arXiv Science⌕ Search

arXiv subjects

Thomas Green

Publications and source records attributed to Thomas Green.

5 recordsLinked to original sources

NVIDIA Vera: A Microarchitecture-Level Comparison with Grace, Turin, and Granite Rapids

The majority of new HPC systems are centered around NVIDIA, AMD, or Intel CPUs. Vera is NVIDIA's latest data-center CPU, including its first custom data-center core design, Olympus. This paper investigates Vera in comparison to AMD Turin and Intel Granite Rapids, using NVIDIA Grace as a baseline. We navigate up the performance ladder, from microbenchmarks that isolate performance metrics, through NAS kernels and miniapps, to single-socket performance of a range of scientific codes. We find that Vera achieves the best performance of the architectures evaluated in most codes tested, and that this gain is primarily attributed to its high memory bandwidth, effective Spatial Multithreading implementation, and low-latency memory subsystem aided by the architecture's predictive structures. While Vera represents a major leap over its predecessor, we show that its comparatively modest floating-point performance trails Intel and AMD in a small number of compute-bound workloads reflecting NVIDIA's GPU focus for such workloads.

cs.DC↗

Multi-tenant Kubernetes Use Cases for AI, Secure Computing and Data Services, and More

Kubernetes, as a container orchestration engine, has been widely used in cloud-native ecosystems for several years. In supercomputing ecosystems, especially where bare-metal performance for compute and network devices are considered, the adoption is somewhat limited. However, with the increasing diversity of use cases such as AI, secure and confidential computing for sensitive data, and mixed workload orchestration, a traditional, single-tenant batch computing system does not offer the flexibility and reproducibility to which public cloud users are accustomed. Note that Kubernetes is not considered a replacement for batch scheduling systems, which have powerful features for large-scale MPI jobs with thousands of network end points. Rather, it is a complementary service provided as part of a national AI Research Resource. We evaluate Kubernetes deployment on a Hewlett Packard Enterprise (HPE) Cray EX supercomputerwith HPE Slingshot interconnect, called Isambard-AI, with co-design use cases. One is a Trusted Research Environment used for medical and health sciences. The other combines KubeRay, Ray, and vLLM to provide a distributed, sandboxed, persistent AI model hosting service targeting multi-tenant confidential computing. We discuss challenges and lessons learned, and where further development is needed to offer a production Kubernetes-as-a-Service on HPE Cray EX (and later) platforms.

cs.DC↗

Particle Builder A Board Game for the Teaching of the Standard Model of Particle Physics at a Secondary Level

We present Particle Builder, an online board game which teaches students about concepts from the Standard Model of Particle Physics at a high school level. This short activity resulted in a gain of 0.16, indicating that students learned a significant amount of particle physics knowledge. Students found the activity was more engaging and less difficult than a normal classroom lesson.

physics.ed-ph↗

Particle Builder -- Learn about the Standard Model while playing against an AI

Particle Builder Online is a web-based education game designed for high school physics students. Students can play against an AI opponent or peers to familiarise themselves with the Standard Model of Particle Physics. The game is aimed at a high school level and tailored to the International Baccalaureate and the Australian Curriculum. Students from four schools in Canberra took pre/post-tests and a survey while completing a lesson where they played Particle Builder. Students' understanding of particle physics concepts improved significantly. Students found the game more enjoyable and effective than regular classroom lessons.

physics.ed-ph↗

Perovskite-perovskite tandem photovoltaics with optimized bandgaps

We demonstrate four and two-terminal perovskite-perovskite tandem solar cells with ideally matched bandgaps. We develop an infrared absorbing 1.2eV bandgap perovskite, $FA_{0.75}Cs_{0.25}Sn_{0.5}Pb_{0.5}I_3$, that can deliver 14.8 % efficiency. By combining this material with a wider bandgap $FA_{0.83}Cs_{0.17}Pb(I_{0.5}Br_{0.5})_3$ material, we reach monolithic two terminal tandem efficiencies of 17.0 % with over 1.65 volts open-circuit voltage. We also make mechanically stacked four terminal tandem cells and obtain 20.3 % efficiency. Crucially, we find that our infrared absorbing perovskite cells exhibit excellent thermal and atmospheric stability, unprecedented for Sn based perovskites. This device architecture and materials set will enable 'all perovskite' thin film solar cells to reach the highest efficiencies in the long term at the lowest costs.

cond-mat.mtrl-sci↗