1080 Ti vs RTX 2080 Ti vs Titan RTX Deep Learning Benchmarks with TensorFlow - 2018 2019 2020

In this post, we are comparing the most popular deep learning graphics cards: GTX 1080 Ti, RTX 2060, RTX 2070, RTX 2080, 2080 Ti, Titan RTX, and TITAN V.

We are excited to see how new NVIDIA's architecture with the tensor cores will perform compared to "old-style" NVIDIA GTX-series without tensor cores.

• We evaluate the performance on each GPU by measuring FP32 and FP16 throughput (# of training samples processed per second) while training common models on synthetic data.
• For each model, we are running 10 training experiments and measured # of images processed per second and averaged the results of the 10 experiments.

• Ubuntu 18.04 (Bionic)
• TensorFlow 1.4.0
• CUDA 9
• CuDNN 7.0
• NVIDIA driver 415 (NVIDIA TITAN RTX support)

Networks

• VGG16
• Resnet152

Benchmark

Yasuko benchmark (https://github.com/u39kun/deep-learning-benchmark).

Test bench:

BIZON G3000 4 GPU deep learning desktop
More details: https://bizon-tech.com/bizon-g3000.html

Tech specs:

• CPU: 14-Core 3.30 GHz Intel Core i9-9940X (latest generation).
• Cooling: CPU liquid cooling system (whisper-quiet)
• RAM: 128 DDR4 3000 MHz.
• Storage: 1 TB PCIe SSD.

BIZON G7000 8 GPU deep learning server
More details: https://bizon-tech.com/bizon-g7000.html

Tech specs:

• CPU: Dual 18-Core 2.30 GHz Intel Xeon Gold 6140 (latest generation).
• RAM: 256 DDR4 2666 MHz ECC.
• Storage: 1 TB PCIe SSD.

GPUs list:

• NVIDIA GTX 1080 Ti
• NVIDIA RTX 2060
• NVIDIA RTX 2070
• NVIDIA RTX 2080
• NVIDIA RTX 2080 Ti
• NVIDIA TITAN RTX
• NVIDIA TITAN V