In an early Halloween treat, Disney announced on Tuesday the intent to buy Lucasfilm for $4 billion dollars. Even more surprising, was the unveiling that a long-awaited, but never expected sequel to the Star Wars saga, Episode VII, is in “early stage development” to be released by 2015. While many fans are excited to see new life breathed into the Star Wars franchise, early doubters fear that Disney will do more harm than good. Regardless of opinions, there is no doubt in my mind that blade server technology will be an important part of the success of this new merger.
This week celebrated the 3rd anniversary of BladesMadeSimple, so it should be no surprise that I’m biased toward blade servers however when it comes to shear compute capacity blade servers are the best choice. Before I get into that argument, let me take a step back and provide some substance to my early comment that blade servers will help the Disney-Lucasfilm merge be successful. When we look back over the years, the advancement in technology, especially CPU technology, has helped not only improve the quality of movies being produced but also the speed at which they are released.
You probably do not know that I have had the privilege of being a member of the Disney social media group for my other blog (http://MoviesMadeSimple.Wordpress.com) and it is has been astonishing to see the rate at which Disney is releasing movies (10 movies in 2012). Compared to years past, that’s nearly DOUBLE the rate. With the announcement of Star Wars Episode VII being released in May (if they follow Star Wars movie history) of 2015, Disney has approximately 30 months to write the screenplay, cast the actors, shoot the film, create the CG effects and render. May of 2015 seems impossible – especially when you look at movies like The Avengers 2 being slated for release at the same time. If Disney fails at hitting this target, it could fuel the fire of the early doubters and cause investors to question this acquisition. While releasing Star Wars Episode VII in 30 months may seem like an impossible goal, I believe with today’s blade server architecture, it can be done. To back up this statement, let’s look at some past examples:
- Star Wars Episode III took approximately 3 years (2002 – 2005) to develop. According to my research, it was created on systems with dual core AMD 64bit (Athlon 64FX) CPUs. These systems had a maximum of 16GB of RAM and 4 cores per system. Assuming a full 42U rack of servers, the maximum capacities these rendering farms could have would have been 256 processing cores and 1TB of RAM.
- Although Avatar took approximately 4 years to create (2006 – Feb 2010), they spent 30 days of 24×7 rendering to complete. They used over 40,000 processor cores and 104 terabytes of RAM. They used blade servers with dual Intel Xeon 5400 CPUs with quad cores for a total of 1024 cores and 4TB of RAM per 42u rack. (Interesting fact – Avatar required 12 megabytes per frame, so each second stamped onto the film equated to 288 megabytes or 17.28 gigabytes per minute.)
Let’s compare the above with the capabilities we have today with a blade server:
- 32 servers x 4 chassis = 128 nodes
- 128 servers x 2 CPUs = 256 CPUs * 8 cores = 2,048 cores per rack
- 128 servers x 192GB RAM (max with 32GB DIMMs) = 24TB RAM
- 128 servers x 96GB RAM (using 16GB DIMMs) = 12TB RAM
- 128 servers x 2 – 200GB SAS Drives = 51TB local storage
With 2x the CPU processing capacity and 6x the memory capacity, not to mention capabilities of up to 40GbE network communications, blade servers are a clear winner for platform of choice in my opinion. As a last step, let’s look at performance differences:
- Star Wars Episode III: Dual Core AMD Opteron 285 – CPUBenchmark.net score of 1477
- Avatar: Intel Xeon X5460 – CPUBenchmark.net score of 4539
- Move Made Today: Intel Xeon E5-2450 – CPUBenchmark.net score of 11,494
Hopefully, if you made it this far, you understand my earlier comment. If Disney wants to meet their aggressive target of creating the next Star Wars film in under 30 full moons, they’re going to have to rely on getting a dense compute rendering farm using the most current technology. The obvious choice for me = blade servers. Good luck Disney. Thanks for bringing the Star Wars saga back to life.
Kevin Houston is the founder and Editor-in-Chief of BladesMadeSimple.com. He has over 15 years of experience in the x86 server marketplace. Since 1997 Kevin has worked at several resellers in the Atlanta area, and has a vast array of competitive x86 server knowledge and certifications as well as an in-depth understanding of VMware and Citrix virtualization. Kevin works for Dell as a Server Sales Engineer covering the Global 500 market.