Performance Analysis - A Theoretical Two Latencies Model

Recent technological advances requires more sophisticated systems architecture to support the growing demands in performance and power requirements. A key challenge to chip designers, architects, engineers, users and scholars is how to effectively evaluate the performance of a system. In this paper, we present a theoretical model to analyze system performance based on two components: transport and execution latencies. This model was used to analyze the performance on a multithreaded memory bound workload. The results correlate with measurement data collected on an Intel Haswell architecture.

Computer Science Engineering

View graph Source

Paper

Full text

Performance Analysis - A Theoretical Two Latencies Model

Semantic Scholar · Computer Science · 2018

Abstract

Recent technological advances requires more sophisticated systems architecture to support the growing demands in performance and power requirements. A key challenge to chip designers, architects, engineers, users and scholars is how to effectively evaluate the performance of a system. In this paper, we present a theoretical model to analyze system performance based on two components: transport and execution latencies. This model was used to analyze the performance on a multithreaded memory bound workload. The results correlate with measurement data collected on an Intel Haswell architecture.

Source of record · Semantic Scholar

Similar papers

Computational system performance evaluationG. V. Petushkov2025 · Proceedings of Southwest State University In Library

The Latency Floor Model: Predictive Performance Diagnostics for Real-Time AI Inference on Edge AcceleratorsAmar Madhukar Balutkar2026 · Zenodo (CERN European Organization for Nuclear Research) In Library

The Latency Floor Model: Predictive Performance Diagnostics for Real-Time AI Inference on Edge AcceleratorsAmar Madhukar Balutkar2026 · Open MIND In Library

A Uniform Latency Model for DNN Accelerators with Diverse Architectures and DataflowsLinyan Mei, Huichu Liu, Tony F. Wu et al.2022 In Library

Writeback Modeling: Theory and Application to Zipfian WorkloadsWesley Smith, Daniel Byrne, Chen Ding2021 In Library

Performance Modeling of Non-Uniform Heterogeneous PlatformsSteven Harris, Roger D. Chamberlain, Christopher Gill2025 In Library

Non-Uniform L2 Cache Latency Across the Streaming Multiprocessors of an NVIDIA L40Faruk Alpay, B. Başaran2026 · 0 citations In Library

CloserToMe: A Unified Framework for Accurate and Transferable Latency Prediction Across Heterogeneous DevicesCheng Tang, Guochong Sui, Wenqi Lou et al.2026 · Proceedings of the AAAI Conference on Artificial Intelligence · 1 citation In Library

System-Level Performance Modeling of Photonic In-Memory ComputingJebacyril Arockiaraj, Sasindu Wijeratne, Sugeet Sunder et al.2026 · arXiv (Cornell University) In Library

System-Level Performance Modeling of Photonic In-Memory ComputingJebacyril Arockiaraj, Sasindu Wijeratne, Sugeet Sunder et al.2025 In Library

Source of record

Semantic Scholar

44B

Welcome.

This panel sticks with you. Pick anything below and it opens right beside it, so you can dig through 60,000-plus records without ever losing your spot here.

Travel 44B

LibraryPapers, policy, standards, statute LabsEvery organization building AI ModelsIntelligence, price, and speed BenchmarksThe evaluation catalog IncidentsAI failures and harms SearchOne field across everything DashboardThe state of AI accountability in NY 44B RegistryThe Art. 44-B compliance portal