supported by the pre-doctoral program AGAUR-FI ajuts (2024 FI200424) Joan Oró offered by Secretaria d’Universitats i Recerca del Departament de Recerca i Universitats de la Generalitat de Catalunya., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publicación Contribución de congreso (ConferenceOutput)

Reusable Verification Environment for a RISC-V Vector Accelerator

  • Quiroga, Josue
  • Genovese, Roberto Ignacio
  • Díaz, Ivan
  • Yano, Henrique
  • Ali, Asif|||0009-0005-6636-280X
  • Sommez, Nehir
  • Palomar Pérez, Óscar|||0000-0001-6729-4187
  • JImenez, Victor
  • Rodriguez, Mario
  • Dominguez, Marc
This paper presents a reusable verification environment developed for the verification of an academic RISC-V based vector accelerator that operates with long vectors. In order to be used across diverse projects, this infrastructure intends to be independent of the interface used for connecting the accelerator to the scalar processor core. We built a verification infrastructure consisting of a Universal Verification Environment (UVM) which is capable of validating the design performing co-simulation of the vector instructions. Moreover, we provided a set of tests and an automated test generation, simulation and error reporting infrastructure. This paper shares our experience on verifying a complex accelerator used in two distinct projects, with different interfaces., This research has received funding from the European High Performance Computing Joint Undertaking (JU) under Framework Partnership Agreement No 800928 (European Processor Initiative) and Specific Grant Agreement No 101036168 (EPI SGA2) and No 956702 (eProcessor) . The JU receives support from the European Union’s Horizon 2020 research and innovation programme and from Croatia, France, Germany, Greece, Italy, Netherlands, Portugal, Spain, Sweden, and Switzerland. The EPI-SGA2 project, PCI2022-132935_N1618737 is also co-funded by MCIN/AEI /10.13039/501100011033 and by the UE NextGenerationEU/PRTR, Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publicación Artículo científico (JournalArticle)

Vitruvius+: An area-efficient RISC-V decoupled vector coprocessor for high performance computing applications

  • Minervini Minervini, Francesco
  • Palomar Pérez, Óscar|||0000-0001-6729-4187
  • Unsal, Osman Sabri
  • Reggiani, Enrico
  • Quiroga Esparza, Josué Vladimir
  • Marimon Illana, Joan
  • Rojas Morales, Carlos|||0000-0002-7714-0277
  • Figueras Bagué, Roger
  • Ruíz Ramírez, Abraham Josafat
  • González Trejo, Alberto
  • Mendoza Escobar, Jonnatan
  • Vargas Valdivieso, Ivan|||0000-0002-5092-3829
  • Hernández Calderón, César Alejandro
  • Cabre Olive, Joan
  • Khoirunisya, Lina
  • Bouhali, Mustapha
  • Pavón Rivera, Julián
  • Moll Echeto, Francisco de Borja|||0000-0002-1290-3253
  • Olivieri, Mauro
  • Kovac, Mario
  • Kovac, Mate
  • Dragic, Leon
  • Valero Cortés, Mateo|||0000-0003-2917-2482
  • Cristal Kestelman, Adrián|||0000-0003-1277-9296
The maturity level of RISC-V and the availability of domain-specific instruction set extensions, like vector processing, make RISC-V a good candidate for supporting the integration of specialized hardware in processor cores for the High Performance Computing (HPC) application domain. In this article,1 we present Vitruvius+, the vector processing acceleration engine that represents the core of vector instruction execution in the HPC challenge that comes within the EuroHPC initiative. It implements the RISC-V vector extension (RVV) 0.7.1 and can be easily connected to a scalar core using the Open Vector Interface standard. Vitruvius+ natively supports long vectors: 256 double precision floating-point elements in a single vector register. It is composed of a set of identical vector pipelines (lanes), each containing a slice of the Vector Register File and functional units (one integer, one floating point). The vector instruction execution scheme is hybrid in-order/out-of-order and is supported by register renaming and arithmetic/memory instruction decoupling. On a stand-alone synthesis, Vitruvius+ reaches a maximum frequency of 1.4 GHz in typical conditions (TT/0.80V/25°C) using GlobalFoundries 22FDX FD-SOI. The silicon implementation has a total area of 1.3 mm2 and maximum estimated power of ~920 mW for one instance of Vitruvius+ equipped with eight vector lanes., This research has received funding from the European High Performance Computing Joint Undertaking (JU) under Framework Partnership Agreement No 800928 (European Processor Initiative) and Specific Grant Agreement No 101036168 (EPI SGA2). The JU receives support from the European Union’s Horizon 2020 research and innovation programme and from Croatia, France, Germany, Greece, Italy, Netherlands, Portugal, Spain, Sweden, and Switzerland. The EPI-SGA2 project, PCI2022-132935 is also co-funded by MCIN/AEI/10.13039/501100011033 and by the UE NextGen- erationEU/PRTR. This work has also been partially supported by the Spanish Ministry of Science and Innovation (PID2019-107255GB-C21/AEI/10.13039/501100011033)., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publicación Contribución de congreso (ConferenceOutput)

Software development vehicles to enable extended and early co-design: a RISC-V and HPC case of study

  • Mantovani, Filippo|||0000-0003-3559-4825
  • Vizcaíno Serrano, Pablo
  • Banchelli Gracia, Fabio
  • Garcia Gasulla, Marta|||0000-0003-3682-9905
  • Ferrer Ibañez, Roger
  • Ieronymakis, Georgios
  • Dimou, Nikolaus
  • Papaefstathiou, Vassilis
  • Labarta Mancho, Jesús José|||0000-0002-7489-4727
Prototyping HPC systems with low-to-mid technology readiness level (TRL) systems is critical for providing feedback to hardware designers, the system software team (e.g., compiler developers), and early adopters from the scientific community. The typical approach to hardware design and HPC system prototyping often limits feedback or only allows it at a late stage. In this paper, we present a set of tools for co-designing HPC systems, called software development vehicles (SDV). We use an innovative RISC-V design as a demonstrator, which includes a scalar CPU and a vector processing unit capable of operating large vectors up to 16 kbits. We provide an incremental methodology and early tangible evidence of the co-design process that provide feedback to improve both architecture and system software at a very early stage of system development., This research received funding from the EU-HPC-JU under FPA N. 800928 (EPI) and SGA N. 101036168 (EPI-SGA2). The JU receives support from some member countries. The EPI-SGA2 project, PCI2022-132935 is also co-funded by MCIN/AEI/10.13039/501100011033 and by the UE NextGenerationEU/PRTR., Peer Reviewed
Recercat. Dipósit de la Recerca de Catalunya
Publicación Contribución de congreso (ConferenceOutput)

Software development vehicles to enable extended and early co-design: a RISC-V and HPC case of study

  • Mantovani, Filippo
  • Vizcaíno Serrano, Pablo
  • Banchelli Gracia, Fabio
  • Garcia Gasulla, Marta
  • Ferrer Ibañez, Roger
  • Ieronymakis, Georgios
  • Dimou, Nikolaus
  • Papaefstathiou, Vassilis
  • Labarta Mancho, Jesús José
Prototyping HPC systems with low-to-mid technology readiness level (TRL) systems is critical for providing feedback to hardware designers, the system software team (e.g., compiler developers), and early adopters from the scientific community. The typical approach to hardware design and HPC system prototyping often limits feedback or only allows it at a late stage. In this paper, we present a set of tools for co-designing HPC systems, called software development vehicles (SDV). We use an innovative RISC-V design as a demonstrator, which includes a scalar CPU and a vector processing unit capable of operating large vectors up to 16 kbits. We provide an incremental methodology and early tangible evidence of the co-design process that provide feedback to improve both architecture and system software at a very early stage of system development., This research received funding from the EU-HPC-JU under FPA N. 800928 (EPI) and SGA N. 101036168 (EPI-SGA2). The JU receives support from some member countries. The EPI-SGA2 project, PCI2022-132935 is also co-funded by MCIN/AEI/10.13039/501100011033 and by the UE NextGenerationEU/PRTR., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Artículo científico (JournalArticle)

SpChar: characterizing the sparse puzzle via decision trees

  • Sgherzi, Francesco
  • Siracusa, Marco
  • Fernández Vega, Ivan
  • Armejach Sanosa, Adrià
  • Moretó Planas, Miquel
Sparse matrix computation is crucial in various modern applications, including large-scale graph analytics, deep learning, and recommender systems. The performance of sparse kernels varies greatly depending on the structure of the input matrix, making it difficult to gain a comprehensive understanding of sparse computation and its relationship to inputs, algorithms, and target machine architecture. Despite extensive research on certain sparse kernels, such as Sparse Matrix-Vector Multiplication (SpMV), the overall family of sparse algorithms has yet to be investigated as a whole. This paper introduces SpChar, a workload characterization methodology for general sparse computation. SpChar employs tree-based models to identify the most relevant hardware and input characteristics, starting from hardware and input-related metrics gathered from Performance Monitoring Counters (PMCs) and matrices. Our analysis enables the creation of a characterization loop that facilitates the optimization of sparse computation by mapping the impact of architectural features to inputs and algorithmic choices. We apply SpChar to more than 600 matrices from the SuiteSparse Matrix collection and three state-of-the-art Arm Central Processing Units (CPUs) to determine the critical hardware and software characteristics that affect sparse computation. In our analysis, we determine that the biggest limiting factors for high-performance sparse computation are (1) the latency of the memory system, (2) the pipeline flush overhead resulting from branch misprediction, and (3) the poor reuse of cached elements. Additionally, we propose software and hardware optimizations that designers can implement to create a platform suitable for sparse computation. We then investigate these optimizations using the gem5 simulator to achieve a significant speedup of up to 2.63× compared to a CPU where the optimizations are not applied., This research was supported by: the Spanish Ministry of Science and Innovation (MCIN) through contracts PID2019-107255GB-C21, PID2019-107255GBC21 (MCIN/AEI/10.13039/501100011033) and the European Processor Initiative (EPI) under grant agreement No. 101036168. M. Siracusa has been partially supported by an FI fellowship (2022FI B 00969). Finally, A. Armejach is a Serra Hunter Fellow, Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Contribución de congreso (ConferenceOutput)

Batched DGEMMs for scientific codes running on long vector architectures

  • Banchelli Gracia, Fabio
  • Garcia Gasulla, Marta
  • Mantovani, Filippo
In this work, we evaluate the performance of SeisSol, a simulator of seismic wave phenomena and earthquake dynamics, on a RISCV-based system utilizing a vector processing unit. We focus on GEMM libraries and address their limited ability to leverage long vector architectures by developing a batched DGEMM library in plain C. This library achieves speedups ranging from approximately 3.5× to 32.6× compared to the reference implementation. We then integrate the batched approach into the SeisSol application, ensuring portability across different CPU architectures. Lastly, we demonstrate that our implementation is portable to an Intel CPU, resulting in improved execution times in most cases., Supported by the EuroHPC Joint Undertaking (JU): FPA N. 800928 (EPI), SGA N. 101036168 (EPI-SGA2), and GA N. 101093038 (ChEESE-2P CoE). The JU receives support from the EU Horizon 2020 research and innovation programme and from Croatia, France, Germany, Greece, Italy, Netherlands, Portugal, Spain, Sweden, Denmark and Switzerland. The EPI-SGA2 project, PCI2022-132935 is also co-funded by MCIN/AEI /10.13039/501100011033 and by the UE NextGenerationEU/PRTR. Supported by the pre-doctoral program AGAUR-FI ajuts (2024 FI200424) Joan Oró offered by Secretaria d’Universitats i Recerca del Departament de Recerca i Universitats de la Generalitat de Catalunya., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Contribución de congreso (ConferenceOutput)

Practically tackling memory bottlenecks of graph-processing workloads

  • Jamet, Alexandre Valentin
  • Vavouliotis, Georgios
  • Jiménez, Daniel A.
  • Álvarez Martí, Lluc
  • Casas, Marc
Graph-processing workloads have become widespread due to their relevance on a wide range of application domains such as network analysis, path- planning, bioinformatics, and machine learning. Graph-processing workloads have massive data footprints that exceed cache storage capacity and exhibit highly irregular memory access patterns due to data-dependent graph traversals. This irregular behaviour causes graph-processing workloads to exhibit poor data locality, undermining their performance. This paper makes two fundamental observations on the memory access patterns of graph-processing workloads: First, conventional cache hierarchies become mostly useless when dealing with graph-processing workloads, since 78.6% of the accesses that miss in the L1 Data Cache (L1D) result in misses in the L2 Cache (L2C) and in the Last Level Cache (LLC), requiring a DRAM access. Second, it is possible to predict whether a memory access will be served by DRAM or not in the context of graph-processing workloads by observing strides between accesses triggered by instructions with the same Program Counter (PC). Our key insight is that bypassing the L2C and the LLC for highly irregular accesses significantly reduces latency cost while also reducing pressure on the lower levels of the cache hierarchy. Based on these observations, this paper proposes the Large Predictor (LP), a low-cost micro-architectural predictor capable of distinguishing between regular and irregular memory accesses. We propose to serve accesses tagged as regular by LP via the standard memory hierarchy, while irregular access are served via the Side Data Cache (SDC). The SDC is a private per-core set-associative cache placed alongside the L1D specifically aimed at reducing the latency cost of highly irregular accesses while avoiding polluting the rest of the cache hierarchy with data that exhibits poor locality. SDC coupled with LP yields geometric mean speed-ups of 20.3% and 20.2% on single- and multi-core scenarios, respectively, over an architecture featuring a conventional cache hierarchy across a set of contemporary graph-processing workloads. In addition, SDC combined with LP outperforms the Transpose-based Cache Replacement (T-OPT), the state-of-the-art cache replacement policy for graph-processing applications, by 10.9% and 13.8% on single-core and multi-core contexts, respectively. Regarding the hardware budget, SDC coupled with LP requires 10KB of storage per core., This work has been partially supported by the European HiPEAC Network of Excellence, by the Spanish Ministry of Science and Innovation MCIN/AEI/10.13039/501100011033 (contracts PID2019- 107255GB-C21 and PID2019-105660RB-C22) and by the Generalitat de Catalunya (contract 2021-SGR-00763). This work is supported by the National Science Foundation through grant CCF-1912617 and generous gifts from Intel. Marc Casas has been partially supported by the Grant RYC-2017-23269 funded by MCIN/AEI/10.13039/501100011033 and by ESF Investing in your future. Els autors agraeixen el suport del Departament de Recerca i Universitats de la Generalitat de Catalunya al Grup de Recerca ”Performance understanding, analysis, and simulation/emulation of novel architectures” (Codi: 2021 SGR 00865). This research has received funding from the European High Performance Computing Joint Undertaking (JU) under Framework Partnership Agreement No 800928 (European Processor Initiative) and Specific Grant Agreement No 101036168 (EPI SGA2). The JU receives support from the European Union’s Horizon 2020 research and innovation programme and from Croatia, France, Germany, Greece, Italy, Netherlands, Portugal, Spain, Sweden, and Switzerland. The EPI-SGA2 project, PCI2022-132935 is also co-funded by MCIN/AEI /10.13039/501100011033 and by the UE NextGenerationEU/PRTR., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Artículo científico (JournalArticle)

Vitruvius+: An area-efficient RISC-V decoupled vector coprocessor for high performance computing applications

  • Minervini Minervini, Francesco
  • Palomar Pérez, Óscar
  • Unsal, Osman Sabri
  • Reggiani, Enrico
  • Quiroga Esparza, Josué Vladimir
  • Marimon Illana, Joan
  • Rojas Morales, Carlos
  • Figueras Bagué, Roger
  • Ruíz Ramírez, Abraham Josafat
  • González Trejo, Alberto
  • Mendoza Escobar, Jonnatan
  • Vargas Valdivieso, Ivan
  • Hernández Calderón, César Alejandro
  • Cabre Olive, Joan
  • Khoirunisya, Lina
  • Bouhali, Mustapha
  • Pavón Rivera, Julián
  • Moll Echeto, Francisco de Borja
  • Olivieri, Mauro
  • Kovac, Mario
  • Kovac, Mate
  • Dragic, Leon
  • Valero Cortés, Mateo
  • Cristal Kestelman, Adrián
The maturity level of RISC-V and the availability of domain-specific instruction set extensions, like vector processing, make RISC-V a good candidate for supporting the integration of specialized hardware in processor cores for the High Performance Computing (HPC) application domain. In this article,1 we present Vitruvius+, the vector processing acceleration engine that represents the core of vector instruction execution in the HPC challenge that comes within the EuroHPC initiative. It implements the RISC-V vector extension (RVV) 0.7.1 and can be easily connected to a scalar core using the Open Vector Interface standard. Vitruvius+ natively supports long vectors: 256 double precision floating-point elements in a single vector register. It is composed of a set of identical vector pipelines (lanes), each containing a slice of the Vector Register File and functional units (one integer, one floating point). The vector instruction execution scheme is hybrid in-order/out-of-order and is supported by register renaming and arithmetic/memory instruction decoupling. On a stand-alone synthesis, Vitruvius+ reaches a maximum frequency of 1.4 GHz in typical conditions (TT/0.80V/25°C) using GlobalFoundries 22FDX FD-SOI. The silicon implementation has a total area of 1.3 mm2 and maximum estimated power of ~920 mW for one instance of Vitruvius+ equipped with eight vector lanes., This research has received funding from the European High Performance Computing Joint Undertaking (JU) under Framework Partnership Agreement No 800928 (European Processor Initiative) and Specific Grant Agreement No 101036168 (EPI SGA2). The JU receives support from the European Union’s Horizon 2020 research and innovation programme and from Croatia, France, Germany, Greece, Italy, Netherlands, Portugal, Spain, Sweden, and Switzerland. The EPI-SGA2 project, PCI2022-132935 is also co-funded by MCIN/AEI/10.13039/501100011033 and by the UE NextGen- erationEU/PRTR. This work has also been partially supported by the Spanish Ministry of Science and Innovation (PID2019-107255GB-C21/AEI/10.13039/501100011033)., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Contribución de congreso (ConferenceOutput)

Reusable Verification Environment for a RISC-V Vector Accelerator

  • Quiroga, Josue
  • Genovese, Roberto Ignacio
  • Díaz, Ivan
  • Yano, Henrique
  • Ali, Asif
  • Sommez, Nehir
  • Palomar Pérez, Óscar
  • JImenez, Victor
  • Rodriguez, Mario
  • Dominguez, Marc
This paper presents a reusable verification environment developed for the verification of an academic RISC-V based vector accelerator that operates with long vectors. In order to be used across diverse projects, this infrastructure intends to be independent of the interface used for connecting the accelerator to the scalar processor core. We built a verification infrastructure consisting of a Universal Verification Environment (UVM) which is capable of validating the design performing co-simulation of the vector instructions. Moreover, we provided a set of tests and an automated test generation, simulation and error reporting infrastructure. This paper shares our experience on verifying a complex accelerator used in two distinct projects, with different interfaces., This research has received funding from the European High Performance Computing Joint Undertaking (JU) under Framework Partnership Agreement No 800928 (European Processor Initiative) and Specific Grant Agreement No 101036168 (EPI SGA2) and No 956702 (eProcessor) . The JU receives support from the European Union’s Horizon 2020 research and innovation programme and from Croatia, France, Germany, Greece, Italy, Netherlands, Portugal, Spain, Sweden, and Switzerland. The EPI-SGA2 project, PCI2022-132935_N1618737 is also co-funded by MCIN/AEI /10.13039/501100011033 and by the UE NextGenerationEU/PRTR, Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Contribución de congreso (ConferenceOutput)

NVIDIA Grace superchip early evaluation for HPC applications

  • Banchelli Gracia, Fabio
  • Vinyals Ylla-Català, Joan
  • Pocurull Serra, Josep
  • Clascà Ramírez, Marc
  • Peiró Conde, Kilian
  • Spiga, Filippo
  • Garcia Gasulla, Marta
  • Mantovani, Filippo
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., This research has received funding from the European Commission via the Horizon Europe research and innovation funding programme, under grant agreement 101092993 (RISER), the EuroHPC H2020 Industrial Leadership programme grant agreement 956416 (exaFOAM), and the European High Performance Computing Joint Undertaking (JU) under Framework Partnership Agreement No 800928 (European Processor Initiative) and Specific Grant Agreement No 101036168 (EPI SGA2). The JU receives support from the European Union’s Horizon 2020 research and innovation programme and from Croatia, France, Germany, Greece, Italy, Netherlands, Portugal, Spain, Sweden, and Switzerland. The EPISGA2 project, PCI2022-132935 is also co-funded by MCIN/AEI /10.13039/501100011033 and by the UE NextGenerationEU/PRTR., Peer Reviewed, Postprint (author's final draft)