NVIDIA Quadro RTX 4000 Review A Versatile AI and Professional GPU

NVIDIA Quadro RTX 4000 Deep Learning benchmarks

As we continue to innovate on our review format, we are now adding deep learning benchmarks. In future reviews, we will add more results to this data set. At this point, we have a fairly nice data set to work with.

ResNet-50 Inferencing in TensorRT using Tensor Cores

ImageNet is an image classification database launched in 2007 designed for use in visual object recognition research. Organized by the WordNet hierarchy, hundreds of image examples represent each node (or category of specific nouns).

In our benchmarks for Inferencing, a ResNet50 Model trained in Caffe will be run using the command line as follows.

nvidia-docker run --shm-size=1g --ipc=host --ulimit memlock=-1 --ulimit stack=67108864 --rm -v ~/Downloads/models/:/models -w /opt/tensorrt/bin nvcr.io/nvidia/tensorrt:18.11-py3 giexec --deploy=/models/ResNet-50-deploy.prototxt --model=/models/ResNet-50-model.caffemodel --output=prob --batch=16 --iterations=500 --fp16

Options are:
–deploy: Path to the Caffe deploy (.prototxt) file used for training the model
–model: Path to the model (.caffemodel)
–output: Output blob name
–batch: Batch size to use for inferencing
–iterations: The number of iterations to run
–int8: Use INT8 precision
–fp16: Use FP16 precision (for Volta or Turing GPUs), no specification will equal FP32

We can change the batch size to 16, 32, 64, 128 and precision to INT8, FP16, and FP32.

The results are in inference latency (in seconds.) If we take the batch size / Latency, that will equal the Throughput (images/sec) which we plot on our charts.

We also found that this benchmark does not use two GPU’s; it only runs on a single GPU. You can, however, run different instances on each GPU using commands like.
```NV_GPUS=0 nvidia-docker run ... &
NV_GPUS=1 nvidia-docker run ... &```
With these commands, a user can scale workloads across many GPU’s. Our graphs show combined totals.

We start with Turing’s new INT8 mode which is one of the benefits of using the NVIDIA RTX cards.

NVIDIA Quadro RTX 4000 ResNet Inferencing INT8Using the precision of INT8 is by far the fastest inferencing method if at all possible, converting code to INT8 will yield faster runs. Installed memory has one of the most significant impacts on these benchmarks. Inferencing on NVIDIA RTX graphics cards does not tax the GPU’s to a great deal. However, additional memory allows for larger batch sizes.

Inferencing on NVIDIA RTX graphics cards does not tax the GPUs a great deal, however additional memory allows for larger batch sizes.

Let us look at FP16 and FP32 results.

Here we see performance quite a bit better than both the NVIDIA GeForce RTX 2070 super. Perhaps more interesting is that with the extra power headroom, and the larger form factor, the NVIDIA Quadro RTX 4000 performs quite a bit better than the more expensive NVIDIA T4.

This intuitively makes sense. NVIDIA is working with content creation software providers such as Adobe to accelerate workflows of tools such as Adobe Sensei that use AI to greatly speed up certain tasks. For many professional users, having great AI inferencing performance will directly impact daily content creation workflows.

Training with ResNet-50 using Tensorflow

We also wanted to train the venerable ResNet-50 using Tensorflow. During training the neural network is learning features of images, (e.g. objects, animals, etc.) and determining what features are important. Periodically (every 1000 iterations), the neural network will test itself against the test set to determine training loss, which affects the accuracy of training the network. Accuracy can be increased through repetition (or running a higher number of epochs.)

The command line we will use is:

nvidia-docker run --shm-size=1g --ipc=host --ulimit memlock=-1 --ulimit stack=67108864 -v ~/Downloads/imagenet12tf:/imagenet --rm -w /workspace/nvidia-examples/cnn/ nvcr.io/nvidia/tensorflow:18.11-py3 python resnet.py --data_dir=/imagenet --layers=50 --batch_size=128 --iter_unit=batch --num_iter=500 --display_every=20 --precision=fp16

Parameters for resnet.py:
–layers: The number of neural network layers to use, i.e. 50.
–batch_size or -b: The number of ImageNet sample images to use for training the network per iteration. Increasing the batch size will typically increase training performance.
–iter_unit or -u: Specify whether to run batches or epochs.
–num_iter or -i: The number of batches or iterations to run, i.e. 500.
–display_every: How frequently training performance will be displayed, i.e. every 20 batches.
–precision: Specify FP32 or FP16 precision, which also enables TensorCore math for Volta and Turing GPUs.

While this script TensorFlow cannot specify individual GPUs to use, they can be specified by
setting export CUDA_VISIBLE_DEVICES= separated by commas (i.e. 0,1,2,3) within the Docker container workspace.

We will run batch sizes of 16, 32, 64, 128 and change from FP16 to FP32. Our graphs show combined totals.

After the NVIDIA T4 review that we did (called Tesla T4 at the time) and is linked above, we discussed how the RTX 2070 (Super) may be a better value for some looking for lower-cost training and inferencing. The Quadro RTX 4000 is perhaps that next-level of intriguing since it offers a single-slot design. While not low-profile, nor passively cooled as are required for many servers, this does put the Quadro RTX 4000 into an interesting position.

Let’s move on to power and thermal testing and continue with our final words.