I recently had the privilege of presenting at the Greenlight Summit as did Tim and Alex. Well, not so much Tim. Fortunately, my internet connection did not wig out during the presentation so not only was I able to get through my presentation but I also did not suffer the awful humiliation of having my voice chipmunked by being played back at twice the normal speed (EDITOR: thanks for bringing it up again). At any rate, I thought I would share my presentation in blog form for your enlightenmentz.
There have been some interesting trends in graphics performance between mobile and console recently. NVidia, Qualcomm, and John Carmack have made some provocative predictions that mobile graphics performance will overcome current gen console performance in 2014. I think this claim merits some explanation and a discussion of its implications since it is and will continue to drive a tectonic shift in the entire multi-billion dollar game industry.
Let’s start with a chart produced by NVidia:
Comparing hardware performance is notoriously difficult as benchmark data can be conflicting between different tests and there are multiple ways of measuring performance of graphics architectures. However, it is still very valuable to overlook the detailed uncertainty when looking at big picture trends. In the graph above NVidia has created an unknown metric for performance and normalized it to 1.0, probably relative to the Xbox in 2001. The graph is logarithmic and shows PC performance increasing logarithmically, consoles more or less flat, and mobile increasing faster than logarithmically. Hence it shows mobile quickly catching up to these more traditional markets.
While of course I think NVidia is a reputable source of information I want to take a look at one specific metric that I think is very important which is texels per second. This metric can limit both the structural complexity and visual complexity of a frame. As mobile screens become larger and higher resolution I feel it is a fair comparison between mobile and console.
Below, I chart mega texels per second between Xbox, Playstation, and the iPad using a linear scale. I chose the iPad since it has the largest screen.
This seems to be telling a different story, especially considering a possible next-gen console data point (Based on a best guess from internet rumors of an AMD 7000 series GPU). Are Carmack, NVidia, and Qualcomm flat out wrong?
The key piece of data is that like the mobile industry in general, the iPad is releasing new generations of hardware every year and the performance of that hardware has been doubling every year. When considering this we get the chart below:
While mobile will probably never be on the bleeding edge of performance, it is certainly catching up very quickly and will not lag that far behind consoles.
It’s fair to question how realistic it is to expect a doubling of performance every year. I think it’s realistic given the following points: First, while there are certainly hard problems to be solved around power and size, mobile has the advantage of following in the footsteps of its larger brethren in the console and PC space. R&D dollars spent in mobile graphics do not need to be spent solving architectural problems. Second, graphics performance can be increased by increasing die size. Finally, given the explosion of the mobile market across the world, lots of R&D dollars will be focused on mobile graphics. Hence the scale of production will bring down cost making it more affordable to dedicating larger silicon to graphics. Certainly a doubling of performance will require die size and power compromises but there will be large market forces driving decisions to favor graphics performance in order to outdo rivals and satisfy consumers.
So what does this mean for the gaming industry? First, I think more traditional console developers will be entering the mobile market just as Epic has done. As they do they will naturally bring their existing production quality with them. Increases in performance will enable the quality bar on mobile games to be raised and the console developers with their highly skilled production staff will raise it. Second, mobile game users will become more sophisticated and focus more of their money on AAA titles. While these new AAA mobile game titles will expand the mobile game market they will also consume a larger percentage of the total market.
Producing a AAA title game is an expensive endeavor requiring a large, highly trained staff. As mobile consumers move to these titles it will raise the barrier to entry in the mobile games market. This in turn will make it harder for the smaller independent game studio to find the market traction and attention they are currently receiving. While there will always be room for innovation with good, fun games, it will be more difficult for new small studios to repeat recent success stories of 2 or 3 people teams raking in millions of dollars of sales. These successes were driven by finding a market early in the right environment which got them atop the app charts enabling them to ride the wave to huge profits. As the app charts become more dominated by AAA games, that opportunity will diminish.
There are serious implications for the console manufactures as well. In my opinion, they are missing a current market opportunity to refresh their hardware allowing mobile to be the story and attract the attention of content developers looking to expand their markets. Of course, console manufactures will release new console hardware which will receive attention but they risk having their market dwarfed by a massive worldwide mobile society.
In order to stay relevant, console manufactures will first need to refresh their hardware more frequently, perhaps every other year. Second they will need to continue to find competitive advantages to keep people in their living rooms to play games. Microsoft has taken the lead on this with Kinect which will probably save the Xbox platform.
The mobile game industry will mature very rapidly over the next 3 years. I think that by 2015 we will see our first $20M mobile game title. So if you’re thinking about starting a mobile game company to serve the core market, give up now, its a fools errand. Oh wait…