AMD Releases Beta OpenCL 1.2 Driver For Developers, Announces HSA Runtime For Linux
Update: AMD let us know that it published the driver at http://developer.amd.com/resources/heterogeneous-computing/opencl-zone/
AMD released its beta Windows OpenCL 1.2 development driver today, and is officially announcing its plans to release Heterogeneous System Architecture (HSA) support for Linux. The company plans to implement this support in the second quarter of 2014 through HSA runtime software that will support multiple programming languages.
While HSA is still in its infancy, it holds the promise of unlocking significant compute performance potential that can be applied to many types of workloads. About 47 percent of the Kaveri die is dedicated to GPU resources. Typically, that logic sits unused unless it's called on for graphics processing tasks. In a perfect world, though, developers could use those transistors to help accelerate general-purpose tasks, yielding a performance boost and improved efficiency. That's the ultimate purpose of heterogeneous computing, using the right device for a given task, and the potential benefits as they've been presented to us are certainly exciting.
With Kaveri, AMD facilitates true HSA-capable hardware for the first time. But as with all hardware, it's only as good as the software that supports it. And software support requires investment from developers. This is the giant hurdle that AMD needs to overcome before the benefits of HSA can be realized. If heterogeneous devices are the cart, then a commitment from the development community is the horse.
This isn't the first time AMD has outlined its strategy to enable HSA on Linux, nor HSA support for multiple languages like Java, C++ AMP, Python, and OpenCL 2.0. The company is putting a lot of effort into developer tools, so it's understandably trying to generate developer interest on an ongoing basis. At the end of the day, it all comes down to those developers. No matter how much potential HSA has, it can only aspire to be as pervasive as the software that supports it. If important applications ever deliver big performance boosts with this technology, AMD's APUs will potentially become a lot more desirable. On the other hand, if HSA never catches on with developers, the enablement work won't be the big pay-off AMD is hoping for.
Right now, the general message is repetitive. We heard it delivered at the AMD Developer Summit 2013, and more recently during the Kaveri tech day in Las Vegas. To recap, Java already has limited support for HSA through Aparapi, an API for expressing parallelized workloads. The company expects the subsequent release of Java 8 to arrive in the middle of 2014 with support for Lambda language expressions and parallel acceleration (thanks to the Sumatra project, which sets out to let Java applications leverage GPUs and APUs) through the HSAIL intermediate language, culminating in native Java Virtual Machine (JVM) support via Java 9 in 2015. OpenCL 2.0 is expected to be the primary path for most apps on both Windows and Linux platforms, though, and AMD plans to deliver this functionality to selected developers in Q3 of this year, with optional features enabled in Q4 that should leverage even more HSA-specific optimizations and functionality.
As I said though, almost all of this information has been discussed; it's a work in progress. To us, the real news today is the release of the Windows Beta OpenCL 1.2 Driver for developers who want to dip their toes into HSA waters now, rather than waiting for full OpenCL 2.0 availability later this year. This development driver is the same one used to drive the demos we saw during the recent Kaveri launch, so it gives widespread access to early adopters for the opportunity to test the potential benefits that HSA has for their software.
With this driver in the wild, we may see better indications of how much developer interest AMD has drummed up over the next few months. Of course, the company already made its driver available to targeted ISVs previously. But now that it's in the domain of anyone who wants a crack at it, we will see what the community comes up with. We plan to follow up this announcement with tests and developer feedback in the days and weeks to come, so stay tuned.

Hundreds of PRO apps already support Cuda and migrating them to Cuda6/maxwell gpus isn't as hard as getting new support for HSA. Game over man...If you're going to cause a lot more work for a dev, you'd better be able to explain how it blows away what they already have invested in heavily. If all it does is copy the same features what is the point in supporting it? I need to gain something right? It's also a lot easier to push a new standard from a position of money/power/market share, which AMD has no such position in anything. Broke/weakness can't push anything up a hill.
Hundreds of PRO apps already support Cuda and migrating them to Cuda6/maxwell gpus isn't as hard as getting new support for HSA. Game over man...If you're going to cause a lot more work for a dev, you'd better be able to explain how it blows away what they already have invested in heavily. If all it does is copy the same features what is the point in supporting it? I need to gain something right? It's also a lot easier to push a new standard from a position of money/power/market share, which AMD has no such position in anything. Broke/weakness can't push anything up a hill.
I disagree with you but you already countered your own point. OpenCL is supported by both Nvidia and AMD. If I was investing money it would not be in a proprietary solution that only works with one type of hardware especially if I was trying to sell something as a package to others like how Photoshop CS6 dropped CUDA for OpenCL support. You will see this trend out over the years as CUDA disappears from the landscape. The same will be said of Mantle years from now as it dies off because OpenGL and DirectX are going to finally step up to the plate. Both CUDA and Mantle will have done something great which is to force the hand of the tech industry to create standards.
The same thing will be the case with Gsync and Freesync five or so years down the road when the basic functionality is baked into every GPU and monitor as a VESA standard as it naturally will because of power savings with a nice plus for gamers.
Hundreds of PRO apps already support Cuda and migrating them to Cuda6/maxwell gpus isn't as hard as getting new support for HSA. Game over man...If you're going to cause a lot more work for a dev, you'd better be able to explain how it blows away what they already have invested in heavily. If all it does is copy the same features what is the point in supporting it? I need to gain something right? It's also a lot easier to push a new standard from a position of money/power/market share, which AMD has no such position in anything. Broke/weakness can't push anything up a hill.
I disagree with you but you already countered your own point. OpenCL is supported by both Nvidia and AMD. If I was investing money it would not be in a proprietary solution that only works with one type of hardware especially if I was trying to sell something as a package to others like how Photoshop CS6 dropped CUDA for OpenCL support. You will see this trend out over the years as CUDA disappears from the landscape. The same will be said of Mantle years from now as it dies off because OpenGL and DirectX are going to finally step up to the plate. Both CUDA and Mantle will have done something great which is to force the hand of the tech industry to create standards.
The same thing will be the case with Gsync and Freesync five or so years down the road when the basic functionality is baked into every GPU and monitor as a VESA standard as it naturally will because of power savings with a nice plus for gamers.
Adobe didn't drop CUDA. It's still in there, works fine. AMD just paid them to support OpenCL so they could join the party. I'm pretty sure there isn't any point in saying more when you don't even know the product you're talking about. OpenCL isn't supported by Nvidia. It may function, but that isn't the same as being optimized (supported). If you buy NV cards for WORK, you already know you should be running CUDA period. Cuda loses NOTHING vs. OpenCL in the real world (meaning stuff you make money with). Stupid tests like mining, F@H, the dumb OpenCL vid/photo tests tomshardware runs etc, can show NV in a bad light, but only because NV knows you'll use CUDA instead and doesn't seem to care what an OpenCL benchmark shows (workstation users know the real story, and websites can't show anything different so they'll keep buying CUDA). OpenCL is for poor people that can't afford a CUDA PRO card (and support that comes with it) or the expensive software you BUY a Cuda card for
NV has no reason to optimize for this junk that makes ZERO profits for business where they rule 90% of workstations already. Why would you waste resources optimizing for ANY app/API you already dominate with your current tech? You won't see MS supporting Mantle either unless the market forces them to (there is no way for AMD to pay for that to happen). Adobe dropped nothing, they are merely ADDING support for AMD because AMD paid them to. Speaking of which, we heard about OpenCL support for Adobe a year ago, where is it? AMD didn't pay enough I suppose to keep it on Adobe's front burner (you don't see adobe pitching it front and center on their pages either) or it just sucks right? AMD should have gotten that in Adobe software 5-7 years ago. Too late to take out the guy who has been spending on cuda for 7yrs and STILL is. If it's working & anywhere near as fast as AMD's fireprographics site shows how come we get no tests vs. Cuda and AMD is SILENT on this great stuff they have?? AMD made a lot of noise but hasn't shown anything in a year. It supposedly works fine in Adobe CC and CS6 (2012 app), yet we get no benchmarks from tomshardware, anandtech etc or even AMD themselves (one chart that doesn't show much)...Odd? I don't think so, they just both hate proprietary stuff and push OpenCL junk instead even though they KNOW cuda wins. I like OPEN standards when they are BETTER, but not when they get their clocks cleaned by proprietary stuff. I can understand neither site wanting AMD to get any weaker, but BSing the public at large (or hiding info we need and KNOW you must have knowledge of) only hurts your credibility in the end. Toms runs adobe tests, but pretends it doesn't support CUDA. I guess that's why you think it doesn't. Hence my BSing the public comment. They might not come out and say Cuda sucks (you can't that's a total lie and provable if they said that), but NOT saying what they know is true (it rocks in PRO apps) is the same thing as spouting lies IMHO.
Again, why do you think Toms, anandtech etc never test CUDA vs. OpenCL? Because it's no contest. NV has poured BILLIONS into cuda, schools to teach it etc. AMD has lost 6BILLION+ over the last 10yrs. Read that again. They haven't made a profit over the last 10yrs. Why do you think they are 7yrs late to the PRO party? Why do you think their OpenCL claim to fame is a SINGLE app from Adobe? Why do you think they have no Moblie SOC yet? You can add up all the years AMD has been in business and they've never made a penny over the life of their company. Read that again. Come back when you can explain to me how a company who has NEVER made a dime over the life of their company & has billions of debt, will take down any company with ZERO debt, billions in the bank and leading technology.
Come back when all of your fantasies are reality. Until then CUDA and it's dominance is reality. Come back when AMD owns 90% of the workstation market and can push a new standard. Come back when OpenCL is winning VS cuda in Adobe photo or video editing tests. Tomshardware and anandtech never respond when I quiz them on why they don't put Cuda vs. OpenCL EVER.
When you ask people who USE this stuff and not just read websites that hide cuda or never pit it vs. OpenCL what do you get? :
http://www.reduser.net/forum/showthread.php?107980-Adobe-CUDA-vs-Open-CL-780ti-vs-R9-290
Mantle will die because others that are entrenched (DX/OpenGL) can do the same stuff and are backed by a TON of people & money already and every game company knows this tech, while they have to learn mantle and by all accounts it probably isn't worth wasting the effort for 7-12% and a small portion of the market.
You are investing due to "feelings" where I don't care about proprietary as a stock buyer, I only care that what I buy WINS and makes me money. I will sit here and tell you I hate NV owning workstation graphics (but love the perf), but in the same second buy another 1000 shares...LOL. My feelings about apple has NEVER stopped me from buying the crap out of their shares when I think they will make me money (I do not like their tactics, despite great products for some uses). It never stopped me from buying MSFT either, despite me basically despising them even though my job depends on them (for now).
Gsync will win because it's backed by a profitable company and is ALREADY here now, usable for GAMERS. AMD can't even fund what they have going now (phase3 drivers come to mind, cpu race given up, running to low end, can't field a mobile device etc), but you somehow think they'll manage to fund mantle to victory, fund OpenCL to take out the KING Cuda, and fund Freesync (not free, need a monitor too, maybe should be called Onedayyoumightsync?) to conquer already here Gsync. How can you live so far from reality and mention buying stocks?
Funding problems are killing AMD's dreams. They couldn't fund a decent cooling solution for their ref cards. They can't fund freesync into a product, can't get Mantle out of beta, can't get out of phases for drivers, can't get OpenCL in much of anything of value (mining, F@H don't count), can't fund an attack on mobile with ARM (only server, where everyone else already coming with 20nm while seattle will sit at 28nm first), stuck at 28nm because they have to go tried/true rather than risk a move to 20nm etc. Worse if AMD is REALLY selling to mining people, gamers aren't even getting mantle capable cards in their hands. Translation, NV will steal more share where it matters; People who play games or use PRO apps (even grid going up, now 200 companies testing it, up from 47 testing it). Mining is a temp world and 437mil stolen by the bitcoin guy shows it's fools gold if you hold these things, note the grins in photos...LOL I said this would happen), while gamers are the REAL WORLD for GPU makers. What does AMD do? They put up a slide showing MINING...ROFL. AMD appears to not understand WE GAME, very few of us MINE anything. Very few of us (165K out of 7Bil people) F@H. Many of those downloads are NOT home users also. AMD should have cut that crap out and concentrated on GAMING which is what sells cards long term.
I suggest you never buy stocks. If you can't buy purely because of LOVE for MONEY, you're going to lose your shirt. The second you love the company you bought you're dead. You can love a company as a USER, not as a stock buyer.
Go to Adobe.com and punch Nvidia Cuda into the search box. Tell me again they dumped it. They have a post on a Cuda 5.5x fix hours ago...I digress...