Infinitely scale up your…budget?

Posted on April 18, 2016 in Blog News

In the news today, Autodesk announced a partnership with Google to incorporate cloud rendering services directly into Maya. Google bought out a portion of this capability when they acquired Zync not too long ago. This move doesn’t surprise us. For starters, Autodesk has proven they don’t really know how to do cloud rendering, so they had to outsource it to someone already in that space. When they looked for partners, I’m sure they figured there was only one possible partner out there that fits their…ahem…girth. Big companies partnering with other big companies isn’t news, though, as it happens all the time. Google was the only “big company” Autodesk thought they could partner with. Time will tell, but this will probably prove to be a “big” yawn.

We still have many people asking us about our review and comparison against Amazon’s Ec2 compute cloud. Although their prices continue to inch downward, they’re still a ways off from our highest Power level (and hence most expensive rate). With today’s news, we thought it would be great to do a slightly similar comparison of the Autodesk/Google/Zync-automaton (hereafter referred to as AGZ) to Pixel Plow from just a couple perspectives. First, let’s look at software selection.

AGZ supports:
Maya (with V-Ray, Arnold, Renderman, and 1 plug-in)
Nuke
<queue the crickets sound>

Pixel Plow supports:
Way too much to list here, see our Render App Support page and expand a few of the render apps
<queue triumphant applause>

Score one for Pixel Plow. Now lets talk turkey.

To even begin to compares rotten apples to oranges, we’ll need to examine what the AGZ “vcpu” thing is all about. From their own page, a vcpu is:

…a virtual CPU is implemented as a single hardware hyper-thread on a 2.6GHz Intel Xeon E5 (Sandy Bridge), 2.5GHz Intel Xeon E5 v2 (Ivy Bridge), or 2.3 GHz Intel Xeon E5 v3 (Haswell)

OK, so an AGZ zync-16vcpu-32gb node would have, effectively, 8 CPU cores and 8 Hyperthreads running at around 2.5Ghz. Let’s head over to the Pixel Plow Pricing page and plug those values in (as though they were on-hardware performance…even though they’re not) to simulate a 100 frame job that takes an AGZ node 1 hour to render. Notice as you move the Power slider around, we show you the effect on job cost. There is a corresponding effect on job completion time, but we can’t simulate that very effectively yet.

The important part is that the job will cost somewhere between $11.52 and $144 depending on the Power level you choose. Put the same values into the pricing estimator on AGZ’s site choosing Renderman as the render engine and you get a job cost of $121. Picking Arnold as the render engine bumps that up to $208. Woof. Pixel Plow is competitive to crushing at even our most expensive rate. At our low-to-mid Power levels, AGZ has put itself out to pasture.

Has Zync (now AGZ) been used for major movie projects? It surely has. Many big budget films have utilized their platform to render some big projects, with the key being “big budget”. Imagine how much more profitable those same films could have been if they used Pixel Plow instead?

We’re not blaming Autodesk for their selection of Google/Zync as a partner for a cloud rendering platform. They wanted someone in the industry with some apparent size, and the clout that goes along with their size. No other render farm service, including Pixel Plow, looks to the fleeting glance of an executive as having size. We fly under their radar because they don’t really want to be efficient, they just want to be big. We congratulate you, Autodesk, for getting…err…bigger. Does your new render farm partner make your butt look big? Yes, yes it does, and we’re not afraid to say it.

If you’re like us, with efficiency and capability winning your attention, then Pixel Plow is for you. Make the smart choice, not the big choice, and join us today. Your wallet/purse/accountant will thank you.

Render Endlessly!

Facebook Icon Twitter Icon Visit Our Vimeo Page