-
6G EdgeAI: Performance Evaluation and Analysis
Authors:
Chien-Sheng Yang,
Yu-Jen Ku,
Yuan-Yao Lou,
Nathan Tenny,
Alex C. -C. Hsu
Abstract:
Generative AI (GenAI) services powered by large language models (LLMs) increasingly deliver real-time interactions, yet existing 5G multi-access edge computing (MEC) architectures often treat communication and computing as separate domains, limiting their ability to meet stringent latency requirements. To address this challenge, we introduce an Integrated Communication and Computing (ICC) framewor…
▽ More
Generative AI (GenAI) services powered by large language models (LLMs) increasingly deliver real-time interactions, yet existing 5G multi-access edge computing (MEC) architectures often treat communication and computing as separate domains, limiting their ability to meet stringent latency requirements. To address this challenge, we introduce an Integrated Communication and Computing (ICC) framework where computing capabilities are enabled to reside directly in radio access network (RAN) nodes and jointly manage bandwidth and computing resources. Our queueing-theoretic analysis shows that ICC outperforms 5G MEC, achieving higher service capacity (defined as the maximum arrival rate that maintains a specified fraction of jobs completed within a given delay budget) by 98%. We corroborate these gains through system-level simulations that account for transformer-based LLM workloads, realistic GPU specifications, and a priority-based scheduling scheme. The simulations show that ICC improves service capacity by 60%, demonstrating its potential to enable efficient, cost-effective real-time GenAI services in 6G.
△ Less
Submitted 23 April, 2025;
originally announced April 2025.
-
A finite-volume matrix Hamiltonian model for a Delta -> nucleon-pion system
Authors:
J. M. M. Hall,
A. C. -P. Hsu,
D. B. Leinweber,
A. W. Thomas,
R. D. Young
Abstract:
A matrix Hamiltonian model is developed to address the finite-volume effects appearing in studies of baryon resonances in lattice QCD. The Hamiltonian model includes interaction terms in a transparent way, and can be readily generalized to address multi-channel problems. The eigenvalue equation of the model is exactly solvable, and can be matched onto chiral effective field theory. The model is in…
▽ More
A matrix Hamiltonian model is developed to address the finite-volume effects appearing in studies of baryon resonances in lattice QCD. The Hamiltonian model includes interaction terms in a transparent way, and can be readily generalized to address multi-channel problems. The eigenvalue equation of the model is exactly solvable, and can be matched onto chiral effective field theory. The model is investigated in the case of Delta -> nucleon-pion scattering. A robust method for determining the resonance parameters from lattice QCD is developed. It involves constraining the free parameters of the model based on the lattice spectrum in question. The method is tested in the context of a set of pseudodata, and a picture of the model-dependence is obtained by examining a variety of regularization schemes in the model. A comparison is made with the Luescher method, and it is found that the matrix Hamiltonian method is equally robust. Both methods are tested in a more realistic scenario, where a background interaction corresponding to direct nucleon-pion scattering is incorporated into the pseudodata. The resulting extraction of the resonance parameters associated with the Delta baryon resonance provides evidence that an effective field theory style of approach yields a successful realization of finite-volume effects in the context of baryon resonances.
△ Less
Submitted 13 May, 2013; v1 submitted 18 March, 2013;
originally announced March 2013.
-
Baryon resonances and hadronic interactions in a finite volume
Authors:
J. M. M. Hall,
A. C. -P. Hsu,
D. B. Leinweber,
A. W. Thomas,
R. D. Young
Abstract:
In a finite volume, resonances and multi-hadron states are identified by discrete energy levels. When comparing the results of lattice QCD calculations to scattering experiments, it is important to have a way of associating the energy spectrum of the finite-volume lattice with the asymptotic behaviour of the S-matrix. A new technique for comparing energy eigenvalues with scattering phase shifts is…
▽ More
In a finite volume, resonances and multi-hadron states are identified by discrete energy levels. When comparing the results of lattice QCD calculations to scattering experiments, it is important to have a way of associating the energy spectrum of the finite-volume lattice with the asymptotic behaviour of the S-matrix. A new technique for comparing energy eigenvalues with scattering phase shifts is introduced, which involves the construction of an exactly solvable matrix Hamiltonian model. The model framework is applied to the case of $Δ\rightarrow Nπ$ decay, but is easily generalized to include multi-channel scattering. Extracting resonance parameters involves matching the energy spectrum of the model to that of a lattice QCD calculation. The resulting fit parameters are then used to generate phase shifts. Using a sample set of pseudodata, it is found that the extraction of the resonance position is stable with respect to volume for a variety of regularization schemes, and compares favorably with the well-known Luescher method. The model-dependence of the result is briefly investigated.
△ Less
Submitted 13 December, 2012; v1 submitted 15 July, 2012;
originally announced July 2012.
