# GANAX: A Unified MIMD-SIMD Acceleration for Generative Adversarial Networks

arXiv: https://arxiv.org/abs/1806.01107 · HTML: https://mlsanity.com/paper/1806.01107/

## Key facts

- Cited by 11 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 10 May 2018.

## Built on by (most cited first, 11 of 11)

- [FLAT: An Optimized Dataflow for Mitigating Attention Bottlenecks](https://mlsanity.com/paper/2107.06419/) (2021)
- [Mind Mappings: Enabling Efficient Algorithm-Accelerator Mapping Space Search](https://mlsanity.com/paper/2103.01489/) (2021)
- [A Survey on Deep Learning Hardware Accelerators for Heterogeneous HPC Platforms](https://mlsanity.com/paper/2306.15552/) (2023)
- [Generative AI for the Optimization of Next-Generation Wireless Networks: Basics, State-of-the-Art, and Open Challenges](https://mlsanity.com/paper/2405.17454/) (2024)
- [A Competitive Edge: Can FPGAs Beat GPUs at DCNN Inference Acceleration in Resource-Limited Edge Computing Applications?](https://mlsanity.com/paper/2102.00294/) (2021)
- [Rethinking Floating Point Overheads for Mixed Precision DNN Accelerators](https://mlsanity.com/paper/2101.11748/) (2021)
- [EcoFlow: Efficient Convolutional Dataflows for Low-Power Neural Network Accelerators](https://mlsanity.com/paper/2202.02310/) (2022)
- [Kernel-Segregated Transpose Convolution Operation](https://mlsanity.com/paper/2209.03704/) (2022)
- [PhotoGAN: Generative Adversarial Neural Network Acceleration with Silicon Photonics](https://mlsanity.com/paper/2501.13828/) (2025)
- [Register Your Forests: Decision Tree Ensemble Optimization by Explicit CPU Register Allocation](https://mlsanity.com/paper/2404.06846/) (2024)
- [Unified Kernel-Segregated Transpose Convolution Operation](https://mlsanity.com/paper/2502.20493/) (2025)

## Abstract (arXiv)

Generative Adversarial Networks (GANs) are one of the most recent deep learning models that generate synthetic data from limited genuine datasets. GANs are on the frontier as further extension of deep learning into many domains (e.g., medicine, robotics, content synthesis) requires massive sets of labeled data that is generally either unavailable or prohibitively costly to collect. Although GANs are gaining prominence in various fields, there are no accelerators for these new models. In fact, GANs leverage a new operator, called transposed convolution, that exposes unique challenges for hardware acceleration. This operator first inserts zeros within the multidimensional input, then convolves a kernel over this expanded array to add information to the embedded zeros. Even though there is a convolution stage in this operator, the inserted zeros lead to underutilization of the compute resources when a conventional convolution accelerator is employed. We propose the GANAX architecture to alleviate the sources of inefficiency associated with the acceleration of GANs using conventional convolution accelerators, making the first GAN accelerator design possible. We propose a reorganization of the output computations to allocate compute rows with similar patterns of zeros to adjacent processing engines, which also avoids inconsequential multiply-adds on the zeros. This compulsory adjacency reclaims data reuse across these neighboring processing engines, which had otherwise diminished due to the inserted zeros. The reordering breaks the full SIMD execution model, which is prominent in convolution accelerators. Therefore, we propose a unified MIMD-SIMD design for GANAX that leverages repeated patterns in the computation to create distinct microprograms that execute concurrently in SIMD mode.