This paper proposes Gleam, a framework for efficient GPU sharing across local-area CUDA devices, reducing bandwidth overhead, improving API call latency, and ensuring context consistency.

This paper presents Festina, a profiling-guided, power-aware control plane for minimizing energy consumption in serverless large language model (LLM) serving.

This paper proposes CrossPool, a serving engine for cold Machine Learning Models (LLMs) that separates weights and KV-cache into two GPU memory pools to improve GPU memory utilization and long-context…

This paper identifies suboptimalities in default GPU-first scheduling for AI tool workloads and proposes an agentic scheduler that adaptively assigns tools to GPU or CPU based on runtime factors.

This paper proposes a decentralized, confidential computing platform using Intel TDX, ITA, and NVIDIA CC for secure and affordable AI workloads.

This paper introduces DGNA, a methodology to unveil the Non-Uniform Memory Access (NUMA) architecture of GPU memory hierarchy through microbenchmarking and data analysis.

This paper explores methods to reduce latency in GPU collective communications for large language model inference, achieving near-optimal designs with barrier-free synchronization and efficient use of…

The paper introduces Rotary GPU, an exploratory execution approach demonstrating that large Mixture-of-Experts models can be run locally on consumer GPUs with limited VRAM, achieving usable decode thr…

This paper demonstrates a novel GPU-side privilege escalation attack, showing that Rowhammer can be used to target and tamper with page tables to gain unauthorized access to co-tenant memory and ultim…

This paper compares and optimizes CUDA strategies for a shallow neural network, achieving a 1.41x speedup on a large dataset.

The paper proposes using hardware fingerprints instead of vulnerable cryptographic keys to enhance the security and robustness of GPU location verification for governing advanced AI development.

TileSight is a tile-centric performance-modeling tool that predicts single-GPU kernel latency and cache hit rates with low error, outperforming state-of-the-art baselines and transferring well across…

First implementation of a 3D Gaussian renderer on an Intelligence Processing Unit (IPU) with 1,472 independent tiles using only on-chip SRAM.

Hawkeye is a system that allows perfect, precision-preserving reproduction of GPU-level matrix multiplication operations on a CPU, enabling efficient and trustworthy third-party auditing of machine le…

The paper proposes n-VM, a novel Layer-1 architecture that unifies multiple heterogeneous virtual machines (VMs) onto a shared consensus and state layer, solving cross-chain fragmentation issues.

This paper presents four novel parallel algorithms for maintaining a rooted spanning forest in dynamic graphs, achieving a throughput of 2 million insertions and 1.4 million deletions per second.

This paper introduces SwarmX, a system for scheduling agentic AI applications in GPU-CPU clusters using neural predictors, reducing tail latency by up to 61.5% and sustaining up to 2x the throughput o…

The paper proposes a method to improve the performance of fine-grained offloads on servers by overlapping the offload with other requests using server-side routing.

This paper proposes FSZ, a GPU error-bounded lossy compressor with three innovations for higher compression ratios and throughput within a single CUDA kernel.

The paper introduces BSGS-Diagonal, a memory-efficient algorithm, and GPU-optimized kernels to significantly accelerate and reduce the resource overhead of encrypted face recognition using Fully Homom…