Red good

Red good, green bad.

Back when I was a member of the technical press, I used to attend press briefings for new AMD (née ATI) and Nvidia GPUs. Herded by each company’s PR folks, the pre-launch briefings usually followed a well-worn formula: fly everyone either to company HQ (Nvidia) or some nice location (ATI), sit everyone down in front of presentations about the top-level of what’s new, allow the interested press some time with engineers to ask questions, then send everyone home a day or so later with a board or two to test and write about, as the clock on the embargo runs down to T-0.

Attend enough of them and you get to know many folks at each IHV, but none more so than the PR person you go through to get access to the engineers and the information you needed. Back in my day, ATI had a great guy in the UK called Andrzej Bania. I think he’s still active in the industry somewhere, so if you happen to read this and know him, let him know I said hi.

He was my go-to person when I needed to talk to ATI about something, be it samples or drivers or whatever was going on at the time. When we’d attend events as members of the press, he’d herd us groups of cats around and make sure we were in the right places at the right times, to hear the things ATI had brought us a long way to hear.

I don’t quite remember when he started saying it, but the missive at the top of this post used to come out of his mouth whenever there was a lull in conversation, and I’ve never forgotten it. Almost every time I’ve read a comparative analysis of GPUs from the two vendors in the last almost 15 years, the words have followed my thoughts around regardless of whether it was an ATI/AMD or Nvidia product that came out on top. Red good, green bad.

I guess it was subliminal in other ways too, because after seven years at PowerVR, I’m very shortly off to AMD to work inside its Radeon Technologies Group. RTG is AMD’s graphics division. And after seven years of mostly competitive analysis, with several healthy side orders of other stuff alongside, I’m switching gears to work on developer-facing graphics tech in RTG’s Games Engineering group, with the rest of its team here in Europe.

It’s been quite the ride at Imagination, home of PowerVR. I’ve watched the company grow from around 350 folks when I joined in 2010, to 1500 or so today. I saw both the Caustic Graphics and MIPS acquisitions, and helped the transition from SGX to Rogue (and beyond very soon!). I saw a CEO resign and the new guy reshape the entire company behind a new philosophy and approach to the core businesses.

The company is over 20 years old, but it really grew up over the last 5 years or so, taking on the kind of growing pains that any mid-sized company needs to deal with, fuelled in large part by the money brought in via the success of the iPhone. That huge cash influx let Imagination expand rapidly, especially in terms of head count when we bought MIPS, and rapid expansion is always hard to keep under control and do properly. As an employee it was exciting, and sometimes exhilarating. As a shareholder, though, it was often maddening. Andrew Heath, the new CEO, has swung things back around for the most part, and now Imagination is a much leaner and more focused beast that isn’t spread so thin.

I’m leaving just before another probable growth spurt, as the company regains profitability and has new success in a growing customer base. The smaller PowerVR GPU cores are now incredibly competitive against their ARM equivalents, to the point that on a purely technical head-to-head evaluation, PowerVR with Series8XE should always win today. That’s allowing PowerVR to tap into volume revenue in cheaper segments with new customers, muscling ARM out of the way in sockets that it’d never have won in the past. I hope it uses that revenue very wisely.

I’m not a shareholder any more today, my organic share awards drying up as I leave, but I’ll become a shareholder again in the not too distant future as a private investor. Things look great for them and I wish everyone at Imagination very well and lots of success in the future, especially at PowerVR.

I’m sad to leave in a lot of ways, but leave I must to pursue something new and fun to work on in this next bit of my career. I’ve already met my new team at RTG and I can’t wait to get started working with them. The change of pace from licensable mobile IP back to desktop-focused discrete products under AMD’s own Radeon umbrella is also going to be a lot of fun.

When I joined PowerVR, desktop GPUs were mostly simple in concept when compared to the TBDR philosophy that underpins PowerVR’s microarchitectures. These days the lines are increasingly blurred and the difference between GPUs across the board, mobile are not, aren’t nearly as big as they used to be.

That means I have something really fun and interesting to work on with my team. Simpler designs mean mostly simple enabling technologies, and simpler optimisation techniques. More complex designs mean more to do across the board, with much more hanging fruit to help the game developer pick. Some of it will be low, some of it will be very hard to reach. All of it will be new and very interesting.

AMD are also doing really well these days, and I’m joining as the company pulls itself out of a long rut. I’ll happily ride those coattails! Ryzen launches tomorrow as I write this and should broadly be a price-performance hit from what I can tell. On the GPU side, Radeon RX Vega is well on the way and the first thing for me to work on when I get there later this month.

So it’s a fond farewell to PowerVR and competitive analysis after 7 years, and hello to RTG and games technology.

Andrzej was ultimately right; red is good.