Weird iMac 2015 vs 2019 results

[Jamers99]

I primarily use my 2015 iMac for editing and rendering videos shot on my 4 year old Sony Handycam PJ340 set to 1080p video resolution. Not sure exactly what codec it is but according to FCPX it's AVC1. I did some side by side editing and rendering. They editing felt virtually the same in FCPX between the 2 systems. I performed a few rendering tests and the 2019 only edged out the 2015 by 23 seconds when I rendered a 2 minute video clip. I calculated it as a 22% difference. My render file output settings are shown below. I guess was expecting the render time to be cut in half considering the benchmarks on the 2019 are nearly twice that of the 2015. I even did a boot test the the 2019 only beat the 2015 by 4 seconds for startup.

What's more, I used my IR gun to take some readings on the back near the exhaust fans during rendering and the 2019 was easily 20 degrees fahrenheit hotter than the 2015. The 2019 fans kicked up to every bit as loud as the 2015 fans (to my ears). Also, I noticed the fans kicked in loud on the 2019 simply importing media from flash cards in FCPX whereas the fans remained silent during the same process on the 2015.

I decided to upgrade my iMac because all the reviews were claiming it quieter, cooler, and faster. For what I'm doing it's not really any of those. Quieter="Not really" Cooler="Not according to my IR Gun" Faster="A mere 22% for the type of video I was encoding"

Perhaps I would have better performance results with different video codecs or maybe when using 4k files. I haven't tested that yet but for the extra $1750 out of pocket it is costing me to upgrade to the 2019 I'm starting to think it's just not work it. There's only 2 TB3 ports, no BT 5.0, no cooling improvements, and now I hear that Apple is coming out with a standalone 31.6" display which will likely make it's way into a completely redesigned 2020-2021 iMac with slimmer bezels and possibly no or smaller chin.

What do you guys think?

Old 2015 iMac
4.0 Ghz i7, 24gb ram, 512 SSD, Radeon R9 395x 4 GB

New 2019 iMac
3.6 Ghz i9, 16 GB Ram, 512 SSD, Radeon Vega 48 8 GB

 

[SecuritySteve]

To be clear, you are upset that the performance gains you received are not as high as you initially perceived they would be?

Your last paragraph expressing being upset about aesthetics makes me question if you really care more about the performance, or if you simply wish that the iMac had been given an aesthetic touch up.

To me, 20% improvement over 2 years is pretty good. With the exception of GPU performance, most computer hardware is not doubling in performance anymore.

I'm sure a 2021 iMac will look better, and if you can get by with a 2015 iMac I wouldn't have bothered with the 2019. I tend to believe computers last 3-5 years depending on how heavy your usage is. Replacing faster than that is only warranted if your needs change to be more extreme.

Just my thoughts.

 

[Jamers99]

To be clear, you are upset that the performance gains you received are not as high as you initially perceived they would be?

Your last paragraph expressing being upset about aesthetics makes me question if you really care more about the performance, or if you simply wish that the iMac had been given an aesthetic touch up.

To me, 20% improvement over 2 years is pretty good. With the exception of GPU performance, most computer hardware is not doubling in performance anymore.

I'm sure a 2021 iMac will look better, and if you can get by with a 2015 iMac I wouldn't have bothered with the 2019. I tend to believe computers last 3-5 years depending on how heavy your usage is. Replacing faster than that is only warranted if your needs change to be more extreme.

Just my thoughts.

It's actually a 3.5 year upgrade since my old iMac is a Late 2015. I guess when taking everything into consideration I'm better off with the 2015 for now. This 2 day old 2019 iMac will be going back to Apple. I'll save my money for the next iMac upgrade.

And with respect to CPU performance my geekbench scores went from ~17581 to 32398 which at 84% is nearly double the performance. Like I said, I guess I was expecting much more improvement on the video rendering times.

 

[SecuritySteve]

It's actually a 3.5 year upgrade since my old iMac is a Late 2015. I guess when taking everything into consideration I'm better off with the 2015 for now. This 2 day old 2019 iMac will be going back to Apple. I'll save my money for the next iMac upgrade.

And with respect to CPU performance my geekbench scores went from ~17581 to 32398 which at 84% is nearly double the performance. Like I said, I guess I was expecting much more improvement on the video rendering times.

My apologies for some reason I was thinking 2017 iMac instead of 2015. Must not have been fully awake yet.

 

[Bohemien]

And with respect to CPU performance my geekbench scores went from ~17581 to 32398 which at 84% is nearly double the performance. Like I said, I guess I was expecting much more improvement on the video rendering times.

Just checking:

• Are you using the newest version of FCPX (because newer versions will be optimized for more cores)?
• What about other tests, like converting a video with Handbrake, is the performance gain in the same range?
• Maybe you could check Activity Monitor, your new iMac might be performing some "new machine" tasks in the background, like indexing or even encrypting the drive, that eat up performance.
• And finally, you have loaded both machines with RAM, but still you could check the RAM usage, 16 vs. 24GB might make a difference for rendering.

Not saying you're doing anything wrong, just want to make sure you're getting valid results before coming to a decision. After all, both processor and GPU should have a significant performance bump over your old machine, you're right about that IMHO.

 

[Jamers99]

Just checking:

• Are you using the newest version of FCPX (because newer versions will be optimized for more cores)?
• What about other tests, like converting a video with Handbrake, is the performance gain in the same range?
• Maybe you could check Activity Monitor, your new iMac might be performing some "new machine" tasks in the background, like indexing or even encrypting the drive, that eat up performance.
• And finally, you have loaded both machines with RAM, but still you could check the RAM usage, 16 vs. 24GB might make a difference for rendering.

Not saying you're doing anything wrong, just want to make sure you're getting valid results before coming to a decision. After all, both processor and GPU should have a significant performance bump over your old machine, you're right about that IMHO.

Where is your evidence that newer versions of FCPX are optimized for more cores? I looked thru the version history release notes and there isn't one reference to cores or cpu. Which version instituted this optimization? I don't use handbrake, I use FCPX so my tests were limited. But it doesn't matter to me what other program might do, I only use FCPX. Nothing else was running in the background. I doubt 16 vs 24GB would have any measurable affect on the temperature or performance in this instance. When I also consider that it's only got (2) TB3 ports, only Bluetooth 4.2, huge bezels and chin, 720 FaceTime camera, subpar cooling system, I have made the decision to wake for a real redesign. Honestly, if at least had (4) TB's and BT 5.0 then that make have been just enough for me to justify it and keep it. But also knowing a future redesigned iMac will likely have more TB3 ports, better cooling and a 31.6" 6k display with minimal bezels then I decided I can wait.

 

[Bohemien]

Where is your evidence that newer versions of FCPX are optimized for more cores?

No evidence, just supposing after the release of the iMac Pro in 2017 they would've done so to utilize the Pro's higher core count. After all, the FCPX user base will be the Pro's primary target audience.

Just trying to understand your results... but you seem to have come to your conclusion anyway, so never mind.

 

[bondavi]

I'm assuming you did a migration assistant install of data between iMacs including FCPX as opposed to a fresh standalone install of the latest version w/ updates?

This is what the i9 does in Cinebench, pretty respectable.

If the former is true, I would uninstall completely and reinstall.

 

[Jamers99]

I did a fresh install of FCPX. Yes the benchmarks are impressive but the real world rendering results weren't as good for the type of video I was working with. Plus it still gets pretty hot. It's back in the box now to return tomorrow.

 

[priitv8]

Can it be, that with the setting you use (Better Quality), FCPX does not use HW accelerated encoding and uses CPU only?
In that case 22% speed increase is not insignificant.
Would be interesting to know, if Intel QuickSync HW-accelerated (Faster Encode) has similar or bigger speed difference?
Source: https://www.idownloadblog.com/2016/01/09/how-to-achieve-faster-video-exports-in-final-cut-pro-x/

 

[macduke]

I doubt you'll notice a lot of difference with footage as outdated as 1080p. I mean, even the iPhone has had 4K going way back to the 6s. I'm pretty sure my ancient iPad 2 could edit 1080p in iMovie, lol.

To put this into perspective, it's kinda like someone sending back their 2012 iMac because it didn't open email noticeably faster than their 2008 iMac. But if you don't need it, you don't need it, just like someone back then could have kept a 2008 iMac around for longer if they were doing such low-power tasks. You made the right choice. That being said, something still seems off with your results and I wonder if it could have been better optimized somehow via the settings or something that needs to be fixed with regard to how the software utilizes the new hardware for 1080p rendering on that codec.

For someone like me who was coming from a 2012 rMBP at home, it was a big upgrade. And yes, compared to the 2017 i7 iMac I had at work at my old job, the 2019 iMac is much quieter and cooler. If you don't understand then you didn't have one, lol. Be glad for that! But even if it does occasionally get a little louder during renders, I'd rather have that than a CPU that is throttled to hell. You just can't expect to get whisper quiet beast mode performance out of a standard iMac. Apple does sell a quieter professional iMac for users who care about that.

Also, about that 31.6" display, don't expect it in the standard iMac anytime soon. I expect it will first come out as a standalone product sold alongside the modular Mac Pro. Then it will likely come to the iMac Pro as a step-up model, probably around $1000 more than the 27" model. So be prepared to pay $6000 for a 31.6" iMac Pro. Then a few years after that, it may eventually come to the standard iMac and replace the 27" model as top-tier and the 27" model will become the cheap model and both models will probably see price bumps which they will justify with the bigger displays while increasing their ASP which Apple loves doing. I think Apple wants to differentiate the iMac Pro and iMac even more, and I think that display will be key. I think in the future Apple will continue to differentiate the iMac Pro with things like mini-LED with local backlight dimming, Pro Motion 120/144Hz, better color gamut, etc. Right now they are pretty much the same but knowing Apple I doubt that will continue once the new display tech arrives.