Top-Down Models across CPU Architectures: Applicability and Comparison in a High-Performance Computing Environment
Top-Down models are defined by hardware architects to provide information on the utilization of different hardware components. The target is to isolate the users from the complexity of the hardware architecture while giving them insight into how efficiently the code uses the resources. In this paper...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-10-01
|
| Series: | Information |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2078-2489/14/10/554 |
| _version_ | 1797573554474057728 |
|---|---|
| author | Fabio Banchelli Marta Garcia-Gasulla Filippo Mantovani |
| author_facet | Fabio Banchelli Marta Garcia-Gasulla Filippo Mantovani |
| author_sort | Fabio Banchelli |
| collection | DOAJ |
| description | Top-Down models are defined by hardware architects to provide information on the utilization of different hardware components. The target is to isolate the users from the complexity of the hardware architecture while giving them insight into how efficiently the code uses the resources. In this paper, we explore the applicability of four Top-Down models defined for different hardware architectures powering state-of-the-art HPC clusters (Intel Skylake, Fujitsu A64FX, IBM Power9, and Huawei Kunpeng 920) and propose a model for AMD Zen 2. We study a parallel CFD code used for scientific production to compare these five Top-Down models. We evaluate the level of insight achieved, the clarity of the information, the ease of use, and the conclusions each allows us to reach. Our study indicates that the Top-Down model makes it very difficult for a performance analyst to spot inefficiencies in complex scientific codes without delving deep into micro-architecture details. |
| first_indexed | 2024-03-10T21:10:43Z |
| format | Article |
| id | doaj.art-55f41feb566043159e01811bb2f9a33b |
| institution | Directory Open Access Journal |
| issn | 2078-2489 |
| language | English |
| last_indexed | 2024-03-10T21:10:43Z |
| publishDate | 2023-10-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Information |
| spelling | doaj.art-55f41feb566043159e01811bb2f9a33b2023-11-19T16:48:06ZengMDPI AGInformation2078-24892023-10-01141055410.3390/info14100554Top-Down Models across CPU Architectures: Applicability and Comparison in a High-Performance Computing EnvironmentFabio Banchelli0Marta Garcia-Gasulla1Filippo Mantovani2Barcelona Supercomputing Center, Plaça Eusebi Güell, 1-3, 08034 Barcelona, SpainBarcelona Supercomputing Center, Plaça Eusebi Güell, 1-3, 08034 Barcelona, SpainBarcelona Supercomputing Center, Plaça Eusebi Güell, 1-3, 08034 Barcelona, SpainTop-Down models are defined by hardware architects to provide information on the utilization of different hardware components. The target is to isolate the users from the complexity of the hardware architecture while giving them insight into how efficiently the code uses the resources. In this paper, we explore the applicability of four Top-Down models defined for different hardware architectures powering state-of-the-art HPC clusters (Intel Skylake, Fujitsu A64FX, IBM Power9, and Huawei Kunpeng 920) and propose a model for AMD Zen 2. We study a parallel CFD code used for scientific production to compare these five Top-Down models. We evaluate the level of insight achieved, the clarity of the information, the ease of use, and the conclusions each allows us to reach. Our study indicates that the Top-Down model makes it very difficult for a performance analyst to spot inefficiencies in complex scientific codes without delving deep into micro-architecture details.https://www.mdpi.com/2078-2489/14/10/554performance modelstop-down modelHPC applicationsMareNostrum 4A64FXPower 9 |
| spellingShingle | Fabio Banchelli Marta Garcia-Gasulla Filippo Mantovani Top-Down Models across CPU Architectures: Applicability and Comparison in a High-Performance Computing Environment Information performance models top-down model HPC applications MareNostrum 4 A64FX Power 9 |
| title | Top-Down Models across CPU Architectures: Applicability and Comparison in a High-Performance Computing Environment |
| title_full | Top-Down Models across CPU Architectures: Applicability and Comparison in a High-Performance Computing Environment |
| title_fullStr | Top-Down Models across CPU Architectures: Applicability and Comparison in a High-Performance Computing Environment |
| title_full_unstemmed | Top-Down Models across CPU Architectures: Applicability and Comparison in a High-Performance Computing Environment |
| title_short | Top-Down Models across CPU Architectures: Applicability and Comparison in a High-Performance Computing Environment |
| title_sort | top down models across cpu architectures applicability and comparison in a high performance computing environment |
| topic | performance models top-down model HPC applications MareNostrum 4 A64FX Power 9 |
| url | https://www.mdpi.com/2078-2489/14/10/554 |
| work_keys_str_mv | AT fabiobanchelli topdownmodelsacrosscpuarchitecturesapplicabilityandcomparisoninahighperformancecomputingenvironment AT martagarciagasulla topdownmodelsacrosscpuarchitecturesapplicabilityandcomparisoninahighperformancecomputingenvironment AT filippomantovani topdownmodelsacrosscpuarchitecturesapplicabilityandcomparisoninahighperformancecomputingenvironment |