In exclusive testing which has never been done for Mac, the team here at Create Pro, who spend their days working on Mac upgrades, have compiled an assessment that reveals that Apple could well have been slowing down not only the iPhone, but Macs too!
There’s a whole heap of maybes in this feature, and we’d like to point out that this is very much a discussion-focused article. So, if you’re as intrigued as we were to find out more on just what all the fuss is about whether Apple’s machines are being slowed, then check out our very own tests and make up your own minds on the situation.
On 27th May 2016, Nvidia released Pascal, its latest GPU architecture. Apple has been AMD-based for years, and the last time Nvidia mentioned Mac drivers was in a post from 2015- https://blogs.nvidia.com/blog/2015/08/31/mac-driver/. On 6th April 2017, they announced a Pascal Driver specifically for the Mac community. Here’s the official 2017 Driver announcement if you needed to refresh your memory – https://blogs.nvidia.com/blog/2017/04/06/titan-xp/.
There are some ifs and buts surrounding the Mac-specific Pascal Driver, and one stands out more than most. We’re all wondering if there was an agreement with Apple that yes, they can have the Nvidia cards functioning in the Mac, but Apple-based applications have to favour AMD technology. So, was the performance slowed down? This is, of course unconfirmed, so is a definite maybe at this stage that an agreement was made with Nvidia, but from the test, we’ve conducted it certainly looks like it.
Why don’t we get to the juicy bits and let you make up your minds for yourselves from the tests we’ve done.
Reduction of Performance
In terms of reduction of performance, issues arose when we updated to
Nvidia Pascal Enabled Driver – 10.12.4 – (16E195) 378.05.05.05f0, (16E195) 378.05.05.05f02.
We carried out speed tests using a Nvidia Titan X Maxwell card which was benchmarked at an average of 22 seconds in Bruce X. From our analysis in the following operating systems: 10.10.4, 10.11.6, 10.12.1, 10.12.2, and 10.12.3, using Final Cut Pro X 10.2 / 10.3 / 10.4 where supported, we found the benchmarks stayed consistent at 32 seconds. It doesn’t take a genius to see that this is a notable 37% reduction in performance! These benchmarks did not waiver when we conducted the same test and analysed the performance for; Davinci Resolve, LuxMark, Geekbench, Cinebench and Heaven.
Conclusively, there was a link between this driver revision and how well the card performs in Final Cut. What are your thoughts on these findings? We’d love to hear what you think in the comments!