high-performance computing
PulseAugur coverage of high-performance computing — every cluster mentioning high-performance computing across labs, papers, and developer communities, ranked by signal.
- 2026-05-15 research_milestone Publication of a paper detailing a new heuristic-based method for merging HPC traces to extend hardware counter coverage for ML performance prediction. source
9 day(s) with sentiment data
-
Liquid cooling proves essential for next-gen AI compute density
Single-phase direct liquid cooling is emerging as a critical solution for managing the intense heat generated by modern AI and high-performance computing systems. As processors exceed 1,000 watts and racks dissipate ove…
-
AI infrastructure firm Cornelis raises $205M
Cornelis, a company specializing in AI infrastructure and high-performance computing network solutions, has secured $205 million in a funding round led by IAG Capital Partners. The company plans to utilize this investme…
-
New framework uses ML to speed up HPC performance autotuning
Researchers have developed a new framework to accelerate the process of performance autotuning in High Performance Computing (HPC) systems. This framework utilizes a machine learning-based ensemble LLVM Intermediate Rep…
-
Data centers boost density to meet AI compute demands
Data center operators are increasingly focusing on densifying existing facilities rather than building new ones to accommodate rising compute demands, particularly from AI and high-performance computing workloads. This …
-
AI and HPC converge via Stencil Computations with Triton and SYCL
This article discusses the intersection of AI and high-performance computing (HPC) through the lens of stencil computations. It highlights how techniques like vector outer products and the Triton programming model can b…
-
AI coding assistants fail to verify software trust signals, study finds
A new study investigated whether AI coding assistants verify software supply chain trust signals before installation. Researchers found that these assistants rarely checked for signals like software bills of materials o…
-
Diamond Agent streamlines HPC resource management across federated clusters
Researchers have developed Diamond Agent, a novel system designed to manage and orchestrate High-Performance Computing (HPC) resources across federated, heterogeneous clusters. This agentic system addresses challenges i…
-
AI drives new fire protection challenges in data centers
The increasing power density and new cooling technologies driven by AI are creating new fire protection challenges in data centers. Higher concentrations of computing power in smaller footprints, along with the adoption…
-
TornadoVM 6.0.0 launches JVMCI-free, JNI-free, and faster
TornadoVM has released version 6.0.0, a significant update that removes dependencies on JVMCI and JNI. This new version is compatible with JDK distributions from 21 to 27 and boasts a 37% reduction in cold start times. …
-
3C4AI project aims for seamless AI workload migration across cloud, edge, and HPC
The 3C4AI project is developing a vision for seamless AI workload migration across cloud, edge, and high-performance computing (HPC) environments. This initiative seeks to integrate unified orchestration, federated data…
-
AI collector interface architecture debated: speed vs. reliability
The author reflects on the architectural design of collector interfaces for AI systems, specifically focusing on the trade-offs between system speed and data reliability in observability. The posts explore concepts like…
-
HPC organizations urged to set AI agent usage policies now
Organizations utilizing high-performance computing (HPC) are urged to establish immediate internal policies regarding the use of AI and AI agents by their technical and administrative staff. These policies should addres…
-
Liquid cooling options emerge for AI-driven data center heat challenges
Data centers are increasingly adopting liquid cooling solutions to manage the intense heat generated by AI and high-performance computing workloads. These advanced cooling methods are necessary as rack densities have su…
-
HPC and AI workloads clash on shared clusters, solutions explored
Jeff Layton's exploration delves into the challenges of managing both high-performance computing (HPC) and AI workloads on shared clusters. The article discusses potential solutions for reconciling the distinct demands …
-
Agentic AI optimizes Python code for scientific simulations, saving millions
A new paper details an approach using Claude agentic AI to optimize Python code for scientific simulations, specifically in project scheduling. This method significantly reduced testing runtime from over 1,200 seconds t…
-
New research explores reprogramming open-weight LLMs for proactive behavior
Researchers have explored methods to reprogram the behavior of open-weight large language models, moving them away from passive assistant roles towards more proactive, Socratic interaction. Through extensive hyperparame…
-
Bitcoin miner BitFarms pivots data centers to AI and HPC workloads
BitFarms, a Bitcoin mining company, has transitioned its data centers to accommodate AI and high-performance computing workloads. The company has no current AI tenants and has ceased US Bitcoin mining operations, but ho…
-
New LLM research covers developer interaction, privacy, self-modeling, and HPC
Recent research explores various facets of Large Language Model (LLM) development and application. One study investigates dynamic LLM conversations for software development, finding that proactive guidance can increase …
-
Survey details methodologies for accelerating deep learning on heterogeneous architectures
This paper provides a comprehensive survey of methodologies and tools designed to accelerate deep learning on heterogeneous architectures. It covers hardware-software co-design, automated synthesis, domain-specific comp…
-
Microchip and Micron showcase PCIe Gen 6 storage architecture for AI and data centers
Microchip Technology and Micron Technology have collaborated to showcase an end-to-end PCIe Gen 6 storage architecture. This joint demonstration, featuring Microchip's Switchtec PCIe Gen 6 switch chip and Micron's 9650 …