NVIDIA GeForce RTX 3090 Review A Compute Powerhouse

7

NVIDIA RTX 3090 FE 3DMark Suite Testing

Here we will run the GeForce RTX 3090 through graphics-related benchmarks.

NVIDIA RTX 3090 FE Port Royal
NVIDIA RTX 3090 FE Port Royal
NVIDIA RTX 3090 FE Time Spy
NVIDIA RTX 3090 FE Time Spy
NVIDIA RTX 3090 FE Fire Strike
NVIDIA RTX 3090 FE Fire Strike

In our 3DMark testing, it is clear to see the GeForce RTX 3090 crushes the previous-generation GPUs. In only a few instances, NVLINK configurations outperform a single GeForce RTX 3090 but at a higher power and heat load cost.

NVIDIA RTX 3090 FE Unigine Testing

NVIDIA RTX 3090 FE Unigine Superposition
NVIDIA RTX 3090 FE Unigine Superposition
NVIDIA RTX 3090 FE Unigine Heaven
NVIDIA RTX 3090 FE Unigine Heaven
NVIDIA RTX 3090 FE Unigine Valley
NVIDIA RTX 3090 FE Unigine Valley

The Unigine benchmarks are older and often do not perform well with current graphics cards. We can see this in the Unigine Valley graph. With the latest Superposition benchmark, the RTX 3090 FE easily takes the top spot.

Next, we are going to look at the GPU with several Deep Learning benchmarks.

7 COMMENTS

  1. Are you using the Tensorflow 20.11 container for all the machine learning benchmarks? It contains cuDNN 8.0.4, while the already released cuDNN 8.0.5 delivers significant performance improvements for the RTX 3090.

  2. Chris Hubick

    Misha, as well known AMD shill, is being facetious – and this is a graphics card and not a compute card like the Ampere A100 – which trades the RT cores for FP64…

  3. Hi,

    I LOVE the GeForce and Threadripper compute reviews (especially the youtube video reviews!)

    However, for our work load, we really need to know how the hardware performs for double-precision memory-bound algorithms.

    The best benchmark that matches our problems (computational physics) is the HPCG benchmark.

    Would it be possible to add HPCG results for the reviews? (http://www.hpcg-benchmark.org/)

    Also, for some other computational physicists, having the standard LinPack benchmark (for compute-bound algorithms) would be really nice to see as well (https://top500.org/project/linpack/)

    – Ron

LEAVE A REPLY

Please enter your comment!
Please enter your name here

This site uses Akismet to reduce spam. Learn how your comment data is processed.