HoliBench is a modular benchmarking and deployment toolkit that jointly measures accuracy, latency, and energy for foundation models across a wide range of devices, from single-board computers to GPU servers. It provides a platform abstraction layer that calibrates cross-device measurements and supports multiple model modalities, inference engines, and quantization levels. Using HoliBench, the authors evaluated 20 models on 7 device types, revealing tradeoffs such as limited latency gains from quantization on low‑precision hardware and diminishing accuracy returns relative to energy consumption, while also showing that single-model profiles can predict multi-model pipeline performance within a few percent.
arXiv:2606. 28565v1 Announce Type: cross Abstract: As large language models (LLMs) move into production serving, practitioners must rapidly evaluate inference performance across diverse hardware, models, and serving parameters to meet cost and latency targets.
arXiv:2607. 20806v1 Announce Type: new Abstract: Lightweight large language models (LLMs) are increasingly being deployed locally on personal computers and are expected to play a growing role in resource-constrained edge and mobile environments.
arXiv:2609.08307v1 Announce Type: cross Abstract: Large language models (LLMs) are increasingly used as backends for intelligent web services, but serving them across the edge continuum requires bala...
The paper argues that AI deployment performance depends on interactions among compression, compiler transformations, and serving policies rather than just model architecture. It introduces a three‑layer taxonomy—model‑level techniques, compiler transformations, and system policies—and frames deployment as a constrained multi‑objective optimization problem over accuracy, latency, throughput, memory footprint, and energy. The authors propose an evidence protocol for comparable benchmarking and synthesize data from edge and data‑center platforms to show that cross‑layer interactions drive deployment outcomes, concluding with a constraint‑aware selection procedure and open research problems.
Dooly is a profiling tool for large language model (LLM) inference that eliminates the need to re‑profile every operation for each configuration. By labeling input dimensions with taint propagation and only profiling operations missing from its latency database, Dooly builds a configuration‑agnostic, redundancy‑aware latency database that can be used as a drop‑in backend for simulators. In experiments across two GPU platforms, three attention backends, and multiple model architectures, Dooly achieves simulation accuracy within 5% MAPE for time‑to‑first‑token and 8% for total‑processing‑time while cutting profiling GPU‑hours by 56.4% compared to traditional methods.
Lightweight large language models (LLMs) are increasingly being deployed locally on personal computers and are expected to play a growing role in resource-constrained edge and mobile environments. In such settings, energy consumption, execution time, and memory usage directly affect practical usability, yet existing evaluations of LLM efficiency largely rely on proxy descriptors such as parameter count or FLOPs, often decoupled from task precision.
arXiv:2606. 02963v1 Announce Type: new Abstract: Production inference increasingly targets a heterogeneous mix of accelerators.
arXiv:2606. 04023v1 Announce Type: cross Abstract: While large language models (LLMs) have been extensively evaluated on code generation tasks for general-purpose programming and GPU-accelerated environments (e.
DynaFlow is a framework that enables transparent and flexible intra‑device parallelism by separating logical model definitions from physical execution schedules. It offers a flexible frontend with annotations for graph partitioning and a programmable interface for custom parallelism strategies, while its efficient backend handles asynchronous control/data‑flow, custom memory management, and remains compatible with CUDA Graphs and TorchInductor. The authors demonstrate that DynaFlow can integrate representative parallelism strategies into six state‑of‑the‑art ML systems with minimal code changes, achieving up to a 1.29× throughput improvement.
arXiv:2607. 27231v1 Announce Type: cross Abstract: Large language models (LLMs) have significantly increased the demand for efficient accelerator kernels, but kernel development remains a highly specialized and labor-intensive task.
arXiv:2607. 05475v1 Announce Type: cross Abstract: Deploying Large Language Models (LLMs) on mobile devices enhances privacy and reduces latency, but is severely bottlenecked by hardware inefficiency.
AgentPerfBench is a new benchmarking suite designed to evaluate the inference performance of agentic large language models (LLMs) that handle multi‑turn, tool‑using, and context‑expanding tasks. It builds on real traces from agentic benchmarks such as SWE‑Bench and TerminalBench, and generates synthetic profiles that reflect realistic input/output lengths and turn counts. The suite also provides kernel‑level Nsight Compute traces and a multi‑dimensional roofline model to identify hardware bottlenecks and quantify the gap between traditional chat benchmarks and agentic workloads.