mlsanityData to 14 Oct 2026

Optimizing ML Workload Partitioning between CPUs and CIM Accelerators for Heterogeneous Computing

arXiv 2607.05240 ↗First posted 6 Jul 2026cs.ET cs.AI cs.ARcited by 0

Not on the map. Why?

Abstract (arXiv)

Computing-in-Memory (CIM) accelerators execute Matrix-Vector Multiplications (MVMs) in memory, making them a compelling solution for Machine Learning (ML) workloads. However, existing ML workload partitioning approaches for CIM accelerators do not fully account for Resistive Random Access Memory (RRAM) constraints such as limited memory, high write latency, and limited endurance. They also neglect parallelism, low-level architectural effects, or the Central Processing Unit (CPU) as a complementary compute resource. To address these limitations, we propose an Integer Linear Programming (ILP)-based workload partitioning framework for heterogeneous CPU-CIM systems. It minimizes end-to-end inference latency under RRAM constraints, captures parallelism, and combines empirical profiling with analytical models. Using our framework, heterogeneous CPU-CIM execution achieves speedups of up to 30.9x over CPU-only execution on an edge CPU and 7.3x over a high-performance CPU. A Design Space Exploration (DSE) yields further design insights for future CIM accelerators.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 24 of its 40 references that are in the map. First posted on arXiv 6 Jul 2026.

Builds on

24 of 40 references.

References · 24

  1. 2017On Calibration of Modern Neural Networkscited by 4,960
  2. 2025V-JEPA 2: Self-Supervised Video Models Enable Understanding, Prediction and Planningcited by 809
  3. 2023Argoverse 2: Next Generation Datasets for Self-Driving Perception and Forecastingcited by 707
  4. 2019MultiPath: Multiple Probabilistic Anchor Trajectory Hypotheses for Behavior Predictioncited by 420
  5. 2024DINO-WM: World Models on Pre-trained Visual Features enable Zero-shot Planningcited by 386
  6. 2022Motion Transformer with Global Intention Localization and Local Movement Refinementcited by 270
  7. 2019CoverNet: Multimodal Behavior Prediction using Trajectory Setscited by 259
  8. 2024OGBench: Benchmarking Offline Goal-Conditioned RLcited by 238
  9. 2023MotionDiffuser: Controllable Multi-Agent Motion Prediction using Diffusioncited by 213
  10. 2020DLow: Diversifying Latent Flows for Diverse Human Motion Predictioncited by 185
  11. 2016Learning in an Uncertain World: Representing Ambiguity Through Multiple Hypothesescited by 97
  12. 2019Overcoming Limitations of Mixture Density Networks: A Sampling and Fitting Framework for Multimodal Future Predictioncited by 96
12 more
  1. 2021LookOut: Diverse Multi-Future Prediction and Planning for Self-Drivingcited by 96
  2. 2016Stochastic Multiple Choice Learning for Training Diverse Deep Ensemblescited by 77
  3. 2023Forecast-MAE: Self-supervised Pre-training for Motion Forecasting with Masked Autoencoderscited by 73
  4. 2022ScePT: Scene-consistent, Policy-based Trajectory Predictions for Planningcited by 64
  5. 2025What Drives Success in Physical Planning with Joint-Embedding Predictive World Models?cited by 43
  6. 2020Trajectory-wise Multiple Choice Learning for Dynamics Generalization in Reinforcement Learningcited by 35
  7. 2023Tree-structured Policy Planning with Learned Behavior Modelscited by 24
  8. 2016DISCO Nets: DISsimilarity COefficient Networkscited by 18
  9. 2026VJEPA: Variational Joint Embedding Predictive Architectures as Probabilistic World Modelscited by 17
  10. 2024Denoising with a Joint-Embedding Predictive Architecturecited by 14
  11. 2025Continuous Visual Autoregressive Generation via Score Maximizationcited by 9
  12. 2026UWM-JEPA: Predictive World Models That Imagine in Belief Spacecited by 3

Built on by

Not cited yet.