Puget Systems print logo

https://www.pugetsystems.com

Read this article at https://www.pugetsystems.com/guides/1544
Article Thumbnail

Redshift 2.6.41 GPU Roundup: NVIDIA GeForce RTX SUPER Performance

Written on August 9, 2019 by William George
Share:

Introduction

Redshift is a production-quality, GPU-accelerated renderer. Traditionally this type of rendering was done on CPUs, but graphics processors (GPUs) are ideal for highly parallel tasks like this - and it is easier to fit multiple video cards in a single computer, to boost performance, than multiple CPUs.

It was written to use NVIDIA's CUDA graphics programming language, and since NVIDIA recently refreshed their GeForce series with new 2060, 2070, and 2080 "SUPER" cards we thought it would be a good time to re-test the whole RTX lineup.

Looking for a Rendering Workstation?

Puget Systems offers a range of poweful and reliable systems that are tailor-made for your unique workflow.

Configure a System!

Labs Consultation Service

Our Labs team is available to provide in-depth hardware recommendations based on your workflow.

Find Out More!

Test Hardware

Here is a list of the hardware we tested Redshift on. Everything except the video card was kept the same throughout all the test runs, to ensure that no other specs could impact performance. The operating system was a fully updated installation of Windows 10 Pro.

Benchmark Details

Redshift has a handy benchmarking tool which tests how long it takes to render an included scene. We ran this three times on each card, and used the fastest (shortest) time for each card in the graphs below.

Benchmark Results

Here are the Redshift 2.6.41 benchmark render times all of the GeForce RTX video cards, along with the TITAN RTX for good measure, sorted from most to least expensive - with the new SUPER series cards in a darker shade of green:

Redshift 2.6.41 NVIDIA GeForce RTX, RTX SUPER, and TITAN RTX rendering performance

And here is another way of looking at the results, as percentages relative to the GeForce RTX 2060's performance:

Redshift 2.6.41 NVIDIA GeForce RTX, RTX SUPER, and TITAN RTX rendering performance relative to the RTX 2060

Analysis

The new RTX 2070 SUPER and 2080 SUPER cards are just incremental improvements, but since they are the same price as the cards they are replacing that is nothing to complain about. However, the RTX 2060 SUPER is a different story: it is a little more expensive than the vanilla 2060 - but it is also over 25% faster! Moreover, it has 8GB of memory instead of 6GB (just like the 2070 and 2080 models).

Nothing changed at the top-end of NVIDIA's line, though. The TITAN RTX is still the fastest card overall, but considering that it is twice the price of the RTX 2080 Ti, only about 5% faster, and isn't available in a single-fan (blower style) card it isn't really a good deal. Seemingly the only reason to go for the TITAN in this application would be if you needed the larger VRAM (24 vs 11GB) and you only wanted to use a single GPU (or maybe two at most, in a system with tons of airflow). If you need lots of memory space but also want multiple GPUs, take a look at the Quadro RTX line instead.

Conclusion: What Are the Best Consumer GPUs for Rendering in Redshift?

Because of the Titan RTX's cooling limitations, and how well Redshift and other GPU rendering applications scale across multiple cards, the best choice for most users is going to be the GeForce RTX 2080 Ti. The new RTX 2060 SUPER also offers great performance for the price, and the other SUPER series cards provide modest increases in rendering speed for the same price as previous models.

Looking for a Rendering Workstation?

Puget Systems offers a range of poweful and reliable systems that are tailor-made for your unique workflow.

Configure a System!

Labs Consultation Service

Our Labs team is available to provide in-depth hardware recommendations based on your workflow.

Find Out More!
Tags: Rendering, Benchmark, Performance, NVIDIA, GeForce, RTX, Turing, Redshift, Titan RTX, RTX 2080, RTX 2080 Ti, RTX 2070, RTX 2060, GPU
Jorge Flores Carrillo

Hello I'm using ggogle translator, you can recommend me if it's better 1 pc with 8 GPU rtx 2080ti or 2 pcs with 4 GPU srtx 2080ti, it's for maya with redshift using thinkbox deadline

Posted on 2019-08-25 11:19:27

I'm sorry, but we have not done any multi-system testing - so I am not sure if it ends up being faster to have all eight GPUs in one system or split across two. Having a single workstation should mean lower licensing costs for some software, though, and less complicated setup / management... while two systems could give you flexibility (if you needed to work on two different projects at once) and redundancy (in case one of the systems failed). So there are pros and cons to both approaches :/

Posted on 2019-08-26 16:42:44
andrew

would be good to see a comparison with 1080TI and older gen cards. the 1080 is a great second hand buy still.

Posted on 2020-02-23 05:06:39

I would hesitate to recommend a 10-series card on modern systems, since only the 20-series cards have the new RTX features that are starting to get traction in GPU rendering - and have a huge performance impact. However, if you wanted to get an idea of how cards like the 1080 Ti stack up, you could cross reference the results from our last test that included them and use the 2080 or 2080 Ti (which were also in that set of tests) to see how the older models would sit alongside newer ones:

https://www.pugetsystems.co...

Posted on 2020-02-24 19:15:57
King Gore

It seems redshift 2.6.41 doesn't support RT cores. So 10 Series cards are still nice choices.

Posted on 2020-03-12 16:11:48

Correct: Redshift 2.X versions don't yet have hardware RT support. That is coming in version 3, which I believe is available as an experimental build for paying customers - but I'm not sure when it will reach release status, and thus be available for us to test.

https://www.redshift3d.com/...

Posted on 2020-03-13 18:13:19
andrew

Still its useful info... For those that dabble in redshift but spend most of their time in open GL, I still see the last gen cards as a good buy.
I'll probably buy a P5000 soon to replace my 1080TI because of that reason.

Posted on 2020-05-15 10:41:09
Alex

I'm curious as to anyone's thoughts on two 2080 Super cards vs one 2080 Ti. I know you guys did another test showing that RTX cards offer performance that scales linearly with the number of cards. It would seem that two 2080 Supers would be faster than 1 2080 Ti...

Posted on 2020-05-06 18:26:32

Two 2080 Supers would definitely be faster, but also have a little less VRAM (which may or may not impact your usage).

Posted on 2020-05-06 18:40:14