Allen: A High-Level Trigger on GPUs for LHCb

Aaij, R.; Albrecht, J.; Belous, Mikhail; Billoir, P.; Boettcher, T.; Rodríguez, A. Brea; Bruch, Dorothea Vom; Pérez, Daniel Hugo Cámpora; Vidal, A. Casais; Craik, D. C.; Declara, P. Fernandez; Funke, L.; Gligorov, Vladimir; Jashal, B. K.; Kazeev, N.; Santos, D. Martínez; Pisani, F.; Pliushchenko, D.; Popov, S.; Quagliani, R.; Rangel, M. S.; Reiss, F.; Mayordomo, C. Sánchez; Schwemmer, R.; Sokoloff, M. D.; Stevens, H.; Ustyuzhanin, A.; Cardona, Xavier Vilasís; Williams, M.

doi:10.1007/s41781-020-00039-7

Cited by 59 publications

(55 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It also carries out a partial reconstruction of charged tracks and muon identification. HLT1 selections are based on the properties of single tracks and two-track composite objects and permit a reduction of the event rate by around a factor 30, depending on the tuning of the selection cuts [4].…”

Section: The Lhcb Trigger Upgradementioning

confidence: 99%

Flexible selections at 30 MHz in the LHCb Upgrade trigger

Daussy-Renaudin¹

2021

Proceedings of 40th International Conference on High Energy Physics — PoS(ICHEP2020)

View full text Add to dashboard Cite

The first LHCb upgrade will take data at an instantaneous luminosity of 2 × 10 33 cm −2 s −1 starting in 2022. Due to the high rate of beauty and charm signals LHCb has chosen as its baseline to read out the entire detector into a software trigger running at the LHC collision frequency of 30 MHz. This High Level Trigger (HLT) will enable unprecedented flexibility for trigger selections. To provide a flexible trigger, the key features of the trigger selection framework are discussed here as well as the current performance.

show abstract

Section: The Lhcb Trigger Upgradementioning

confidence: 99%

Flexible selections at 30 MHz in the LHCb Upgrade trigger

Daussy-Renaudin¹

2021

Proceedings of 40th International Conference on High Energy Physics — PoS(ICHEP2020)

View full text Add to dashboard Cite

show abstract

“…Allen's framework design allows the chain of algorithms that computes the full HLT1 to compile for both GPU and CPU architectures [2]. Allen is designed and optimized for GPUs; compilation for CPUs in Allen supports basic multithreading and does not vectorize the algorithms.…”

Section: Spmd and Ispcmentioning

confidence: 99%

“…The throughput achieved with different GPUs is highlighted for comparison: the GPUs achieve higher throughput for the Compass implementation. 2…”

Section: Spmd Compass Trackingmentioning

confidence: 99%

Compass SPMD: a SPMD vectorized tracking algorithm

Declara

García

2020

EPJ Web Conf.

View full text Add to dashboard Cite

Compass is a SPMD (Single Program Multiple Data) tracking algorithm for the upcoming LHCb upgrade in 2021. 40 Tb/s need to be processed in real-time to select events. Alternative frameworks, algorithms and architectures are being tested to cope with the deluge of data. Allen is a research and development project aiming to run the full HLT1 (High Level Trigger) on GPUs (Graphics Processing Units). Allen’s architecture focuses on data-oriented layout and algorithms to better exploit parallel architectures. GPUs already proved to exploit the framework efficiently with the algorithms developed for Allen, implemented and optimized for GPU architectures. We explore opportunities for the SIMD (Single Instruction Multiple Data) paradigm in CPUs through the Compass algorithm. We use the Intel SPMD Program Compiler (ISPC) to achieve good readability, maintainability and performance writing “GPU-like” source code, preserving the main design of the algorithm.

show abstract

“…The ALICE experiment at the LHC is planning to use GPUs for real-time processing and data compression of their Time Project Chamber subdetectors ( ALICE Collaboration and Rohr, 2020 ). The LHCb experiment at the LHC is considering using GPUs for the first level of their trigger system ( Aaij et al, 2020 ).…”

Section: Introductionmentioning

confidence: 99%

GPU-Accelerated Machine Learning Inference as a Service for Computing in Neutrino Experiments

et al. 2021

View full text Add to dashboard Cite

Machine learning algorithms are becoming increasingly prevalent and performant in the reconstruction of events in accelerator-based neutrino experiments. These sophisticated algorithms can be computationally expensive. At the same time, the data volumes of such experiments are rapidly increasing. The demand to process billions of neutrino events with many machine learning algorithm inferences creates a computing challenge. We explore a computing model in which heterogeneous computing with GPU coprocessors is made available as a web service. The coprocessors can be efficiently and elastically deployed to provide the right amount of computing for a given processing task. With our approach, Services for Optimized Network Inference on Coprocessors (SONIC), we integrate GPU acceleration specifically for the ProtoDUNE-SP reconstruction chain without disrupting the native computing workflow. With our integrated framework, we accelerate the most time-consuming task, track and particle shower hit identification, by a factor of 17. This results in a factor of 2.7 reduction in the total processing time when compared with CPU-only production. For this particular task, only 1 GPU is required for every 68 CPU threads, providing a cost-effective solution.

show abstract

Allen: A High-Level Trigger on GPUs for LHCb

Cited by 59 publications

References 15 publications

Flexible selections at 30 MHz in the LHCb Upgrade trigger

Flexible selections at 30 MHz in the LHCb Upgrade trigger

Compass SPMD: a SPMD vectorized tracking algorithm

GPU-Accelerated Machine Learning Inference as a Service for Computing in Neutrino Experiments

Contact Info

Product

Resources

About