News Gaming Update : AMD Is Working On A New Linux Graphics Driver To Catch Up With Nvidia

There’s no doubt about it: AMD’s Linux graphics drivers are behind Nvidia’s, something that will start mattering a lot more when Valve’s first Linux-based Steam Machines start hitting the market this November.

AMD hasn’t turned the ship around yet, and big-name games are still only supporting Nvidia hardware when they launch on Linux. But AMD hasn’t been sitting on its hands. AMD’s developers are working on a new Linux driver architecture that will result in better open-source drivers, too—eventually.

How has AMD been doing?

Before we dive into that, though, let’s recap what’s happened since out last look at the subject of Linux graphics drivers.

Nvidia is still maintaining its lead over AMD on Linux, and new games still target Nvidia hardware. Middle-earth: Shadow of Mordor recently launched on Linux thanks to Feral Interactive, but it only officially supports Nvidia graphics cards. The official FAQ says you’ll experience poor performance if you attempt to run it on an AMD graphics card. 

Phoronix recently discovered you can boost the performance of Counter-Strike: Global Offensive on Linux when you’re using an AMD graphics card just by renaming the “csgo_linux” binary to “hl2_linux”. This will give you as much as a 40 percent graphics boost. The AMD Catalyst driver has application profiles designed for Source engine games, but AMD’s developers haven’t bothered adding csgo_linux to the application profiles—despite Counter-Strike: Global Offensive having been out for a year at this point.

Application profiles are commonly used across operating systems and drivers, so this is normal. What looks bad for AMD here is how slow it’s been to maintain these application profiles when compared to Nvidia on Linux and even AMD’s own profiles on Windows.

It’s not all bad for AMD users. AMD released Catalyst 15.5 Linux in early June. Phoronix ran some benchmarks and put it bluntly: “Metro Last Light Redux and Metro 2033 Redux no longer run like garbage on AMD Linux.” That’s an improvement, but the AMD Catalyst graphics drivers are still behind Nvidia’s. And that was the only game that noticeably improved in performance with the new drivers.

AMD’s new graphics driver architecture :
Currently, there are two main AMD graphics drivers on Linux. There’s the open-source “Radeon” driver and the closed-source “Catalyst” driver. As with Nvidia’s drivers, the open-source driver is fine for just using a graphical desktop with AMD graphics cards, but you’ll want the closed-source driver to get maximum gaming performance.

AMD has now been pursuing a “unified” Linux driver strategy and writing an entirely new driver. This driver, known as “AMDGPU,” will have a single Linux kernel module, which will be open-source. The closed-source Catalyst code will continue to exist, but it will be a smaller “binary blob” that runs in userspace. Open-source fans who don’t need maximum gaming performance can skip the Catalyst blob and use an entirely open-source driver.

 This driver will only be used for new AMD graphics cards. It will only support the very latest GPUs and future AMD graphics hardware.

The new structure could help a lot. Rather than two entirely separate drivers with separate kernel modules, there will be a single open-source kernel driver. The closed-source Catalyst part of the driver becomes much smaller and confined to userspace. AMD won’t have to update the Catalyst driver whenever there’s a new Linux kernel or X.org X server release. It will automatically be compatible because the Catalyst driver is a smaller piece of code that hooks into the open-source AMD driver included in the projects themselves.

While AMD isn’t going completely open-source as Intel did with its graphics drivers, it’s much more open-source-friendly than Nvidia’s strategy of going it alone. Linux developers have wanted closed-source kernel modules to go away for a long time, too. For more details, read Phoronix’s report on AMD’s new Linux driver strategy.

This driver should appear over 2015, and the “AMDGPU” kernel driver is set to debut in Linux 4.2. (Yes, we’re already past Linux 4.0 !) However, the driver is in a very early state and has a long way to go. Don’t expect to be using it any time soon.

In the long run, this could be what helps AMD close the gap with Nvidia when it comes to Linux graphics drivers. We should all hope so, anyway—it would be best for Steam Machines if AMD and Nvidia were competitive.

Posted by Unknown Wednesday, August 05, 2015

News Gaming Industry Update : NVIDIA Releases GeForce GTX 960M and GTX 950M Mobile Graphics

NVIDIA has announced new GPUs to round out their 900-series mobile lineup, and the new GTX 960M and GTX 950M are based on the same GM107 core as the previous 860M/850M parts.

Both GPUs feature 640 CUDA Cores and are separated by Base clock speed, with the GTX 960M operating at 1096 MHz and GTX 950M at 914 MHz. Both have unlisted maximum Boost frequencies that will likely vary based on thermal constraints. The memory interface is the other differentiator between the GPUs, with the GTX 960M sporting dedicated GDDR5 memory, and the GTX 950M can be implemented with either DDR3 or GDDR5 memory. Both GTX 960M and 950M use the same 128-bit memory interface and support up to 4GB of memory.

As reported by multiple sources the core powering the 960M/950M is a GM107 Maxwell GPU, which means that we are essentially talking about rebadged 860M/850M products, though the unlisted Boost frequencies could potentially be higher with these parts with improved silicon on a mature 28nm process. In contrast the previously announced GTX 965M is based on a cut down Maxwell GM204 GPU, with its 1024 CUDA Cores representing half of the GPU core introduced with the GTX 980.

New notebooks featuring the GTX 960M have already been announced by NVIDIA's partners, so we will soon see if there is any performance improvement to these refreshed GM107 parts.

Posted by Unknown Friday, March 13, 2015

News Gaming Update : Nvidia Surprise Launches The 12GB GeForce Titan X, The Most Advanced GPU Ever'

 Nvidia chief executive Jen-Hsun Huang strolled on stage at a Game Developer Conference presentation from Epic Games, casually announced the new flagship GeForce Titan X graphics card, autographed it, and left.

So that happened.

After spending the better part of two hours launching its Nvidia Shield gaming set-top box/console on Tuesday night, Huang took everyone by surprise when he launched the Titan X on Wednesday morning.

“We have launched the most advanced GPU ever but and [given] the first one to Tim Sweeney,” the founder of Epic Games, Huang said on stage. After announcing the Titan X and some of its specs, Huang autographed the massive GPU module with a flourish—“To Tim, with love, Jen-Hsun” waved goodbye, and left.

The Titan X is apparently anything but vaporware; Huang said that it “will power GDC 2015,” and that showgoers would see “some amazing demonstrations this week”.

“I cherish the hardware and of course we’ll see what it can do,” Sweeney said.

 So what’s inside the Titan X? Eight billion transistors, making it the “most advanced GPU the world has ever seen,” Huang said. It will contain a massive 12GB frame buffer. And that’s about it: Huang didn’t reveal more details, and Nvidia’s official blog and press releases haven't mentioned it yet.

Given what Huang said about it, however, we can assume that it’s more powerful than the GeForce Titan Z, a $3,000 graphics card which Nvidia launched last March.  That card included 5,760 CUDA cores with two Kepler cores inside of it, 12GB of memory, and 8 teraflops of computing power. Since the GPU has the same amount of memory inside, it’s likely that there are more cores. It’s also not clear what GPU is at the heart of the Titan X; Nvidia is preparing its next-gen Pascal processor, but that’s not due until 2016, Huang said at the time.

More details about the new GeForce Titan X will be shared at Nvidia’s GPU Technology Conference on March 17, a spokesman for the company said. He declined to comment further.

Why this matters: Although Nvidia still makes the bulk of its income from the PC, enterprise products command huge premiums. Nvidia would like to make products like its Iray VCA, a $50,000 virtual computing appliance for rendering images using modeled photons, as the tool for CAD and CGI specialists to fabricate their renderings. What's really interesting is that we should be able to see this on the show floor at the Game Developer Conference this week.

Posted by Unknown Thursday, March 05, 2015

AMD Announces Radeon HD 8000M Series GPUs

AMD has been usually silent when it comes to the GPU market. According to Engadget, AMD is now delivery its newest sequence of GPUs made for laptops—the HD 8000M sequence. The first of these, known as the Radeon HD 8555M, is delivery within the Asus Vivobook U38DT. The GPUs will be making their existence sensed in 2013

AMD has also exposed some guarantees about the GPU, which display the Radeon 8800M ruining Nvidia's Geforce 650M in standards. The company hasn't released any more information though, such as energy intake.

All four snacks in the series—the 8500M, 8600M, 8700M and the 8800M—contain AMD's newest 28nm Style Primary Next structure. This gives them advantages when it comes to DirectX 11.1. The GPUs are also set to advantage from energy preserving functions like ZeroCore, which reduces off energy to rarely used cores, and PowerTune, which increases time rate to use low compertition energy.

Back in Nov, AMD had declared its FirePro GPU work stations in Indian. According to AMD, the FirePro W9000 GPU functions amazing estimate and improved storage information and provides better multi-display assistance performance than the competitive remedy. AMD has also released the AMD FirePro W8000, W7000 and W5000 work station graphics cards, which are all designed on the AMD Style Primary Next Architecture, and developed to stability estimate and 3D workloads effectively for computer-aided design and technological innovation, and for press and enjoyment experts. 

AMD FirePro A300 Series multiplied handling device (APU) is developed for entry-level and popular desktop pc work stations. Presenting AMD Eyefinity multi-display technological innovation, the new AMD FirePro A300 Series APUs are developed for customers who need a high-performance handling foundation to energy their computer-aided design, and press and enjoyment workflows.

David Cummings, Mature Home, Engineering WS Style, AMD said, “As experts work with bigger information places that need innovative creation and complicated designs, they need a graphics remedy that is fast, highly effective, and efficient. Qualified for the present programs, the new AMD FirePro work station graphics cards carry a variety of functions and abilities for experts working in electronic signs, transmitted graphics, CAD/CAE and M&E, providing the perfect stability of energy, performance and stability at the right price.”

“Design experts need workstation-class resources that allow performance and versatility in their work-flow, and the AMD FirePro A300 Series APUs allow work station integrators and OEMs an interesting new handling foundation on which to make and develop highly effective, entry-level desktop pc work station options that provide irresistible value for CAD and M&E workflows,” included Mr. Cummings. "The AMD FirePro A300 Series APUs brings together AMD FirePro graphics technological innovation with innovative CPU technological innovation, providing amazing estimate performance, enhanced design versatility and excellent performance."

Earlier this year, AMD had released its Radeon HD 7990. A dual-GPU cards from AMD was unavoidable, as it, just like Nvidia, always has a dual-GPU video cards that is usually the combination of two of its highest-end GPU cores connected by CrossFire on a single PCB.

Posted by Unknown Tuesday, December 18, 2012