# TFApprox: Towards a Fast Emulation of DNN Approximate Hardware Accelerators on GPU

arXiv: https://arxiv.org/abs/2002.09481 · HTML: https://mlsanity.com/paper/2002.09481/

## Key facts

- Cited by 17 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 21 Feb 2020.

## Built on by (most cited first, 17 of 17)

- [Hardware Approximate Techniques for Deep Neural Network Accelerators: A Survey](https://mlsanity.com/paper/2203.08737/) (2022)
- [Control Variate Approximation for DNN Accelerators](https://mlsanity.com/paper/2102.09642/) (2021)
- [AdaPT: Fast Emulation of Approximate DNN Accelerators in PyTorch](https://mlsanity.com/paper/2203.04071/) (2022)
- [ApproxTrain: Fast Simulation of Approximate Multipliers for DNN Training and Inference](https://mlsanity.com/paper/2209.04161/) (2022)
- [TransAxx: Efficient Transformers with Approximate Computing](https://mlsanity.com/paper/2402.07545/) (2024)
- [Positive/Negative Approximate Multipliers for DNN Accelerators](https://mlsanity.com/paper/2107.09366/) (2021)
- [HAWX: A Hardware-Aware FrameWork for Fast and Scalable ApproXimation of DNNs](https://mlsanity.com/paper/2602.16336/) (2026)
- [Combining Gradients and Probabilities for Heterogeneous Approximation of Neural Networks](https://mlsanity.com/paper/2208.07265/) (2022)
- [Leveraging Highly Approximated Multipliers in DNN Inference](https://mlsanity.com/paper/2412.16757/) (2024)
- [HEAM: High-Efficiency Approximate Multiplier Optimization for Deep Neural Networks](https://mlsanity.com/paper/2201.08022/) (2022)
- [DNNShield: Dynamic Randomized Model Sparsification, A Defense Against Adversarial Machine Learning](https://mlsanity.com/paper/2208.00498/) (2022)
- [FAMES: Fast Approximate Multiplier Substitution for Mixed-Precision Quantized DNNs--Down to 2 Bits!](https://mlsanity.com/paper/2411.18055/) (2024)
- [Carbon-Efficient 3D DNN Acceleration: Optimizing Performance and Sustainability](https://mlsanity.com/paper/2504.09851/) (2025)
- [Gradient Estimation Methods of Approximate Multipliers for High-Accuracy Retraining of Deep Learning Models](https://mlsanity.com/paper/2509.10519/) (2025)
- [Approximate Multiplier Induced Error Propagation in Deep Neural Networks](https://mlsanity.com/paper/2512.06537/) (2025)
- [AxMoE: Characterizing the Impact of Approximate Multipliers on Mixture-of-Experts DNN Architectures](https://mlsanity.com/paper/2605.04754/) (2026)
- [FAME: An FPGA-Based Platform for Approximate Multipliers Evaluation with Pattern-Guided DNN Retraining](https://mlsanity.com/paper/2609.17730/) (2026)

## Abstract (arXiv)

Energy efficiency of hardware accelerators of deep neural networks (DNN) can be improved by introducing approximate arithmetic circuits. In order to quantify the error introduced by using these circuits and avoid the expensive hardware prototyping, a software emulator of the DNN accelerator is usually executed on CPU or GPU. However, this emulation is typically two or three orders of magnitude slower than a software DNN implementation running on CPU or GPU and operating with standard floating point arithmetic instructions and common DNN libraries. The reason is that there is no hardware support for approximate arithmetic operations on common CPUs and GPUs and these operations have to be expensively emulated. In order to address this issue, we propose an efficient emulation method for approximate circuits utilized in a given DNN accelerator which is emulated on GPU. All relevant approximate circuits are implemented as look-up tables and accessed through a texture memory mechanism of CUDA capable GPUs. We exploit the fact that the texture memory is optimized for irregular read-only access and in some GPU architectures is even implemented as a dedicated cache. This technique allowed us to reduce the inference time of the emulated DNN accelerator approximately 200 times with respect to an optimized CPU version on complex DNNs such as ResNet. The proposed approach extends the TensorFlow library and is available online at https://github.com/ehw-fit/tf-approximate.