Just want to share some benchmarks I've done with the Zotac GeForce RTX 3070 Twin Edge OC, Tensorflow 1.x and Resnet-50. It looks that FP16 is not working as expected. Also is
Installing Tensorflow 2.x with GPU support on Windows
NVIDIA GeForce RTX 2080 Super Review - Page 6 of 7 - ServeTheHome
CUDA_ERROR_OUT_OF_MEMORY: out of memory with RTX 2070 · Issue #25337 · tensorflow/tensorflow · GitHub
CUDA Toolkit Install on Ubuntu 18.04 LTS / RTX2070 Super / TensorFlow 2 - Qiita
Tutorial | Complete Guide on Installing TensorFlow 1.8 GPU on Ubuntu 17 – Beeren Sahu
RTX Titan TensorFlow performance with 1-2 GPUs (Comparison with GTX 1080Ti, RTX 2070, 2080, 2080Ti, and Titan V) | Puget Systems
RTX 2080 Ti Deep Learning Benchmarks with TensorFlow
Installing Tensorflow 2.x with GPU support on Windows
Tensorflow with RTX 2070 Super - Frameworks - NVIDIA Developer Forums
NVIDIA GeForce RTX 2060 Super Review Entry GPU Compute Leader - Page 6 of 6
RTX Titan TensorFlow performance with 1-2 GPUs (Comparison with GTX 1080Ti, RTX 2070, 2080, 2080Ti, and Titan V) | Puget Systems
NVIDIA RTX 3090 vs 2080 Ti vs TITAN RTX vs RTX 6000/8000 | Exxact Blog