Premiere Pro CC 2017 Mac Pro vs PC Performance

Introduction

If you are thinking of moving your Premiere Pro workstation from Mac to PC, you are likely doing so for one of three reasons:

1. The more robust capabilities available on PC (built-in video capture options, VR headset support, hardware upgrades, etc.)
2. The lower cost of PC workstations
3. The higher performance of a modern PC workstation

In this article we will be specifically looking at number 3 – examining how the top-end Mac Pro (late 2013) compares to two of our Premiere Pro workstation configurations. These two PC workstations were designed to represent both a great value option (priced at roughly half the cost of the Mac Pro) as well as the best overall performance possible with modern hardware. There are a wide variety of tasks we could test in Premiere Pro, but in this article we will specifically be looking at rendering previews, exporting, live playback, as well as a few other miscellaneous tasks.

If you would don’t like looking at charts and charts worth of data, we also have a summary of this article available on Youtube:

Test Setup

To see how the current Mac Pro compares to a modern workstation, we will be testing with the following systems:

Two of the biggest draws to a PC workstation are the price and performance so we decided to compare the Mac Pro to two different PC workstation configurations. The first is a fairly typical workstation configuration we sell for Premiere Pro using an Intel Core i7 6900K, GeForce GTX 1080 8GB video card, and 64GB of RAM. This system comes in at almost exactly half the cost of the Mac Pro so it should be a great indicator of the kind of performance you could expect while significantly reducing the cost of the workstation itself. Even if you have a large budget, freeing up almost $5,000 can open a number of possibilities including upgrading multiple workstations at the same time, upgrading your monitors and other peripherals, adding increased internal storage (which isn’t even an option on the Mac Pro), or taking a nice vacation to Hawaii.

In addition to this more affordable workstation, we will also be testing with one of the best overall hardware configurations for Premiere Pro. This workstation uses the Intel Core i7 6950X, a GTX 1080 Ti 11GB video card, and 128GB of RAM. Note that while some of our testing starts to bump up against needing more than 64GB of system RAM, the 128GB of RAM in this system should not actually impact performance for our testing. We simply included it to show that you can get an extremely capable workstation with twice the RAM of the Mac Pro while still coming in at just over two thirds the cost of the Mac Pro.

One thing we want to point out is that while all these workstations are using a single hard drive, that is not actually what we would typically recommend to our customers. We have found in our testing that using at least a two drive configuration with the media cache and scratch files on a secondary drive can make a big impact when it comes to importing footage and tasks like conforming audio. However, since the Mac Pro is only capable of having a single internal drive we opted to match that limitation on the PC workstations to ensure that we are not being overly favorable towards PC.

Most of the media we will be using is available from the Sample R3D Files and were transcoded to the various codecs we wanted to test.

To test exporting and rendering previews we used a moderately complex timeline involving multiple clips, Lumetri Color, multicam footage, and some other effects like a logo overlay, Gaussian Blur and Cross Dissolves. If you want a more in-depth look at what our timelines look like, we recorded a short video explaining our test process:

Our 4K VR testing was performed using the “Sample 1 – Ring road motorbike ride” footage from the Autopano Video Benchmarking page. We tested using both some built-in Premiere Pro effects (Lumetri Color, text overlay, and cross dissolve) as well as using the Mettle Skybox 360 VR Tools and Skybox 360/VR Transitions plug-ins to apply a number of effects such as Denoise, Rotate Sphere, Sharpen, and Iris Wipe. We typically try to avoid using plug-ins in our testing, but since Premiere Pro only has basic support for VR at the moment we felt it made sense to also look at the popular Mettle plug-in for VR projects.

Render Previews – Standard Footage

Rendering previews is something that you hope to never have to do since it interrupts your workflow, but if you do complex editing it is sometimes unavoidable. Because of this, being able to render previews as quickly as possible is often an important part of a Premiere Pro workstation.

Starting off with standard footage and projects, we found some very interesting results. While this article is about Mac vs PC primarily, we found that we had to test the Mac using both OpenCL and Metal because which one was faster was inconsistent. Metal was very good at 4K with most of the codecs we tested, but 8K DNxHR and RED footage in general performed poorly on Metal. Due to this inconsistency, we are going to examine the performance of our two PC workstations relative to both OpenCL and Metal on the Mac, as well as a hybrid using the fastest option on each individual test.

With that said, compared to the Mac Pro with OpenCL, we found that our Core i7 6900K workstation (which is almost exactly half the cost of the Mac Pro we tested with) was about 26% faster while our high-end workstation was about 47% faster. If we switch the Mac Pro to Metal, our Core i7 6900K workstation was about 24% faster and our Core i7 6950X workstation was about 44% faster.

If we cherry pick the fastest results from both OpenCL and Metal, however, our Core i7 6900K workstation still came in at about 14% faster and our Core i7 6950X workstation at about 32% faster.

Render Previews – VR Footage

VR might not be mainstream quite yet, but it is certainly gaining in popularity so we wanted to include it in our testing. Since it is not widespread, however, we opted to keep the results separate from our testing with standard footage. Note that we opted to test both with the built-in effects and transitions along with using the Mettle plug-in. We typically avoid testing plug-ins since it opens up a huge amount of potential testing, but in the case of VR we thought it was prudent to include some testing with Mettle as that is one of the most popular plug-ins for editing VR footage.

Overall, our Core i7 6900K workstation was about 18% faster than the Mac Pro with OpenCL while our Core i7 6950X workstation was about 39% faster. Interestingly, using Metal on the Mac Pro gives really bad performance when combines with the Mettle plug-in so both of our workstations were more than twice as fast on average as the Mac Pro using Metal.

Picking the best results between OpenCL and Metal, our Core i7 6900K workstation was only about 4% faster than the Mac Pro (so roughly equivalent performance for half the cost) while our Core i7 6950X workstation was about 23% faster than the Mac Pro.

Export to 1080p H.264

1080p might be replaced by 4K at some point, but for now it is still among the most popular resolutions to export to. Unlike the 4K and 8K export testing we will be doing in later sections, we opted to stick with only 4K source footage rather than going all the way up to 6K or 8K since if you working with 8K footage you probably are not going to be terribly concerned with 1080p exporting performance.

When exporting to 1080P H.264, we found that our Core i7 6900K workstation was about 39% faster than the Mac Pro with OpenCL while our Core i7 6950X workstation was about 70% faster. The Mac Pro with Metal results are skewed a bit due to the overall poor performance with the RED codec when using Metal, but on average our Core i7 6900K workstation was 63% faster while our Core i7 6950X workstation was just under twice as fast as the Mac Pro with Metal.

Looking at whichever Mac Pro result was fastest for each individual test (OpenCL or Metal), our Core i7 6900K workstation was still about 23% faster than the Mac Pro while our Core i7 6950X workstation was 50% faster on average.

Export to 4K H.264

With the rise of 4K, we opted to spend a good portion of our testing on exporting to 4K using a wide range of source resolutions and codecs. In addition, while we will focus on exporting to 4K H.264 in this section, we will also look at 4K H.264 for VR as well as 4K DNxHR HQ 8-bit in the next two sections.

When exporting to 4K H.264, we found that our workstation with a i7 6900K and GTX 1080 was about 30% faster than the Mac Pro with OpenCL while our Core i7 6950X workstation with a GTX 1080 Ti was about 46% faster. Metal once again gave us poor performance with RED footage, but it was also worse than OpenCL for 8K footage in general. This means that compared to the Mac Pro with Metal our Core i7 6900K workstation was on average 85% faster while our Core i7 6950X workstation was over twice as fast as the Mac Pro. 

Giving the Mac Pro the best chance possible (switching between OpenCL and Metal depending on performance), our Core i7 6900K workstation was still about 25% faster than the Mac Pro while our Core i7 6950X workstation was 41% faster.

Export to 4k H.264 VR

VR is gaining in popularity and one of the most commonly used combinations of resolution and codec is 4K H.264. As we stated in the "Render Previews – VR Footage" section, we tested using both the built-in effects and transitions as well as the Mettle plug-in. We typically avoid testing plug-ins since it opens up a huge amount of potential testing, but in the case of VR we thought it was prudent to include some testing with Mettle as that is one of the most popular plug-ins for editing VR footage.

When exporting to 4K H.264 for VR content, we found that the Core i7 6900K workstation was a healthy 45% faster than the Mac Pro with OpenCL while our Core i7 6950X workstation was 71% faster. OpenCL was actually faster across the board than Metal, but if for you are using Metal when exporting 4K VR content, we found that our Core i7 6900K workstation was about 56% faster on average while our Core i7 6950X workstation was about 85% faster than the Mac Pro with Metal. 

Since OpenCL was overall much better than Metal, even if you cherry pick the Mac Pro results our Core i7 6900K workstation was still 44% faster and our Core i7 6950X workstation was 71% faster than the Mac Pro.

One quick thing we want to point out is that if you work with 4K H.264 footage, the PC workstations perform much better than the average suggests. In this case, our i7 6950X workstation is able to export almost twice as fast as the Mac Pro while our i7 6900K is only slightly slower – about 80% faster than the Mac Pro.

Export to 4K DNxHR HQ 8-bit

While H.264 might still be one of the most common codecs to export to, it is not a great choice if there is additional editing that needs to be done in a separate application. Since PC workstations do not support exporting to ProRes due to Apple's licensing of that codec, DNxHR is a popular alternative that we wanted to make sure we tested.

Interestingly, while our PC workstations were still significantly faster than the Mac Pro when exporting to 4K DNxHR HQ 8-bit, the difference is not as pronounced as it was when exporting to H.264. In fact, the Mac Pro with Metal was actually faster than our PC workstations when the source footage was also 4K DNxHR HQ – although only by a couple of seconds.

With that said, compared to the Mac Pro with OpenCL our Core i7 6900K workstation was 18% faster on average while our Core i7 6950X workstation was 31% faster. While Metal performed well in some cases, the poor results with RED footage made our Core i7 6900K workstation on average 78% faster and our Core i7 6950X just under twice as fast as the Mac Pro with Metal.

If we look at the best results from the Mac Pro between OpenCL and Metal, our Core i7 6900K workstation was 13% faster than the Mac Pro while our Core i7 6950X workstation was 26% faster.

Export to 8K H.265

8K video is not common quite yet, but there are already a number of people working with 8K capable cameras which is why we wanted to include at least some tests for exporting popular codecs to 8K H.265.

While we only tested three different source codecs, there is still some very interesting information to be had in our 8K testing. Compared to the Mac Pro with OpenCL, our Core i7 6900K workstation was about 41% faster and our Core i7 6950X workstation was about 43% faster. Metal once again had issues with RED footage which made both our PC workstations over twice as fast on average compared to the Mac Pro with Metal when exporting to 8K H.265.

If we take the best results from the Mac (OpenCL vs Metal), our Core i7 6900K workstation was about 39% faster and our Core i7 6950X workstation was about 41% faster.

Live Playback – Maximum Resolution

Performance when exporting and rendering previews is important, but being able to play your footage live at high resolution is equally important. This a challenge for us to accurately test since the resolution you can play at (full, half, quarter, etc.) is highly dependent not only on your source footage but also what effects you have applied to the timeline.

To try to keep things universally applicable, we opted to test our source footage with two relatively simple timelines. The first test we did was to see what playback resolution we were able to play our clips (with no dropped frames) without any effects or transitions applied. After that, we did the same testing only with a Lumetri Color Correction effect applied. Note that if we dropped any frames (even if it was just one or two at the very start of playback) we called it an unsuccessful run even if the rest of it was completely smooth.

It is hard to pull any generalities from this kind of testing, so we advise that you look specifically at whatever type of footage you work with if we happened to test it. Still, there are a few conclusions we can draw from the results. First, Metal on the Mac Pro is not really very good for live playback performance of clips without any effects. It did as well as OpenCL five times, but one of those was when the Mac Pro was not able to successfully play the timeline without dropping frames even at 1/16 resolution. In exactly half of our testing, we had to turn down the resolution one extra step compared to OpenCL in order to play the timeline without any dropped frames.

Between the Mac Pro and our PC workstations, the Mac Pro (with OpenCL) did about as good as our Core i7 6900K workstation. The Mac Pro did better with 4K ProRes 4444 footage, but the Core i7 6900K did better with 8K ProRes 4444 footage so we would call these two systems a wash from an overall playback of basic footage standpoint. The Core i7 6950X system did a bit better and was either better or matched the other systems in all but one of our tests (4K ProRes 4444).

Moving on to our testing with just a Lumetri Color Correction effect applied, the overall results are largely the same. The resolution we were able to play with each system certainly dropped (which is expected since it is a more challenging timeline), but we saw many of the same trends.

Once again, using Metal on the Mac Pro did not fare particularly well although it was able to match the Mac Pro with OpenCL in six of our ten tests. The Mac Pro with OpenCL was again about the same as our Core i7 6900K workstation either matching or beating the other systems in seven of the timelines. Just like the basic clip testing, our Core i7 6950X workstation matched or bettered the other systems in nine of the tests with the Mac Pro only beating it on 4K DNxHR HQ.

Warp Stabilize 4K H.264

While exporting and rendering previews may be the easiest and most common thing to benchmark in Premiere Pro, we have received a lot of feedback that performing a warp stabilize is another task where high performance is important. We have found that the time it takes to complete a warp stabilize analysis does not vary much on the source codec, but the resolution does make a big difference so we will be testing with a 4K H.264 clip as well as a 8K ProRes 4444 clip in the next section.

Since warp stabilize is not well threaded (it does not take great advantage of multiple CPU cores), one method people use to decrease the time it takes to complete the analysis is to split their clip into multiple parts and analyze all of them at the same time. This essentially forces Premiere to make more effective use of all the CPU cores in your system. Since this is trick employed fairly commonly, we decided to not only time how long it takes to apply a warp stabilization effect to a single 10 second clip, but also how long it takes if the clip is split into 2, 4, 8, and 16 "sub-clips" that are all analyzed at the same time.

Note that since the analysis portion of warp stabilize only utilizes the CPU, we did not include testing results for both Metal and OpenCL on the Mac Pro. We did compare the two but found no difference between them so for simplicity's sake we decided to just report generic Mac Pro results.

Warp Stabilization is a very interesting test, especially when we split the clip up into multiple sections. With just running the stabilization on a single clip, the Mac Pro actually holds up fairly well and is only a few percent slower than our PC workstations. However, if we divide the clip into two or more sections the Mac Pro falls pretty far behind. The exact amount changes based on the number of "sub-clips" you create, but compared to our i7 6900K workstation it ranges from either matching performance (with 16 clips) to being 78% slower (with 4 clips). However, even dividing up the clips the Mac Pro is not able to beat our i7 6950X workstation. At best the i7 6950X workstation was 17% faster while at worst it was 87% faster.

If you take all of this as an average, our Core i7 6900K workstation was overall about 27% faster than the Mac Pro while our i7 6950X workstation was about 39% faster.

Warp Stabilize 8K ProRes 4444

Interestingly, the Mac Pro actually did better on our 8K warp stabilization test than it did in the previous section with 4K. This time, our i7 6900K workstation was only about 9% faster on average while our i7 6950X workstation was 13% faster.

Keep in mind that this is just the average of what we tested in this section. If you look at the individual tests with the different number of "sub-clips", the Mac Pro is actually able to match our i7 6900K workstation if you divide the 8K clip into 8 or more sections. However, even doing so our i7 6950X still comes out ahead – although only by about 7-8%.

Miscellaneous Tasks

To round out our testing, we decided to look at a couple of tasks to try to get a feel for how much faster PC workstations are for some of the general day-to-day tasks in Premiere Pro. We only tested how long it took to launch Premiere Pro and open a project file as well as how long it took to open the export window, but if any of our readers have a suggestion for other tasks they would like us to test please let us know in the comments section!

Neither of these tasks utilize the GPU so we can compare the Mac Pro to our PC workstations without the need to differentiate between OpenCL and Metal. Since these tasks are completely different from each other, however, we will look at both individually.

Starting with the time it takes to launch Premiere Pro and open an existing project, we timed how long it took to open a project file directly from the file explorer (or Finder on the Mac Pro) until all media was loaded and the preview window showed the media correctly. We actually timed all ten of the project files used in our export and preview testing, but decided to report on the average since there was not a significant difference between each project. With that said, both of our workstations are able to launch Premiere Pro and get you working on your project about 30% faster than the Mac Pro.

Launching the export window only takes a handful of seconds on the Mac Pro, but it is a good example of how much more responsive a modern PC workstation can be compared to the Mac Pro. In this case, it took the Mac Pro about 3.7 seconds on average to launch the export window, it took our PC workstations just a hair over a second. This is looking at the first time you open the export window after launching Premiere Pro (since subsequent openings will be much faster), but for this task our PC workstations are over three times faster than the Mac Pro.

Conclusion

To get the obvious out of the way: yes, a modern PC workstation is significantly faster than the current Mac Pro. Even without our testing almost any Mac or PC user could tell you that with confidence since the hardware in the Mac Pro is over four years and two or more generations out of date. However, with the large differences we saw between using Metal and OpenCL on the Mac Pro and the difficulty of assessing live playback performance, giving a firm "PCs are X% faster" is actually pretty hard.

In an attempt to make it easier to quantify the Mac vs PC performance argument, we are going to first make the assumption that you would be using a Mac Pro with OpenCL rather than Metal. Yes, Metal was slightly faster than OpenCL in some situations – but in others it gave downright terrible performance. So unless you are willing to switch between Metal and OpenCL based on what you are doing and what codec you are using, as well as periodically testing to make sure Metal actually is faster, right now OpenCL is clearly the better overall choice for Premiere Pro on Mac.

With that assumption made, we can create a chart using the averages of our testing to show how much faster a PC workstation should be for different tasks in Premiere Pro CC 2017:

Keep in mind that we are comparing a top-end Mac Pro (late 2013) which retails for about $9,399 to two of our workstations that cost $4,930 and $6,422. This isn't simply the raw parts costs of our workstations either, this includes all of our standard markup to cover pre-sales consultation (to ensure you are getting exactly the right hardware), our full production process (including assembly, OS/software installation, and our thorough burn-in and QC process), as well as our terrific post-sale support and repair service should you ever have a problem. Even with all of that, the two PC workstations we tested still came in at 52% and 68% respectively of the cost of the Mac Pro. Even if budget isn't a concern, this cost savings can give you significant financial leg room for things like upgrading monitors or increasing internal storage capacity.

If you happen to be one of our more common Premiere Pro customers with about a $5,000 budget, you can expect somewhere around 30% faster export times as well as 20% faster preview render and warp stabilization times by switching to PC. The system should also be much faster for miscellaneous tasks like opening and saving projects although it will only roughly match the Mac Pro in terms of live playback performance. It will be better for a few codecs and resolutions but we will freely admit that the Mac Pro will be better in a few others.

If you have any comments or suggestions, we highly encourage you to leave them in the comments section. We are a PC workstation manufacturer, but we did our best to be as accurate and fair in our testing as possible. We believe that publishing misleading test results can only have a negative impact on everyone (ourselves included) but we heavily rely on our readers to let us know if any of our testing isn't as logical or accurate it could possibly be.