The return
Few job titles are likely to draw as much envy as Gaming Scientist, a role recently filled at AMD by Richard Huddy.
He's served stints at 3D Labs, Nvidia and ATI, but outsiders may not know this is actually Huddy's second round at the company. He first joined after AMD bought ATI in 2006 but left to "dabble" in online poker in 2011. It was a short-lived sojourn, thanks in no small part to the US' refusal to legalize it.
AMD wasn't able to open the doors to Huddy then, so it was on to Intel for a few years. However, as of June 1, 2014, Huddy is back in AMD's arms.
Huddy brings decades of experience and a palpable passion for graphics to AMD's increasingly gaming-centered table, but his most intriguing asset - a willingness to shoot from the hip about things that, frankly, seem to piss him off - could be his most valuable contribution.
He's rejoining AMD at a time when the company is embroiled with rival Nvidia over the latter's GameWorks platform. GameWorks is a set of Nvidia-developed tools aimed at helping game makers enhance their titles' graphics. AMD claims the tools deliberately handicap its products, making Nvidia's look superior by comparison. Nvidia has staunchly denied AMD's accusations.
TechRadar asked Nvidia for comment on the situation, but haven't heard back. We'll update this story if and when we do, but until then read on for more on Huddy's role and what he has to say about GameWorks.
TechRadar: Let's get this one out of the way. What does being AMD's Gaming Scientist entail?
Richard Huddy: I sit in the office of the CTO, reporting to Raja Koduri. It's my job to go out to ISVs, to the games developers, because my focus is on 3D. My first focus is on discrete graphics, the high-end graphics solutions. I talk to those game developers in that area; Crytek, DICE, Epic, all the people who produce quality graphics engines on PC, and bring in information on what it is they want and how they're trying to change their rendering engines.
I bring in an understanding of what's needed to solve the set of problems they're confronted with at that point, and then in the office of the CTO, I get a seat at the table when we are designing our GPUs, when we're making prioritization decisions, how much we spend on certa in features and whether they're in or out. Because of the formalization of the process, it means that AMD guarantees that in our GPU designs, gaming is front and center.
That's emphatically the message from the company as a whole. Gaming is absolutely fundamental, and it's not going to go away from us. The visual experience, the stuff that we have that's almost magical in its ability to represent 3D worlds so quickly and at such a high quality, that's a big deal for AMD.
You're going to see me out there a lot as well. Although my primary role is going to be this communication with ISVs and discussion inside the office of the CTO, there is no doubt that I quite like getting out there and talking to people like yourself. AMD needs to get the message out there.
TR: What do you think AMD can stand to improve? Is there a product or area you want to address in your new role?
RH: There are a couple of things going on in the business that give us opportunities there. I'm not sure if you've written about Mantle [editor's note: I have]. There are some really surprising numbers in there. We at AMD are the authors of Mantle. It doesn't belong to anyone else. There isn't anybody else's IP involved though we will make it an open standard. And yet we have 47 registered developers, seven of whom are public and 40 are through a private, email beta. Forty-seven is a very big number in the PC market place.
Bear in mind that's the preponderance of people who care about graphics deeply. There are some for whom graphics is a side issue; it's just a way of showing the game. And there are others for whom graphics are the differen tiating feature. The vast bulk of those are involved with working with Mantle.
We couldn't get 47 people to sign-up for it if it wasn't a solution to a problem that they care about. We could maybe pay 5-10 people to join in, but that's a very different kind of situation than the one we've got.
And then there's a really surprising number; with DX 11, from it's arrival which I think was in October or November 2009, if you take it to its first birthday, there were nine games which used DX 11 in that year, and that includes folks who patched their previous games and so on. They get nine titles in that year. We've already got nine declared titles that will come in 2014 and I suspect that it will be more like 12.
I think it's inherently surprising that AMD can put together a collection of developers that's actually as big, or in this case marginally bigger than what Microsoft achieved in their first year. I think it's a testament to the fact that we really are so lving a problem that games developers want.
Fighting words
TR: AMD, at least in talking about Mantle, has been focused on conveying a sense of openness to the industry as of late. Is this arms-open approach new for the company or as it always been like this?
RH: I think it's true of AMD that it's kind of in our DNA. That's the kind of the phrase that all companies use when they talk about something which lasts for awhile. But it really is. It's very fundamental to us. We believe in open and fair competition. We believe in exposing things through open standards. So, for example, if I look back, the 64-bit x86 processors that we use in our PCs today, the x86 stuff was invented by AMD and put out as a standard.
We have worked with monitor vendors to improve the quality of gaming monitors so you don't get the kind of visual tearing that's around on monitors at the moment. That will come to market at the end of this year or early next year, and that's coming to market as part of a standard.
I think that we're good listeners. We certainly aim to be good listeners and we absolutely work with the benefit of the ecosystem in mind. I guess everyone's tempted to say that - I don't know whether that sounds terribly cogent and you feel persuaded by me saying that. I don't know, but the truth is we really do work that way. Open standards are a big deal. When I no longer wanted to be at Intel because Intel's focus wasn't on graphics, I looked around the industry and I looked at companies like Nvidia and Microsoft and Sony and I chose AMD because AMD has this open, very ecosystem friendly kind of approach. We don't tie people to our platform.
You've probably seen some fuss about GameWorks recently and we're going after that and we're attacking it as being just the opposite of open. They close everything down. There's DLLs which are delivered to games developers with code in there that they can't see. It's the opposite of what any games developer would want and it is not the way AMD approaches it. We could have [done it] our own. 'We'll do AMD GameWorks, that's the answer. We'll lock everyone into our code.' To be honest, it's a dumb way to drive the industry forwards.
TR: That's not exactly a warm and welcoming attitude.
RH: I don't know if you've had the exciting thrill of watching my video with PC Perspective in the last couple of days. I've been flattered by the amount of attention that it's received because we've gone out there and we've talked very deliberately, very openly about the things that we want and the things that we like and why we're trying to go in particular directions and what we think is wrong with some of the things in the industry.
The stuff we talk about with open standards, it's always gentle and friendly and touchy feely kind of stuff. It's, you know, pushing things in a healthy ecosystem kind of way, but we are absolutely willing to defend ourselves and our principles so we are being pretty bullish about the attack that we're making on GameWorks at the moment as being a closed, propriety way of near as damn as cheating at benchmarks.
It's an artificial way of creating benchmarks which harm AMD, and that's bad for AMD, obviously, but it actually is bad for games players even when they are using Nvidia hardware. It's as bad as that. That's a remarkable state of affairs. So you will see a change from AMD in that we're taking that slightly more bullish attitude and we're going out there and shouting at it and being proud of the things that we do well.
Most corporations have a legal department that try to rein you in and I've certainly felt the wrath of the legal team where they try and make you terribly timid. AMD is more committed, the organization as a whole is more committed, to this bullish attitude, so I've not had any desperate emails from legal about some of the interviews that I've been involved in, nor do I expect them.
We are serious about going out there and talking very proudly about the things that we're doing, which are good for the industry. We believe that things that we do which are good for the industry will be good for AMD.
TR: Can you speak specifically about what sets Mantle and GameWorks apart?
RH: One of the accusations that Nvidia had made about Mantle is that well, look, Mantle is about optimizing for AMD hardware and GameWorks is for optimizing for Nvidia hardware, so why are they making such a fuss?
There's a basic principle underneath it which shows that they're fundamentally diffe rent things. With Mantle, we will allow someone to build a Mantle driver if they like, Nvidia can do it, and there is absolutely nothing, if you think about it logically, there is absolutely nothing that AMD can do to harm a competitor by using Mantle.
If a games developer chose to do a Mantle-only game, we would discourage them. I guess it would be the games developer's choice, but we're not encouraging that. They aught to do a DX game or a GL game, but no matter how fast it runs on Mantle, we can't stop, or influence even, how fast it runs on Nvidia. This is just a way of giving the goodness in our hardware to our consumers.
Unfortunately GameWorks works the other way. It stops the developer doing what they want to do. It stops games players from getting a great experience. The visuals may be improved by frame rates tend to plummet when GameWorks is on, and it creates this artificial situation where they harm us with code that they've written. We can't to that. They're fundamentally different things. Ours is an open standard, theirs is a closed box that is damaging.
[Laughs] Very bullish. And it's often the case that people will futz around a little bit and not try to name competitors, but this stuff looks very malevolent to me, and whether I read their minds correctly or not I don't know, but the fact that they can create a benchmark that is run on our hardware? Huh? Not Right.
No comments:
Post a Comment