More performance from Mac Pro for long form

Hey Gang,

I’m looking into speeding up my rendering process on a lot of the work I am doing. I’m (unfortunately?) running a simple Mac setup (Radeon Pro Vega 2 32GB, 192gb DDR4 & 3.2GHz 16-Core intel) and would love your advice on how I can process this work faster.

My current advice seems to be effectively buy another system and work two machines, which is a possibility, but your thoughts, as ever, are
much loved!

1 Like

Beren –

Can you share details of where you feel your work is hitting a bottleneck? Big Batch setups, large frame sizes, complex timelines?

Hi Chris,

So, i’m working as a VFX Editor and Flame Artist on a large Amazon long form drama. Only HD (from transcodes at the moment) I am performing a lot of work from 3D objects (cars, flies, clouds etc), to machine learning timewarps, splits, screen comps, 3D tracking for advanced clean-up. At this stage, a medley of everything.

Back-end, I’ll be performing a lot of the final screen comps, splits, beauty work, timewarps, all 4.5k 16-Bit EXR.

Also, apologies, whilst I am not necessarily hitting a bottleneck yet, I certainly pre-empt hitting them. Specifically with camera analysis and multiple 3D objects within the scene

I would advise getting a new system. A Linux box on Lenovo P620. This would give you the fastest machine with a reasonable cost. Apart from sheer rendering speed it would speed up your manual work as well being snappy and all. You could share your framestore between the machines and leave some long sequences to render on the Mac, if you get another license. Not sure how Flame + Flame Assist/Flare licensing would work on a system like this.
Good luck with the project, and keep in mind that you can ask for all sorts of help here…


Agreed with @Sinan. For half the price of your current machine you could have a P620 that is almost 3 times faster.


OK thanks kindly gentlemen! Honestly greatly appreciated, I’ll speak to my engineers on this side and see what we can get built. Greatly appreciated

1 Like

Burn can really help, too. If you can set up a linux box for background rendering, you’ll really be able to speed through the work.

Beren I can’t read anything about your frame storage but Finn wrote about NVMe on Mac Pro in the past and if you don’t use this, may be you can speed up your Mac by just using a faster frame store if you currently don’t?

NVMe Raid 0 speed topic on Mac Pro

1 Like

Hi Sam,

Thanks for the heads up! I’m not sure it’s an option at the moment, as my Linux knowledge is very limited, however i have put out a few feelers to freelance engineers about the possibility of supplying the Linux support. Thanks again!

I got a quote for a fully configured burn node from my reseller. A lot less expensive than a whole flame, and with 10G-Ethernet, a very quick renderer.

Didn’t know you could use Burn with Mac flames. Good to know.

1 Like

I didn’t think you could.

You can’t run a burn node on a Mac, but you can send burn renders from a Mac to Linux nodes.


A lot of interesting thoughts here. I am likely going to veer towards getting another Mac setup, to work with someone alongside myself. Effectively just holding a Flare licence, along with Mocha and some of the other 3rd parties. Just got to keep an eye on the new Apple M1, I guess!

Ideally i’d go Linux world, but my knowledge is very limited and I’ve not really the contacts and support to assist me over this way in Edinburgh. Lots to ponder!

1 Like

One more thought Beren, are there any rental options for Flame near you? If so, renting a well-equipped Linux system for the duration of the show would be a nice way to compare the performance difference without committing to a Mac that may be a compromise with M1 Pro machines on the way.

Also, any proper rental house will offer full support in helping you getting their system integrated with yours.

Good luck whichever way you go!

Aye, 'tis a great shout, and so of how I imagined it would go. Unfortunately I do not have the proper support for that set up just yet. Alas, one day!

Thanks Chris!