Interview: Page (1) of 1 - 04/03/15 Email this story to a friend. email article Print this page (Article printing at MyDmn.com).print page facebook

Image Engine Tames "Chappie" Renders with PipelineFX's Qube!

A Q&A with Gino del Rosario, Head of Technology at Image Engine By DMN Staff Writer
1) Can you tell us about your work on "Chappie?" What did it consist of?

Image Engine was the primary VFX facility on "Chappie", so we were tasked with all the fight sequences, hard-surface character creation, live action animation, comps and lighting jobs that Neill Blomkamp and Sony Pictures requested. Our team consisted of about 150 people working out of the Vancouver office for about a year and half.



2) Recently you announced that you employed PipelineFX's Qube! to manage your render farm on this project, what led you to using it?

Qube! has actually been with our facility for a while now; it manages the 500 machines we have in our render farm. We keep using it for a lot of reasons. First of all, it never crashes, even when we throw huge, final sequences its way. For reference, "Chappie"'s final sequence took machines with 192 GB of memory over 24 hours to render a single frame. Thinking about your render farm crashing when you have that kind of load to deliver will give you anxiety, but because we use Qube!, we don't have to feel that way. It's production proven.




Another reason is that you can build custom tools into it. One that we developed ran on all of our Linux workstations and queried artists on their machine's availability. We use our desktops for extra render farm power when we can. Using this tool, Qube! can take them into the fold and put more  machines to work. This helps the farm run at 100% capacity, night and day.



3) Do you think pipeline efficiency is challenging for most studios? How do you think they can solve some of its challenges?


Efficiency is challenging for everyone! As new tools emerge and pipelines become more complex and customized, the bar is constantly being raised. What worked yesterday might need to be rethought today. One way out of that is vigilance; you have to constantly pay attention to bottlenecks that tie up resources or prevent your studio from reaching its full potential. Another strategy - and these methods should definitely be put together as much as possible - is finding a way to automate processes. That's one reason why we use Qube!. Render farms, when they aren't optimized, can really grind things to a halt. If you don't have a tool like Qube! that notices when a render job fails and makes it instantly start again, then efficiently starts slipping and money starts getting wasted. As artists we are all really busy, so taking one thing off our checklist, especially something as big as making sure the renders are getting done, is huge. 



4) In your "Chappie" announcement, you discussed building a customized prioritization system using Qube!'s API tools. Is API accessibility essential for the modern VFX studio or is out of the box more important?


Depends on how your studio operates. If you have a customized pipeline like us, API accessibility is huge. Qube! is shared by all of our in-house production teams (since "Chappie" isn't our only project), so tweaking their own prioritization system helped us put in a lot of really useful things like categorized tags. Examples include: 911, Overnights, Overdue and Finals. With this in place, everyone across our facility knows what jobs are most important, which helps us be efficient even when a lot of different stakeholders are in the mix. "Chappie", by itself, averaged 36,107 render tasks every weekday, so you can imagine the disorder that would emerge without a render farm management system.  It's so necessary nowadays.



5)  With efficiency being such a buzzword nowadays, how do you track and assess your efficiency at Image Engine? 

Again, we like to use Qube! because it can deliver a lot of hard data about your farm. When you mine it (statistics goes into a personal database), you can profile shots, help estimate timelines for upcoming projects and even set new internal priorities a lot more effectively. It's also extremely important when you are doing the post-mortem assessments. Information like it taking us 498 days to complete "Chappie" or that we were running an average of 1,479 render jobs a day lets us benchmark how we are doing now and consider with how we can do better, all of which makes us a much stronger studio, in a strategic and operational sense.



In the future, we plan to use dashboards with the system so we can visualize the data in really easy-to-understand ways. We are going to work with PipelineFX (company that makes Qube!) on this so this feature can come to the entire community. They seem really excited about it. We are too!


Page: 1


Related Keywords:Chappie, Image Engine, Render Farm, CG, Sony Pictures, Qube!

HOT THREADS on DMN Forums

Our Privacy Policy --- @ Copyright, 2015 Digital Media Online, All Rights Reserved