Nostalgia - we all get a bit of it from time to time, wouldn't you agree? It's that time when all that is now old seems more appealing than a free pass to a Victoria's Secret or Mister Venezuela (let's cover the entire taste spectrum) event, and when the mothballs are cast away in shame. For those looking to earn bonus points, mentions of the "Great Wheel of Reincarnation" are uttered. The sad state of present affairs is decried publicly and quite vocally. Which brings us, neatly and on the wings of reasonably dubious prose, to the hook of the day: what will happen with our good friend DirectX and its annoying relative OpenGL, now that people think writing to-the-metal ASM is cool again?
This latent curiosity got a rather swift kick in the seat of the pants when, earlier this year, none other than Richard Huddy himself proclaimed that the evil APIs are keeping GPUs from going firmly over 9000, and they need to go away so that a new golden age of hardware specific non portable code can come! Ok, the last part is not due to Mister Huddy, but rather to our fearful nature, seeded by memories of the good old days of having to waddle through a billion or so different, incompatible, hardware accelerated versions of a particular game, each sponsored by another IHV and fit more snugly to its hardware than a pair of pants on Miss Kim Kardashian. Our recent experiences, when even perfectly valid DirectX can break between IHVs, don't help either.
Casting all fear aside, and donning a memory deprived geek's hat, it's not hard to see that the disdain for API overhead is not without its merits, and that with the soaring focus on treating GPUs as more general purpose compute machines, as opposed to pushers of pixels, the implementation of more daring approaches would require tickling the silicon fancy somewhat more intimately than ID3D11Device and ID3D11DeviceContext. So, we wanted the truth, irrespective of whether or not it was manageable!
After much deliberation and in-depth analysis that took about 30 minutes, we reasoned that the best way to get to the bottom of this was to go and ask people from the front-lines a (large) number of not necessarily easy questions. Indeed, what better way to guess at the shape of the future than ask the actual shapers? What follows is the (partial) result of that outreach, with hopefully even more such goodness to come as more people decide to come forth with their answers.
This latent curiosity got a rather swift kick in the seat of the pants when, earlier this year, none other than Richard Huddy himself proclaimed that the evil APIs are keeping GPUs from going firmly over 9000, and they need to go away so that a new golden age of hardware specific non portable code can come! Ok, the last part is not due to Mister Huddy, but rather to our fearful nature, seeded by memories of the good old days of having to waddle through a billion or so different, incompatible, hardware accelerated versions of a particular game, each sponsored by another IHV and fit more snugly to its hardware than a pair of pants on Miss Kim Kardashian. Our recent experiences, when even perfectly valid DirectX can break between IHVs, don't help either.
Casting all fear aside, and donning a memory deprived geek's hat, it's not hard to see that the disdain for API overhead is not without its merits, and that with the soaring focus on treating GPUs as more general purpose compute machines, as opposed to pushers of pixels, the implementation of more daring approaches would require tickling the silicon fancy somewhat more intimately than ID3D11Device and ID3D11DeviceContext. So, we wanted the truth, irrespective of whether or not it was manageable!
After much deliberation and in-depth analysis that took about 30 minutes, we reasoned that the best way to get to the bottom of this was to go and ask people from the front-lines a (large) number of not necessarily easy questions. Indeed, what better way to guess at the shape of the future than ask the actual shapers? What follows is the (partial) result of that outreach, with hopefully even more such goodness to come as more people decide to come forth with their answers.