# Analyzing Machine Learning Workloads Using a Detailed GPU Simulator

arXiv: https://arxiv.org/abs/1811.08933 · HTML: https://mlsanity.com/paper/1811.08933/

## Key facts

- Cited by 9 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 18 Nov 2018.

## Built on by (most cited first, 9 of 9)

- [Dual-side Sparse Tensor Core](https://mlsanity.com/paper/2105.09564/) (2021)
- [T3: Transparent Tracking & Triggering for Fine-grained Overlap of Compute & Collectives](https://mlsanity.com/paper/2401.16677/) (2024)
- [G10: Enabling An Efficient Unified GPU Memory and Storage Architecture with Smart Tensor Migrations](https://mlsanity.com/paper/2310.09443/) (2023)
- [AI Benchmark Democratization and Carpentry](https://mlsanity.com/paper/2512.11588/) (2025)
- [Machine Learning in Short-Reach Optical Systems: A Comprehensive Survey](https://mlsanity.com/paper/2405.09557/) (2024)
- [Automating Energy-Efficient GPU Kernel Generation: A Fast Search-Based Compilation Approach](https://mlsanity.com/paper/2411.18873/) (2024)
- [gSuite: A Flexible and Framework Independent Benchmark Suite for Graph Neural Network Inference on GPUs](https://mlsanity.com/paper/2210.11601/) (2022)
- [Fake Runs, Real Fixes -- Analyzing xPU Performance Through Simulation](https://mlsanity.com/paper/2503.14781/) (2025)
- [Democratizing AI: A Comparative Study in Deep Learning Efficiency and Future Trends in Computational Processing](https://mlsanity.com/paper/2603.20920/) (2026)

## Abstract (arXiv)

Most deep neural networks deployed today are trained using GPUs via high-level frameworks such as TensorFlow and PyTorch. This paper describes changes we made to the GPGPU-Sim simulator to enable it to run PyTorch by running PTX kernels included in NVIDIA's cuDNN library. We use the resulting modified simulator, which has been made available publicly with this paper, to study some simple deep learning workloads. With our changes to GPGPU-Sim's functional simulation model, we find GPGPU-Sim performance model running a cuDNN enabled implementation of LeNet for MNIST reports results within 30% of real hardware. Using GPGPU-Sim's AerialVision performance analysis tool we observe that cuDNN API calls contain many varying phases and appear to include potentially inefficient microarchitecture behaviour such as DRAM partition bank camping, at least when executed on GPGPU-Sim's current performance model.