Saved in:
Bibliographic Details
Main Author: Das, Anurita
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2604.21026
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914504441331712
author Das, Anurita
author_facet Das, Anurita
contents Deploying large language models to heterogeneous hardware is often constrained by memory, not compute. We introduce MCAP (Monte Carlo Activation Profiling), a load-time per-layer importance estimator that enables dynamic precision and memory placement decisions on the target device. MCAP produces a lightweight per-layer signal that drives both precision dispatch (W4A8 vs. W4A16) and residency tier (GPU, RAM, SSD), allowing a single set of weights to operate across diverse memory budgets. Our system, NVE, achieves 1.5-1.8x higher decode throughput than llama-cpp Q4_0 on NVIDIA T4 and enables models to run in memory regimes previously infeasible without modifying weights.
format Preprint
id arxiv_https___arxiv_org_abs_2604_21026
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle MCAP: Deployment-Time Layer Profiling for Memory-Constrained LLM Inference
Das, Anurita
Machine Learning
Deploying large language models to heterogeneous hardware is often constrained by memory, not compute. We introduce MCAP (Monte Carlo Activation Profiling), a load-time per-layer importance estimator that enables dynamic precision and memory placement decisions on the target device. MCAP produces a lightweight per-layer signal that drives both precision dispatch (W4A8 vs. W4A16) and residency tier (GPU, RAM, SSD), allowing a single set of weights to operate across diverse memory budgets. Our system, NVE, achieves 1.5-1.8x higher decode throughput than llama-cpp Q4_0 on NVIDIA T4 and enables models to run in memory regimes previously infeasible without modifying weights.
title MCAP: Deployment-Time Layer Profiling for Memory-Constrained LLM Inference
topic Machine Learning
url https://arxiv.org/abs/2604.21026