GTX 1070 and GTX 1080 Premiere Pro Performance

Always look at the date when you read an article. Some of the content in this article is most likely out of date, as it was written on June 11, 2016. For newer information, see GTX 1070 and GTX 1080 Premiere Pro Performance

Table of Contents

Introduction

Whenever new video cards are released, the very first thing our customers tend to ask is "how fast are they?" This is a great question, especially since all the online reviews tend to focus on their performance in video games rather than in professional applications. Today we are going to look specifically at Premiere Pro to see just how well the GeForce GTX 1070 and GTX 1080 perform in with a variety of footage codecs and resolutions.

With Premiere Pro, Adobe utilizes the Mercury Playback Engine which uses the video card to vastly improve the performance of certain features. This provides a tremendous boost to performance, but it adds more complexity to the question of "what hardware do I need" since you need to take into consideration how many accelerated effects you use. In this article, we want to explore the performance differences between the new GTX 1070 and GTX 1080 and the previous generation GTX 980 Ti and GTX Titan X using a variety of different codes and resolutions.

Test Setup

For our test system, we used the following hardware:

Testing Hardware
Motherboard:	Asus Z10PE-D8 WS	Asus X99 Deluxe II/U3.1
CPU:	2x Intel Xeon E5-2687W V4 3.0GHz (3.2-3.5GHz Turbo) Twelve Core	Intel Core i7 6950X 3.0GHz (3.4-4.0GHz Turbo) Ten Core
RAM:	8x Samsung DDR4-2133 32GB ECC Reg. (256GB total)
Hard Drive:	Samsung 850 Pro 512GB SATA 6Gb/s SSD
OS:	Windows 10 Pro 64-bit
PSU:	EVGA SuperNOVA 1200W P2
Software:	Premiere Pro CC 2015

Since Premiere heavily utilizes the CPU in addition to the video card, we will actually be testing with two different systems. The first is a high-end dual Xeon system with two E5-2687W V4 CPUs. The second system uses a Core i7 6950X which is a bit more standard of a CPU choice for Premiere Pro.

The different video cards we will be testing are:

Test Video Cards
1-2x NVIDIA GeForce GTX 1080 8GB
1-2x NVIDIA GeForce GTX 1070 8GB
1-2x NVIDIA GeForce GTX Titan X 12GB
1-2x NVIDIA GeForce GTX 980 Ti 6GB

To help with consistency – and since the benchmarks we performed ran for several days – we programmed a custom script using AutoIt to start Premiere Pro, load the relevant project, time how long it takes to export the timeline with the appropriate settings or generate previews, close Premiere Pro to clear any data from the system RAM, then loop while changing the project file.

The files we will be testing with came from a variety of sources:

1080P H.264/CineForm 4K H.264/CineForm	Provided by: Jerry Berg Barnacules Nerdgasm – YouTube
ProRes 4K	Grant Petty Blackmagic Design Forum (available for public download)

In order to make our testing as accurate as possible, we used relatively simply timelines for our testing in this article. In the past, we've loaded on the accelerated effects to show the maximum difference between cards, but we found that this was not representative of real-world performance gains. Our test timelines consisted of:

4-5 clips arranged in series to make a 60 second timeline
A basic transition was applied to each clip
Lumetri color correction effect applied to each clip
Vector-based logo graphic added to the bottom corner of the footage

Exporting to 1080p – Single GPU

While more and more people are starting to shoot in 4K and higher resolutions, 1080p is still by far the most common resolution to export to. Because of this, we thought we would start our testing by looking at how Premiere is able to utilize a single GPU when exporting from a variety of resolutions and codecs to H.264 1080p:

2x Intel Xeon E5-2687W V4

Intel Core i7 6950X

Right off the bat, we are getting some very interesting results. With some codes (1080p H.264 especially), the GTX 1070 and 1080 are much faster than the older GTX 980Ti and Titan X. In fact, on the Dual Xeon with 1080p H.264 source footage, the GTX 1070 is about 9% faster than a GTX 980Ti and 3% faster than a Titan X. The GTX 1080 is even better clocking in at over 13.5% faster than a 980Ti and 7% faster than a Titan X!

However, for the other file formats we only saw minimal performance gains from the new cards. Overall, on the Dual Xeon system the GTX 1070 was actually a hair slower than the 980Ti, although it was still about 2% faster than the Titan X. The GTX 1080, however, was much better at about 3.5% faster than a 980Ti and 6% faster than a Titan X.

Unfortunately, the story isn't quite as good with the Core i7 system. In this case, the GTX 1070 is actually a bit slower than the older cards by about 2-2.5%. The GTX 1080 is still a hair faster than the GTX Titan X, but only by about 2%

Note: You may notice that the charts above (and the dual GPU charts in the next section) often show faster render times on the Core i7 system than on the Dual Xeon system. This is not a mistake! If you read our Adobe Premiere Pro CC 2015 Multi Core Efficiency (Update1) article, you will see that when exporting to 1080p, Premiere Pro is not very good at using a high number of CPU cores and will at times actually see a drop in performance with two physical CPUs verses just a single CPU. Since the Core i7 6950X has an all-core Turbo frequency that is .2GHz higher than the dual Xeon-E5 2687W V4 CPUs, this actually makes the Core i7-6950X faster when exporting to 1080p.

Exporting to 1080p – Dual GPU

2x Intel Xeon E5-2687W V4

Intel Core i7 6950X

With two video cards, our results are quite a bit different than what we saw with a single GPU. This time, both the GTX 1070 and 1080 were about .5% faster than the GTX 980Ti and Titan X. This isn't much, and honestly is close enough that we would deem it a wash between the four cards.

On the Core i7 system, however, the two newer cards performed great. The GTX 1070 was about 1.5-2% faster than the GTX 980Ti and Titan X. The GTX 1080 was even better and outperformed the two older cards by 4-4.5%

The good thing to note is that if you compare these raw results to the results from our single GPU test, it becomes clear that utilizing two video cards when exporting to 1080p can result in great performance gains. On average, we saw about a 20% jump in performance on the Dual Xeon system, and almost a 50% increase in performance on the Core i7 system. So while the difference between the pairs of cards we tested isn't much in many situations, simply having two cards in the first place works nicely when exporting to 1080p.

Exporting to 4K – Single GPU

Although 4K isn't as widespread as 1080p quite yet, more and more Premiere Pro users are either exporting to 4K or are considering exporting to 4K in the near future. Exporting to higher resolutions is more taxing on the system, so it will be interesting to see if the difference between each video card changes when exporting to 4K versus exporting to 1080p:

2x Intel Xeon E5-2687W V4

Intel Core i7 6950X

There were some odd results in this testing – primarily stemming from a couple of times where the GTX Titan X performed much slower than we expected. Overall, on the Dual Xeon system the GTX 1070 was a bit slower than the GTX 980Ti, but about 7.5% faster than the GTX Titan X. The GTX 1080 was similar, although it was about .5% faster than the GTX 980Ti and 8.5% faster than the GTX Titan X.

When using just the single CPU in the Core i7 system, however, the new cards performed much better. On this system, the GTX 1070 was about 8.5% faster than the older cards. The GTX 1080 was a even faster, beating the older cards by about 10%

Exporting to 4K – Dual GPU

2x Intel Xeon E5-2687W V4

Intel Core i7 6950X

With two video cards, there isn't really much to discuss. On the Dual Xeon system the newer cards were about 1.5% faster than the older cards, but on the Core i7 system they were about .5% slower.

If you again compare the single GPU results to the results in this section, you will notice that using two cards is indeed faster than just using a single card when exporting to 4K. Unfortunately, this time dual GPUs was only about 2.5% faster on the Dual Xeon system and about 10% faster on the Core i7 system.

Render Previews – Single GPU

2x Intel Xeon E5-2687W V4

Intel Core i7 6950X

When rendering previews, we saw some interesting results because – for whatever reason – the GTX Titan X was often fairly slower than the other cards. On the Dual Xeon system, this made the GTX 1070 about 8% faster than the GTX Titan X, although it was only roughly equal to the GTX 980Ti. The GTX 1080, on the other hand, was about 2% faster than the GTX 980Ti and about 10.5% faster than the Titan X.

On the Core i7 system, however, the GTX Titan X was right back in line. Because of this, the GTX 1070 was about 1% faster than the GTX 980Ti and roughly equal to the GTX Titan X. The GTX 1080 was about 4% faster than the GTX 980Ti and about 3% faster than the GTX TItan X.

Render Previews – Dual GPU

2x Intel Xeon E5-2687W V4

Intel Core i7 6950X

We saw pretty wide fluctuations in performance between each test, but on the Dual Xeon system the GTX 1070 was on average about 5% faster than the GTX 980Ti and about 15% faster than the GTX Titan X. For whatever reason, the GTX 1080 was actually a bit slower clocking in at about 2.5% faster than a GTX 980Ti and about 12% faster than a GTX Titan X.

The results were not quite as impressive with the Core i7 system, but the GTX 1070 was still about 3.5% faster than the GTX 980Ti. The GTX 1080, on the other hand, was about 8.5% faster than the GTX 980Ti and about 4% faster than the GTX Titan X.

Once again, the big thing we saw was an excellent speedup when going from one GPU to two. In the case of rendering previews, we saw about a 30% increase in performance on the Dual Xeon system and a 40% increase in performance on the Core i7 system.

Conclusion

Summarizing all our results, we saw the following performance gains over the old cards on the two systems we tested with:

2x Xeon E5-2687W V4	Single GPU		Dual GPU
Average Performance Difference	GTX 1070	GTX 1080	GTX 1070	GTX 1080
Exporting to 1080p	.5% slower than 980Ti 2% faster than Titan X	3.5% faster than 980Ti 6% faster than Titan X	.5% faster than 980Ti .5% faster than Titan X	.5% faster than 980Ti .5% faster than Titan X
Exporting to 4K	1% slower than 980Ti 7.5% faster than Titan X	.5% faster than 980Ti 8.5% faster than Titan X	1% faster than 980Ti 2% faster than Titan X	1.5% faster than 980Ti 2% faster than Titan X
Render Previews	.25% faster than 980Ti 8% faster than Titan X	2.5% faster than 980Ti 10.5% faster than Titan X	5% faster than 980Ti 15% faster than Titan X	2.5% faster than 980Ti 12% faster than Titan X
Average	.5% slower than 980Ti 6% faster than Titan X	2% faster than 980Ti 8.5% faster than TItan X	2% faster than 980Ti 6% faster than Titan X	1.5% faster than 980Ti 5% faster than Titan X

Core i7 6950X	Single GPU		Dual GPU
Average Performance Difference	GTX 1070	GTX 1080	GTX 1070	GTX 1080
Exporting to 1080p	2% slower than 980Ti 2.5% slower than Titan X	2% faster than 980Ti 1.5% faster than Titan X	2% faster than 980Ti 1.5% faster than Titan X	4.5% faster than 980Ti 4% faster than Titan X
Exporting to 4K	8.5% faster than 980Ti 8.5% faster than Titan X	10% faster than 980Ti 10% faster than Titan X	1% slower than 980Ti .5% slower than Titan X	.5% slower than 980Ti equal to Titan X
Render Previews	1% faster than 980Ti equal to Titan X	4% faster than 980Ti 3% faster than Titan X	3.5% faster than 980Ti 1% slower than Titan X	8.5% faster than 980Ti 4% faster than Titan X
Average	2.5% faster than 980Ti 2% faster than Titan X	5.5% faster than 980Ti 5% faster than TItan X	1.5% faster than 980Ti equal to Titan X	4% faster than 980Ti 2.5% faster than Titan X

If there is one thing we learned in our testing, it is that the performance difference between the cards can change drastically depending on what source codec and resolution you use as well as what resolution you are exporting to. So if you tend to only work with one of the codes we specifically tested, we highly recommend ignoring these averages and looking solely at that one test.

Averaging everything out, however, we found that

The GTX 1070 is about 1% faster than the GTX 980Ti and about 4% faster than the Titan X.
The GTX 1080 is about 4% faster than the GTX 980Ti and about 6.5% faster than the Titan X.

Again, this is only an average so it won't be true in all cases. Sometimes, the newer cards were actually slower than the older ones; other times they were as much as 20% faster. Also, keep in mind that the more GPU accelerated effects you use the larger the difference should become – so if you tend to use a lot of the accelerated effects the GTX 1070 and 1080 should provide an even larger benefit than what we showed in this article

One last thing we want to point out is that, while this article is primarily about looking at the performance of the GTX 1070 and 1080, one thing we did find was that dual GPU configurations can often work really well for Premier Pro. We didn't see much of a gain when exporting to 4K (only 2.5-10% better performance), but exporting to 1080p and rendering previews was anywhere from 20% to 50% faster with two video cards versus just one.