5 citations · 6 across the 6 of their papers we have counts for
8 papers
E8 Lattice Key Quantization for Low-Bit KV-Cache Compression
Sasho Nedelkoski, Alexander Acker, Soeren Becker +2
Compressing the key-value (KV) cache is essential for serving long-context large language models (LLMs) under memory constraints. While current scalar quantization methods achieve…
Beyond Microservices: Testing Web-Scale RCA Methods on GPU-Driven LLM Workloads
Dominik Scheinert, Alexander Acker, Thorsten Wittkopp +6
Large language model (LLM) services have become an integral part of search, assistance, and decision-making applications. However, unlike traditional web or microservices, the hard…
Distributed LLM Pretraining During Renewable Curtailment Windows: A Feasibility Study
Philipp Wiesner, Soeren Becker, Brett Cornick +3
Training large language models (LLMs) requires substantial compute and energy. At the same time, renewable energy sources regularly produce more electricity than the grid can absor…
What happens when nanochat meets DiLoCo?
Alexander Acker, Soeren Becker, Sasho Nedelkoski +3
Although LLM training is typically centralized with high-bandwidth interconnects and large compute budgets, emerging methods target communication-constrained training in distribute…
Distributed Low-Communication Training with Decoupled Momentum Optimization
Sasho Nedelkoski, Alexander Acker, Odej Kao +2
The training of large models demands substantial computational resources, typically available only in data centers with high-bandwidth interconnects. However, reducing the reliance…
Predicting the Performance of Scientific Workflow Tasks for Cluster Resource Management: An Overview of the State of the Art
Jonathan Bader, Kathleen West, Soeren Becker +5
Scientific workflow management systems support large-scale data analysis on cluster infrastructures. For this, they interact with resource managers which schedule workflow tasks on…
WOW: Workflow-Aware Data Movement and Task Scheduling for Dynamic Scientific Workflows
Fabian Lehmann, Jonathan Bader, Friedrich Tschirpke +6
Scientific workflows process extensive data sets over clusters of independent nodes, which requires a complex stack of infrastructure components, especially a resource manager (RM)…
Towards a Peer-to-Peer Data Distribution Layer for Efficient and Collaborative Resource Optimization of Distributed Dataflow Applications
Dominik Scheinert, Soeren Becker, Jonathan Will +2
Performance modeling can help to improve the resource efficiency of clusters and distributed dataflow applications, yet the available modeling data is often limited. Collaborative…