1 2 3 Previous Next

AMD Business

446 Posts

This post has been authored by Suresh Gopalakhrisnan, a Corporate Vice President and General Manager at AMD.

 

Earlier this year, we welcomed the latest server CPU, formerly codenamed “Warsaw”, to our family of x86 AMD Opteron processors. The AMD Opteron™ 6300 Series CPU was developed after AMD spent  time listening to key customers about what level of performance and power would suit their data centers need. When these customers, such as 1&1 Internet, run thousands of servers, even a single feature can make a lot of difference.

 

1&1 Internet is an innovator in the web hosting industry and as such set extremely high standards for performance and reliability. So it gives me great pleasure to see 1&1 Internet, one of Europe’s biggest and most respected hosting companies deploy servers with AMD Opteron™ 6388P processors for its XL dedicated range of servers.

 

With 1&1 Internet’s XL o12A-32 and o12A-64 dedicated servers, customers can deploy 12-core AMD Opteron processors with up to 64GB RAM and  4TB of storage space giving mid-sized businesses exceptional compute performance and storage at an affordable price. With more cores you not only have performance on tap but it allows businesses more room to grow without having to worry about migrating to new hardware.

 

The AMD Opteron 6388P processors are designed to meet today’s enterprise workloads, making 1&1 Internet’s AMD powered XL servers ideal for dynamically generated websites, xSQL databases, CRM and storage.  To support these heavy duty, compute and memory intensive workloads, 1&1 Internet’s customers will be able to take advantage of key features in the AMD Opteron 6338P processor, including:

 

  • 12 64-bit x86 cores using our “Piledriver” architecture, a tried and tested architecture that designed for today's multi-threaded workloads
  • AMD Turbo CORE technologies boost each core from 2.3GHz to 2.8GHz, providing a power efficient performance boost when it is really needed
  • AMD Virtualization™ technology, to provide superior performance for virtual machines
  • AMD-P, our state-of-the-art power management technology, enabling systems using the AMD Opteron 6338P processor to be extremely power efficient

 

Our AMD Opteron processors have led the way when it comes to supporting high memory capacities and the AMD Opteron 6388P processors used by 1&1 Internet are no exception. I am so pleased to see that 1&1 is taking advantage of this by offering 64GB RAM on the o12A-64, more than double that of our competitor and for less money!

1&1 Internet’s decision to use AMD Opteron 6388P processors in their XL range of dedicated servers is a great choice for enterprise customers. The decision highlights how forward-thinking and innovative 1&1 Internet is in providing their customers with a choice and allowing them to use the technology that is best placed to serve their compute needs. With AMD’s expertise in energy efficient, high performance compute and 1&1 Internet’s state-of-the-art hosting technologies, 1&1’s XL o12A-32 and o12A-64 dedicated servers are the perfect choice for mid-sized enterprises looking for the ultimate in compute and network performance.



Suresh Gopalakhrisnan is Corporate Vice President and General Manager of the Server Business Unit at AMD. His postings are his own opinions and may not represent AMD’s positions, strategies or opinions. Links to third party sites, and references to third party trademarks, are provided for convenience and illustrative purposes only. Unless explicitly stated, AMD is not responsible for the contents of such links, and no third party endorsement of AMD or any of its products is implied.

guestblogger

The New Balance of Power

Posted by guestblogger Jul 25, 2014

By Ben Boehman, Product Development Engineer, AMD


Boehman_headshot_resized.jpg

 

Microprocessor vendors are understandably proud of the innovation that we pour into our latest and greatest offerings, touting processor performance breakthroughs and benchmarks that push the boundaries of compute and IO speed. We balance these performance data points with important metrics about the processor’s energy efficiency, highlighting thermal design power (TDP) and performance-per-watt as critical measures of a processor’s value to the customer.


This is meaningful info that helps enable a discerning customer to make informed purchasing decisions. But these metrics don’t tell the full story about a processor’s performance-per-watt benefits for the customer’s unique design requirements, which can call for dynamic power/performance scaling to efficiently accommodate rapidly-shifting processing workloads and challenging thermal conditions.


What’s needed is the ability to reduce the power of underutilized cores and re-allocate that thermal budget to other cores for improved performance and better efficiency. This is especially important for APUs and other platforms with multiple onboard processing engines and varied functional ‘blocks’. AMD’s Turbo CORE technology utilizes algorithms that assess a variety of frequency, voltage, temperature and logic activity inputs to dynamically determine which core needs a performance boost and how much thermal headroom is available.


The latest generation of AMD Embedded APUs provides a ‘configurable TDP’ capability, essentially giving system designers a knob that they can turn to modify the APU’s TDP to better fit the needs of the target application.  Together, these features improve flexibility of APUs by maximizing performance under a variety of design constraints.

 

To learn more, read AMD’s white paper, ‘Advanced Power Management Helps Bring Improved Performance,’ available here.

 

Ben Boehman is a Product Development Engineer for Embedded Gaming at AMD. His postings are his own opinions and may not represent AMD’s positions, strategies or opinions. Links to third party sites, and references to third party trademarks, are provided for convenience and illustrative purposes only. Unless explicitly stated, AMD is not responsible for the contents of such links, and no third party endorsement of AMD or any of its products is implied.

Several advances in the software ecosystem – particularly those associated with OpenCL™ -- inspired AMD to showcase some great innovations at International Supercomputing Conference (Leipzig, Germany, June 22-26).

 

The conference attracts industry leaders, a variety of exhibitors, and members of academia from around the world and this year it explored the future direction of HPC technologies, life science applications, big data, quantum computing, and the real-world value of HPC.

 

As you might know, OpenCL™ has made GPUs increasingly attractive for running high-performance computing (HPC) workloads in a scalable and efficient manner. At the show, I had the opportunity to share details on AMD’s new workstation and server GPU solution.

photo 3.JPG.jpg

One of the highlights was the announcement of AMD’s new AMD FirePro™ W8100 workstation graphics cards powered by OpenCL™ for personal supercomputing.  Each card is capable of delivering more than 4.0 TFLOPS of peak GPU compute performance for future-ready 4K multi-display workflows, and deliver up to nearly 38X more floating point compute performance than the most comparable competing product(1).

 

Another highlight was a sneak peek of the new AMD FirePro™ S9150 – our newest server GPU for HPC. This forthcoming product offers massive parallel processing power that combines with OpenCL™ to accelerate applications beyond just graphics, and compute intensive workloads are expected to benefit from technologies and features such as:
- More than 2.0 TFLOPS peak double precision performance
- Half rate double precision floating point
- High performance per watt
- 16GB ultrafast GDDR5 memory
- AMD STREAM technology(2)
- Mechanical and thermal design optimized for standard servers

 

We treated attendees to a demonstration of the AMD FirePro™ S9150 running a prototype version of the open source numerical computation package Scilab. This side-by-side dataset comparison took a look at Scilab running on a CPU only, alongside Scilab with the sciGPGPU module providing support for OpenCL™. The demonstration – powered by four AMD FirePro™ S9150 server GPUs in a Supermicro SuperServer chassis – showed how for the first time, users will have access to GPU acceleration for mathematical modeling enabled by the industry standard OpenCL™ API. The Scilab application provides a simple interface to access the power of GPU compute, even for users who aren’t familiar with the intricacies of programming for GPU acceleration.

 

photo 1b.JPG.jpg

 

AMD FirePro™ S9150 GPUs support AMD STREAM technology(2), which was first announced at Computex 2014.  AMD STREAM technology powers the ecosystem that enables AMD FirePro™ graphics cards – including the new AMD FirePro™ S9150 – to be used for compute-intensive workflows leveraging the massively parallel processing power of AMD GPUs.

 

This technology can be used to accelerate many applications beyond just graphics, including critical research and operations for a variety of scientific and mathematic purposes across many industries and disciplines. For example, AMD has supported the University of Frankfurt SANAM Supercomputer with AMD FirePro™ S10000 server GPUs, with the same underlying technology as the AMD FirePro S9150 to enable groundbreaking power and efficiency for specific research projects.

 

AMD STREAM technology(2) includes:
- ECC Memory Support
- Fast Single & Double Precision Performance
- Peer-to-Peer Multi-GPU Support
- Bi-directional PCIe® Memory Transfers
- GPU Optimized OpenCL™ Libraries

 

The goal for AMD STREAM technology enabled on AMD FirePro S-Series server GPUs is to deliver immense compute capability, performance and flexibility to handle the dense, multithreaded workloads and extremely large, complex datasets associated with supercomputing and HPC.

 

If you aren’t doing so already, try working with OpenCL and assess the performance of AMD FirePro™ graphics for your particular compute workloads.

 

Learn about AMD FirePro graphics and University of Frankfurt SANAM Supercomputer


Niles Burbank is a senior manager, server GPU products, at AMD. His postings are his own opinions and may not represent AMD’s positions, strategies or opinions. Links to third party sites are provided for convenience and unless explicitly stated, AMD is not responsible for the contents of such linked sites and no endorsement is implied.

OpenCL and the OpenCL logo are trademarks of Apple Inc. used by permission by Khronos. Apple and Mac Pro are trademarks of Apple Inc., registered in the U.S. and other countries.

 

1 SiSoftware Sandra test details:
System Description: AMD FirePro W8100 vs. Nvidia Quadro K5000 - Dell T3610, Intel Xeon E5-1620 v2 @ 3.60 GHz, 8GB DDR3, Seagate HDD 7200RPM, Win7 64-bit SP1, 1920x1080 resolution
AMD Driver 13.352.1009 | Nvidia Driver 333.11

 

2 AMD STREAM technology is a set of features offered with select AMD FirePro graphics cards for the acceleration of compute-intensive workflows.  Not all products have all features and full enablement of some capabilities may require complementary software.  Check with your system manufacturer for specific capabilities and supported technologies. 

System Admin

The Future of Display

Posted by System Admin Jun 25, 2014

Amidst a cornucopia of visual attractions and distractions, AMD demonstrated an 8K resolution future of displays to attendees at InfoComm 14. The annual event brings together manufacturers, system integrators, end users and multimedia professionals from more than 80 countries to view the latest and greatest in professional audiovisual and information communications.

 

InfoComm 2014 booth1.jpg

 

AMD has always been on the cutting edge of visual solutions and has led the charge on a variety of display technologies. AMD offers industry leading multidisplay technology with AMD Eyefinity, and AMD was first and is still the only vendor to support up to six displays from a single GPU.  In the professional market, AMD enabled edge blending, image warping, and color correction with Scalable Display through our AMD FirePro DOPP (Display Output Post-Processing) feature. AMD was also one of the first to embrace support for Miracast, the industry standards based solution for wireless displays.

 

From that powerful heritage of display and graphics innovation, AMD unveiled new, cutting-edge display technologies at InfoComm 14. With 4K displays now a reality and almost the expected norm in the professional graphics industry, the natural progression is to ask what lies beyond 4K displays. AMD answered that question by demonstrating four 4K displays driven from a single AMD FirePro graphics card - exactly the same number of pixels of an 8K display. In fact the media server, which contained two AMD FirePro™ graphics card, drove a total of eight 4K displays for a total of 66 MPixels!

 

AMD also demonstrated a prototype monitor that supports FreeSync, AMDs technology based on VESA’s DisplayPort Adaptive sync. Vertical synchronization, or v-sync, is the traditional solution to screen tearing, but it introduces its own problems. FreeSync helps solve tearing without those problems or the use of proprietary technology. AMD recently published a blog on the subject of FreeSync as part of an ecosystem around display technologies that helps reduce unwanted visual artifacts such as tearing, stuttering and input latency. AMD has applied that
solution to the professional graphics market with a demonstration at InfoComm 14.

 

InfoComm 2014 booth2.jpg

 

 

Finally, AMD’s Display SW Architect Syed Hussain presented the topic of end-to-end display pipeline on behalf of VESA in the talk entitled “DisplayPort: 4K and Beyond”. He explained the end-to-end system bottlenecks that need to be alleviated in order to support higher than 4K resolutions. Additionally, Syed provided insight on how to support beyond 4K resolutions with different techniques offered by DisplayPort.

 

Learn more about display information on AMD.com

 

Roger Quero is a Solutions Architect, Professional Graphics at AMD. His postings are his own opinions and may not represent AMD’s positions, strategies or opinions. Links to third party sites are provided for convenience and unless explicitly stated, AMD is not responsible for the contents of such linked sites and no endorsement is implied.

bellzey

Converging Technologies

Posted by bellzey Jun 18, 2014

By Kelly Gillilan, AMD Embedded Solutions


At AMD, we view the embedded industry as segmented into different vertical markets (or “verticals”), such as digital signage, industrial control, factory automation, communications, medical imaging, embedded gaming, etc. And each of these verticals has its own ecosystem of suppliers, developers, integrators, and even technologies. Yet at the recent G2E Asia event, an industry event specifically for the casino gaming market, I noticed a convergence of technologies typically associated with different vertical markets at levels I had not seen in the past.


One such example is the “smart” virtual dealer. Many companies demonstrated virtual dealers comprised of video segments of actual dealers who would perform actions (deal cards, collect the cards, call for bets, etc.) depending on the circumstance. Some companies took that technology to the next level by incorporating facial/player recognition technologies to determine the playing audience. By incorporating this type of technology (which was typically found in digital signage applications), the dealers would know who was sitting at which play stations and could provide direct interaction—making direct eye contact, giving words of encouragement,  congratulating the winners, etc.—with those players.

 

Another example is the use of equipment typically found in a factory automation setting to drive some of the games. At G2E, I saw robotic arms running roulette wheels, traditional motors driving gigantic 20’ wheel games, and even platforms designed to shake dice. Pneumatics were also often used to reset ball positions or to help control the flow of objects. It appears that a new level of mechanical components will be playing a key part in future casino games.


And speaking of future casino games, it also became very clear that not only will developers demand hardware solutions capable of driving the high resolution content which players expect, but also capable of handling the additional computation required with the integration of these aforementioned technologies—[shameless plug] something our recently launched 2nd Generation AMD Embedded R-Series APU platform is ideal for [/shameless plug].


At the end of the day, this convergence of technologies enables the development of more efficient, yet greater eye-catching solutions which casino game designers are using to set them apart from the rest of their competitors.


Booth front 1_cropped resized.jpg

 

Kelly Gillilan is a Strategic Marketing Manager for Embedded Gaming at AMD. His postings are his own opinions and may not represent AMD’s positions, strategies or opinions. Links to third party sites, and references to third party trademarks, are provided for convenience and illustrative purposes only. Unless explicitly stated, AMD is not responsible for the contents of such links, and no third party endorsement of AMD or any of its products is implied.

At Computex 2014, AMD outlined new initiatives to support the acceleration of server-based workflows and workloads with its professional graphics products.

 

Press Conference.jpg

AMD SKY Technology and AMD STREAM Technology1 were unveiled as solutions to enhance centralized computing for commercial and consumer environments based on industry-recognized AMD professional graphics products. AMD STREAM Technology is designed to leverage the massively parallel processing power of the AMD FirePro™ S-Series server GPUs plus OpenCL™ to accelerate applications beyond just graphics. AMD SKY Technology enables AMD FirePro S-Series and AMD Radeon™ SKY Series graphics cards to power visual cloud applications in commercial and consumer workflows.

 

 

These AMD technologies are making it easier to take advantage of graphics acceleration for remote workstation, virtual workstations and applications, and cloud gaming deployments with SKY, and for HPC (high performance computing) workloads and compute intensive workflows with STREAM.

 

AMD SKY Technology offers the following support for accelerating graphics in the cloud:
- Hypervisor Support
- Multi-GPU Systems Support
- Multi-VM Shared GPU Support
- Remote Access & Display Software Support
- Hardware Accelerated Codecs2

 

AMD STREAM Technology offers the following support to accelerate compute intensive workloads:
- ECC Memory Support
- Fast Single & Double Precision Performance
- P2P Multi-GPU Support
- Bi-directional PCIe® Memory Transfers
- GPU Optimized OpenCL™ Libraries

 

AMD understands the importance of ongoing collaborations with key industry leaders within the software ecosystem to help extend and sustain our presence in the market.

During the AMD press conference at Computex, I was joined onstage by Barry Chen, General Manager Taiwan, from VMware Inc.

Barry Chen - VMware.jpg

AMD and VMware announced certification for VMware Virtual Shared Graphics Acceleration (vSGA) technology with AMD FirePro™ S-Series server cards including the AMD FirePro™ S7000 and AMD FirePro™ S9000 server graphics cards, and AMD FirePro™ W7000 workstation graphics card. With AMD FirePro cards, vSGA-configured virtual machines with the potential to run thousands of applications developed on Microsoft® DirectX® 9 and OpenGL 2.1 APIs. Designed to help streamline system management, help reduce costs and enhance user productivity, AMD SKY Technology plus AMD FirePro S-Series cards with vSGA support help to facilitate migrations from physical PCs to virtual machines. The result delivers high quality graphics performance while addressing GPU density with one GPU shared among multiple virtual machines.

 

 

Going forward AMD and VMWare plan to work closely together to further develop the ecosystem.

 

AMD is also pleased to announce that HP has chosen the AMD FirePro™ S4000X server module for its HP ProLiant WS460c Graphics Server Blade to reduce per user deployment cost as well as operational cost through its high user density. Support for the AMD FirePro S4000X module brings entry to mid-range workstation graphics performance to the blade server environments, with support for up to six displays per module, allows IT to deliver end-user desktops for CAD and Engineering, as well as financial services professionals, from a secure data center.

 

David Cummings S4000X.jpg

So now with the combination of AMD FirePro cards plus AMD SKY and AMD STREAM technologies, we are committed to increasing adoption of professional graphics cards in data centers for GPU Compute and Visual Cloud -- technologies designed to simplify deployment and ensure a high quality experience. 

 

AMD will be rolling out more news in support for AMD SKY and AMD STREAM technologies at International Supercomputing Conference 2014 in Leipzig, Germany in booth #250 from June 23-25, 2014.

 

Learn more about AMD SKY

Learn more about AMD STREAM

 

David Cummings is a senior director and general manager, Professional Graphics at AMD. His postings are his own opinions and may not represent AMD’s positions, strategies or opinions. Links to third party sites are provided for convenience and unless explicitly stated, AMD is not responsible for the contents of such linked sites and no endorsement is implied.

OpenCL and the OpenCL logo are trademarks of Apple Inc. used by permission by Khronos. Apple and Mac Pro are trademarks of Apple Inc., registered in the U.S. and other countries.

 

1 Not all products feature support for every element of AMD STREAM or AMD SKY technology – check with your component or system manufacturer for specific model capabilities.

 

2 AMD does not provide a license/sublicense to any intellectual property rights relating to any to any standards, including but not limited to any audio and/or video codec technologies such as AVC/H.264/MPEG-4, AVC, VC-1, MPEG-2, and DivX/xVid.

AMD’s Embedded G-Series family has just gotten bigger with our latest x86 System-on-chip (SoC), previously codenamed “Steppe Eagle” and the x86 central processing unit (CPU), previously codenamed  “Crowned Eagle.”

 

Last month we introduced the revolutionary second generation AMD Embedded R-Series solution previously codenamed  “Bald Eagle,” today we continue to bring innovation in the embedded market.

 

The Embedded G-Series “Steppe Eagle” SoC is a marvel of engineering, with a 60 percent CPU intensive  performance jump1 from previous G-Series SoC solutions, with a configurable TDP as low as 5W. “Steppe Eagle” is a highly efficient SoC, with a 96 percent improvement in overall performance-per-watt2 and is designed to work in the harshest environments, delivering dependable performance where it is needed the most.

 

The “Steppe Eagle” SoC includes a graphics processing unit based on AMD’s award winning and proven Graphics Core Next architecture. This allows “Steppe Eagle” to produce stunning graphics and because it supports OpenCL™, provides immense compute capability.

 

Our latest Embedded G-Series SoCs also combine enterprise-grade technology such as ECC support and the AMD platform security processor (PSP). AMD PSP is built upon ARM’s proven TrustZone® architecture and provides protection against malicious software threats.

 

Alongside “Steppe Eagle,” we are launching a new 64-bit x86 CPU previously known as “Crowned Eagle.” This CPU is designed to meet the rapidly growing demand for high performance network services due to the proliferation of the cloud. “Crowned Eagle” is designed for networking and communications infrastructure equipment, with features such as PCI-Express Gen 2.0, USB 3.0 and single-channel DDR3-1600 memory with ECC support and TDP as low as 5W.

 

To address the need for demand for data security protocols such as IPSec, “Crowned Eagle” incorporates an on-chip security processor that allows for fan-less security appliances, such as network infrastructure equipment, Network Attached Storage appliances and storage controllers.

 

We have made it easier than ever for our customers to use different members of the Embedded G-Series family. The new AMD G-Series SoC and CPU solutions

are pin-compatible, meaning customers can design common boards and slot in the G-Series SoC of choice without having to invest in hardware and software tooling.

 

To support our Embedded G-Series silicon, we are a committed member of the Linux community. AMD is a gold-level sponsor of the Yocto Project™ - a Linux Foundation Collaboration Project. We have also recently signed a multi-year agreement with Mentor Graphics, a major contributor in the open source community, giving developers access to customized embedded Linux development and commercial support on our AMD G-Series family of chips.

 

The AMD Embedded G-Series has collected numerous press and industry awards. Today the AMD G-Series gets two new family members, extending AMD’s position of product and innovation leadership in the embedded market and serving applications such as networking, industrial control and automation and cloud-based thin client solutions.


Kamal Khouri is Director of Embedded Products at AMD. His postings are his own opinions and may not represent AMD’s positions, strategies or opinions. Links to third party sites, and references to third party trademarks, are provided for convenience and illustrative purposes only. Unless explicitly stated, AMD is not responsible for the contents of such links, and no third party endorsement of AMD or any of its products is implied.

 

1 CPU performance comparison based on Passmark v7 benchmark. The performance delta of 60% was calculated based on GX-212JC’s Passmark v7 CPU Mark score of 1623 and GX-210JA’s Passmark v7 CPU Mark score of 1016. The AMD Steppe Eagle GX-212JC used an AMD Larne motherboard with 4GB DDR3-1333 memory and 320GB Toshiba HDD. The G-S SOC GX-210JA used an AMD Larne motherboard with 4GB DDR3-1066 and 320GB Seagate HDD. Both systems ran Windows® 7 Ultimate. EMB-99

 

2 Overall performance was measured using a suite of industry benchmarks consisting of 3DMark06, 3DMark11, POVRay v3.7, Passmark v7, PCMark8 v2.0, and BasemarkCL 1.0. The GX-412HC’s TDP is 7W and GX-210HA’s TDP is 9W. The performance delta of 53% was calculated based on GX-412HC’s geometric mean of 555.3 and GX-210HA’s geometric mean of 363.6. The performance-per-watt delta of 96% was calculated based on GX-412HC’s performance-per-watt ratio of 79.3 and GX-210HA’s performance-per-watt ratio of 40.4. The AMD Steppe Eagle GX-412HC and G-S SOC GX-210HA used an AMD Larne motherboard with 4GB DDR3-1333 memory and 320GB Toshiba HDD. The system ran Windows® 7 Ultimate. EMB-104

Bristol, England was the focal point for all things OpenCL™ as the academic communities and key stakeholders gathered together for the annual International Workshop on OpenCL (http://iwocl.org/) AMD was a gold sponsor of the annual meeting about OpenCL™ where users, researchers, developers and suppliers shared best practices, while promoting the evolution and advancement of the OpenCL standard.

 

AMD’s participation at the event is just one example our deep commitment to an open ecosystem that benefits the industry and customers alike. 

 

The event began with a keynote presentation from Simon McIntosh-Smith, head of the Microelectronics Research Group at the University of Bristol. His presentation showcased the use of OpenCL along with the AMD FirePro™ S10000 server graphics card with the Bristol University Docking Engine to exploit the performance of modern many-core processors in drug screening. http://research-information.bristol.ac.uk/en/publications/high-performance-in-silico-virtual-drug-screening-on-manycore-processors(9a90a05e-96ae-41c6-82ea-ff0c85468197).html


The conference was visited by attendees from 14 countries, 35 companies and 18 different academic institutions. And just to underline the interest in OpenCL, attendance almost tripled from last year’s conference.

 

AMD joined other exhibitors in the conference with a unique demonstrations. Developed by the AMD FirePro team, the AMD Waterfall demo featured an FTS Celsius R930 Workstation with the AMD FirePro W9100 and a 4K monitor.
demo1.jpg

 

Designed to showcase OpenCL and OpenGL interoperability, in the waterfall particle simulation, the solution computed the collision of each particle in order to roll it along the rocks. The other simulation with a lake featured a 2D grid simulation showing the movement of each node of the grid being computed depending on its neighbor nodes. Both simulations were computed with OpenCL while the rendering was carried out with OpenGL.

 

The AMD motto was “Be Locked or Be Free”, to emphasize to attendees the open source nature of OpenCL. And with a growing community of programmers  and contributing companies to the open standard platform, software developers have access to a growing body of tools, resources and shared “best practices” they can draw from to create powerful applications.

 

IWOCL 2014 presentations are available for download from Agenda & Slides from 2014 | IWOCL

 

Learn more about OpenCL developer information on AMD.com

locked.jpg


JC Baratault is a senior business development manager, global GPU computing, Professional Graphics at AMD. His postings are his own opinions and may not represent AMD’s positions, strategies or opinions. Links to third party sites are provided for convenience and unless explicitly stated, AMD is not responsible for the contents of such linked sites and no endorsement is implied.

OpenCL and the OpenCL logo are trademarks of Apple Inc. used by permission by Khronos. Apple and Mac Pro are trademarks of Apple Inc., registered in the U.S. and other countries.

This post has been authored by Kevin Tanguay who is a Director of Embedded Gaming Products at AMD

 

AMD’s second generation embedded R-series Accelerated Processing Unit (APU) formerly codenamed  “Bald Eagle” has landed! The second generation embedded R-series APU combines AMD’s expertise in designing high performance, multi-core x86 CPU architectures and multipurpose programmable GPUs into a single chip that supports four independent displays and Heterogeneous System Architecture (HSA), making it a great choice for casino and arcade gaming.

 

AMD’s APUs have already proven themselves in the embedded market, with the first-generation AMD R-series raising the bar for levels of processing power, visualization and power efficiency. The second generation embedded R-series APU builds on the success of the first generation R-series APUs by making use of AMD’s Steamroller CPU architecture, powering up to four x86 CPU cores. Coupled with the latest CPU architecture is AMD’s award-winning AMD Radeon™ HD 9000-series GPU architecture that supports Microsoft DirectX 11.1, OpenGL 4.2 and OpenCL all within a 35 W power envelope.

 

The second generation embedded R-series APU brings together the latest AMD CPU and GPU architectures to create a phenomenal APU for the embedded market. AMD’s second generation embedded R-series APU is the first embedded processor to support HSA. This allows applications to seamlessly make use of the CPU and GPU through Heterogeneous Queuing and access the same memory through Heterogeneous Unified Memory Architecture.

 

AMD’s R-series APUs have always supported GPU compute but with has support, the developer now has even more tools at their disposal to access the phenomenal compute found in the Radeon HD 9000-series GPU in every second generation embedded R-series APU. HSA features enable higher resolution video playback and increased number of video streams, meaning that in a casino the gamer can see more videos at a higher quality, improving their experience.

 

Complementing the revolutionary CPU and GPU architectures and HSA support, our engineers have also added features such as DDR3-2133 and ECC memory support, along with PCI Express Gen 2 and Gen 3 support. What does all this mean? The increase in memory bandwidth and support for higher bandwidth to peripherals means the R-series APU not only can access more data, but can also support a second GPU through AMD Radeon™ Dual Graphics technology. A second GPU can turbocharge graphics performance or provide output for even more displays.

 

AMD’s embedded R-series APUs have been the ultimate solution when it comes to driving multiple displays with eye-popping graphics, and our second generation embedded R-series APUs continue this rich tradition. The second generation R-series APUs support up to four independent displays, driven through today’s most common standards including HDMI, DisplayPort 1.2, UVD 4.2 and VCE 2.0.

 

AMD has been working closely with RAD Game Tools making RAD’s popular industry standard Bink 2 video codec work flawlessly on AMD R-series APUs. Bink 2 can be found in most popular games and the latest generation game consoles such as Microsoft’s Xbox One and Sony’s PlayStation 4, which also feature AMD processors. RAD’s Bink 2 codec is also popular in the casino gaming industry, where the need for high-quality video is essential to meet users’ expectations for a quality gaming experience. AMD’s close relationship with RAD Game Tools has resulted in Bink 2 becoming the first version of the software video decoder to utilize the GPU for computation and makes the R-series APU a superior processor to play back content encoded with Bink 2.

 

RAD Games is now shipping Bink 2.4 with optional GPU decoding.  This version of Bink offloads the video decoding using compute shaders on Windows, Linux, Sony’s PlayStation4 and Microsoft’s Xbox One. This approach results in two to four times faster decoding than using CPU-only decode (and even more for 4K video).1 For example, 4K video frames can be decoded in 2.3 ms on Sony’s PlayStation4 or Microsoft’s Xbox One , and 1.4 ms on a PC!

 

The second generation embedded R-series APUs support Microsoft Windows and Linux, with Microsoft DirectX™ 11.1 and OpenGL 4.3 supported. Naturally, AMD’s R-series APUs support OpenCL, giving developers access to computation power held within the Radeon HD 9000-series GPU. This means whether your applications run on Windows or Linux, you can make use of AMD’s R-series APUs.

 

The second generation embedded R-series APU showcase AMD’s commitment to the gaming industry, with a high-performance embedded APU that includes the latest technology combined with strict industry certification. With dual-core and quad-core R-series “Bald Eagle” APUs available in 17 W and 35 W power envelopes, AMD’s latest generation of x86 R-series APUs provide high-performance and power-efficient processing power to drive multidisplay installations, making them ideal for casino and arcade gaming machines.

 

 

Kevin Tanguay is Director of Embedded Gaming Products at AMD. His postings are his own opinions and may not represent AMD’s positions, strategies or opinions. Links to third party sites, and references to third party trademarks, are provided for convenience and illustrative purposes only. Unless explicitly stated, AMD is not responsible for the contents of such links, and no third party endorsement of AMD or any of its products is implied.

If you follow AMD closely, you likely know that we design and integrate technology that powers millions of intelligent and connected devices, from the world’s fastest graphics card to supercomputers, tablets and game consoles. But what you might not know is our technology also powers countless embedded solutions like some of the latest electronic and radio frequency (RF) test equipment, geographic information systems, retail signage, aerospace solutions and medical equipment. Embedded computing is an integral part of the AMD product portfolio and strategy. It’s a focus area where we are blazing new paths of innovation and that’s gaining significant traction with x86 AMD Embedded G- and R-Series CPU’s, APUs, and SoCs along with the AMD Embedded Radeon™ graphics processing units (GPUs).

 

Now comes news from CoreAVI, a long-time technology partner, that Boeing has selected AMD embedded technology for its next generation of high performance avionics cockpit display systems. Working together with CoreAVI, AMD embedded solutions enable the full capabilities of mission critical visual systems, including compute and graphics processors, multi-independent 3-D display outputs, and H.264/MPEG2 Universal Video Decoders that enable a state-of-the-art visual display system for pilots.


Needless to say, we’re thrilled by Boeing’s selection. It’s an excellent example of AMD embedded technology at work and it validates our approach to creating embedded solutions into an integrated and compelling offering in addition to a number of other examples where our embedded solutions are being used (i.e., digital signage, casino and arcade gaming machines, portable ultrasound systems, DNA analysis, vision control systems and advanced robotics). I think it’s easy to say our technology is used nearly everywhere.


Avionics are at the heart of flight systems today and the level of complexity of these systems is tremendous. As such, this use-case is a great example of the AMD embedded value proposition. AMD embedded solutions are optimized to handle 3-D mapping and image manipulation, high-speed data streaming, as well as the massively parallel processing required for tasks like radar processing and object recognition. What’s more, many are ideal for fanless requirements such as what might be needed in the cockpit. Setting the optimal balance between processing performance, power consumption and heat dissipation is especially crucial, as any significant skewing of these properties can negatively impact the performance profile of the entire cockpit electronics system. AMD Embedded GPUs offer the ability to clock performance up or down as needed to provide greater overall power scalability and thermal control.


The recently introduced AMD Radeon™ E8860 GPU is an excellent example of bringing to market world-class graphics technology for embedded solutions. Providing 768 GFLOPS of precision floating point performance and supporting thermal design power (TDP) of 37 watts, the GPU provides an optimal performance-per-watt profile for applications like cockpit electronics and display systems that require the highest level of graphics clarity and accuracy without sacrificing energy efficiency. This is made possible, in part, by its massive parallel compute capability which optimizes the data processing path to improve real-time video and graphics processing performance.


The news from Boeing is just the latest example of the value provided by AMD embedded solutions. With new processors – both x86 and ARM – this year, the momentum is sure to continue growing.


Kamal Khouri is Director of Embedded Products at AMD. His postings are his own opinions and may not represent AMD’s positions, strategies or opinions. Links to third party sites, and references to third party trademarks, are provided for convenience and illustrative purposes only. Unless explicitly stated, AMD is not responsible for the contents of such links, and no third party endorsement of AMD or any of its products is implied.

llatif

The Best of Times

Posted by llatif Apr 25, 2014

This post was authored by Gary Frost, a Software Fellow at AMD


Here at AMD we are committed to provide GPU compute performance from a variety of programming languages. We understand that not every developer will have the flexibility or inclination to port embarrassingly parallel sections of code to OpenCL in order to take advantage of the energy and performance advantages afforded by modern SIMD-style accelerators.

 

In November of last year I had the opportunity to showcase the state of the OpenJDK Sumatra project at AMD’s APU13 developer summit. “Sumatra” is a joint AMD/Oracle project which allows the Java Virtual Machine’s JIT (Just In Time) compiler infrastructure to generate code suitable for GPU offload. When fully realized this will allow Java developers to see their Java code accelerated by the JVM and dispatched to the GPU automatically at runtime.


At APU2013 Nandini Ramani (Vice President of Java Platform at Oracle Corporation) and Phil Rogers (Corporate Fellow at AMD) made some time in their keynote presentations to call out the work of the Sumatra team and to highlight the hardware and software features that allow Sumatra to bring GPGPU compute to the Java community.


Specifically in Nandini’s keynote I demonstrated our ‘Dickens’ demo. This showed how a simple search algorithm coded using standard Java 8 patterns and idioms could be executed on a HSA enabled platform. The code created a histogram of names present in a subset Charles Dickens’ novels. Thanks to HSA’s shared virtual memory, the Java user interface was able to update in real time as the Java code searched through the text. With the Java 8 Stream APIs the user can switch easily between a sequential and parallel implementation, and on a HSA enabled platform the Sumatra enabled JVM was able to execute parallel fragments directly on the GPU cores.


From a performance point of view, at APU13 we successfully demonstrated   the JVM seamlessly migrating this Java workload from CPU multicore APU graphics cores with very little effort from the Java developer.


This demo was well received, but was essentially unrepeatable by the general public as we were showing an early Sumatra ‘drop’ running on a pre-release Java 8 JVM, on a prerelease internal HSA runtime and on prerelease “Kaveri” APU hardware on a set of Windows drivers specially composed for the demonstration. In the words of Phil Rogers “how could this possibly go wrong?” 


What a difference six months makes. In January of this year AMD released the new AMD A-Series APU (code named “Kaveri”) allowing all sorts of applications to take advantage of the GPU cores on this processor.


At the end of February the HSA Linux driver team within AMD made available the Linux Kernel patches and drivers to allow the Linux kernel to coordinate the execution of code on the GPU compute units within a HSA enabled platform.


Also in February the HSA runtime team from AMD made available an early access Linux HSA runtime, which allows projects such as Sumatra to dispatch HSA kernels on the HW from user mode.


To align with this the Sumatra developers (both at AMD and over at Oracle) added their first round of HSA support to the Graal infrastructure which Sumatra uses to generate code for GPUs.


At the end of March Oracle released Java 8.


So now the stars have aligned and anyone with a HSA compatible Linux kernel on HSA compatible hardware can recreate the demos we showed at APU2013 from publically available software and hardware.


Now in April, we have showcased all of these components running on Fedora at the Red Hat Summit. Congratulations to all the folk at AMD, the HSA foundation, and our friends at Oracle for getting us to this point.


We have Great Expectations for the future of Sumatra. 



Gary Frost is a Software Fellow at AMD. His postings are his own opinions and may not represent AMD’s positions, strategies or opinions. Links to third party sites, and references to third party trademarks, are provided for convenience and illustrative purposes only.  Unless explicitly stated, AMD is not responsible for the contents of such links, and no third party endorsement of AMD or any of its products is implied.

This post was authored by Young-Sae Song, CVP of Product Marketing at AMD


The great thing about OpenStack is its community driven approach that allows the developers to shape and mold the final outcome limited only by their creativity and drive to create useful software.  However, the downside is that the efforts and talents of the developer community can be spread thin and the features and capabilities that are critical for an end-to-end operationally scalable solution do not get the attention they deserve.  Canonical has recognized this problem and with Ubuntu 14.04 LTS delivered an OpenStack cloud platform that is able to deliver any workload to bare metal or virtual machines.  It is designed to scale services faster than any OpenStack platform today for virtual or bare metal machines.  You can read more about Canonical’s release announcement here .

 

AMD’s SeaMicro SM15000™ microserver is ideally suited for massive OpenStack deployments. AMD and Canonical collaborated closely to leverage the SeaMicro SM15000 microserver’s unique capabilities to provide the scalability and flexibility to build and manage large-scale public and private clouds with virtualized or bare metal servers.

SeaMicro SM15000 Microserver

 

  • SeaMicro SM15000 microserver integrates compute, storage and Layer 2 networking and creates pools of resources to build custom bare metal server configurations for OpenStack
    • The Freedom™ fabric disaggregates compute, storage and network I/O to provide the most flexible, scalable and resilient data center infrastructure in the industry
    • The result is an adaptive data center where any server can be mapped to any hard disk/SSD or network I/O to expand capacity or recover from a component failure

 

The SeaMicro SM15000 microserver’s RESTful APIs make bare metal or virtual machine provisioning simple and accelerates mass deployments.  AMD’s SeaMicro SM15000 with Juju and MAAS (Metal as a Service) integration is the fastest, most efficient and most scalable way to build scale out infrastructure. MAAS, part of Ubuntu 14.04 LTS and Ubuntu OpenStack, provisions the server, storage and networking to make physical server provisioning as easy as virtual servers.  The solution is available today and is the most scalable, automated option to deploy OpenStack in hyperscale environments.

 

Key SeaMicro SM15000 and Ubuntu 14.04 LTS Integration Capabilities

 

  • SeaMicro SM15000 with Juju and MAAS accelerates server deployment and setup by making bare metal servers available for OpenStack workloads in minutes.
  • MAAS leverages the Freedom fabric to turn the physical servers on the SeaMicro SM15000 microserver into pools of malleable computing and storage resources to create a flexible server infrastructure that can be expanded independently without adding unnecessary compute or storage capacity.
  • SeaMicro SM15000 with Juju/MAAS integration provides the fastest and most flexible way of deploying workloads, be they web, big data on bare metal or building OpenStack cloud infrastructure. Juju & MAAS gives this flexibility of cloud on bare metal servers to make them available for workloads in minutes

 

The AMD and Canonical collaboration is a big step forward in creating an end-to-end solution that is ready to be deployed by enterprises and service providers. This was achieved by ensuring that the OpenStack software was able to take advantage of the unique capabilities of the underlying hardware that provided unique benefits and provided capabilities that otherwise may not have been possible.  As the OpenStack community continues to plow forward, it is this type of close collaboration among hardware and software companies that will ensure that future OpenStack releases are “deployment ready” versus something that is more suited for small trials.

 

Young-Sae Song is a Corporate Vice President of Product Marketing at AMD. His postings are his own opinions and may not represent AMD’s positions, strategies, or opinions. Links to third-party sites and references to third-party trademarks are provided for convenience and illustrative purposes only. Unless explicitly stated, AMD is not responsible for the contents of such links, and no third-party endorsement of AMD or any of its products is implied.

This post was authored by Young-Sae Song, CVP of Product Marketing at AMD

AMD continues to demonstrate its commitment to OpenStack as a leading member of the Red Hat OpenStack Cloud Infrastructure Partner Network. The SM15000 server has achieved Red Hat Enterprise Linux OpenStack Certification and provides OpenStack In A Box for enterprise-ready, commercial-grade OpenStack infrastructure.

The AMD SM15000 server solution provides 64 servers with up to 64GB of memory and consumes about 55 watts per server. Working closely with Red Hat, AMD offers customers an enterprise-ready and commercial-grade OpenStack infrastructure, with complementary technologies across compute, networking and storage.

AMD’s SeaMicro SM15000 system is the highest-density, most energy-efficient server in the market and is designed from the ground-up for today’s workloads. In 10 rack units, it links 512 compute cores, 160 gigabits of I/O networking, more than five petabytes of storage with a 1.28TB/s high-performance supercompute fabric, called Freedom™ Fabric. The SM15000 server eliminates top-of-rack switches, terminal servers, hundreds of cables and thousands of unnecessary components for a more efficient and simple operational environment.

AMD’s SeaMicro server product family currently supports the next generation AMD Opteron™ (“Piledriver”) processor, Intel® Xeon® E3-1260L (“Sandy Bridge”), E3-1265Lv2 (“Ivy Bridge”), E3-1265Lv3 (“Haswell”) and Intel® Atom™ N570 processors. The SeaMicro SM15000 server also supports the Freedom Fabric Storage products, enabling a single system to connect with more than five petabytes of storage capacity in two racks. This approach delivers the benefits of expensive and complex solutions such as network attached storage (NAS) and storage area networking (SAN) with the simplicity and low cost of direct attached storage.

AMD’s SeaMicro SM15000 is the simplest and fastest way for companies to deploy OpenStack. The certified solution replaces what typically requires multiple racks and cuts energy consumption by up to 50 percent. It combines compute, networking, and storage in a single 10 RU system all linked by a highly efficient supercompute fabric to power the computing and storage needs of the next generation data center.

seamicro_sm15000_openstack.jpg

For more information on how AMD’s SeaMicro SM15000 servers are the ideal solution for OpenStack deployments, please view www.amd.com/openstack.

 

Other resources:

Accelerate Your ROI on OpenStack Private Cloud:

http://seamicro.com/webinar/openstacksolution

 

OpenStack Design and Implementaion:

http://www.youtube.com/watch?v=yEsLgcxwePU&feature=youtu.be

 

Key Learnings from Actual OpenStack Deployments

http://www.youtube.com/watch?v=nLasAGj01ps&feature=youtu.be

This post has been authored by Leendert van Doorn, Corporate Fellow at AMD


Developing a software ecosystem to support new hardware technology is hard work. For our ARM-based AMD Opteron™ A1100 Series processor, code named “Seattle,” AMD is working with several key organizations to enable a standard-based server class software foundation – firmware, operating systems, hypervisors, device drivers, and development tools. For me this effort involves a lot of travel to meetings and technical discussions with strategic partners and customers.

 

One of my recent trips was to Linaro Connect Asia in Macau. In case you didn’t know, Linaro is a not-for-profit engineering organization developing open source software for the ARM architecture. Linaro Connect is a great collaboration event with a mixture of discussion, planning and agreement about engineering projects being run within Linaro. My focus is on making sure the software foundation that is being developed by Linaro adheres to recognized server standards. My goal is to avoid the “wild west” approach that we have seen in the 32-bit ARM world – with specialized software tuned only for specific devices. We need ARM servers to fit with minimal disruption into the existing world of x86 servers – so that data centers can have a clear option on the technology they choose to drive their business.

 

The Linaro connect trip to Macau was exciting for a number of reasons. First of all, more ARM server partners joined Linaro, most notably Qualcomm. All the key ARM server players are now members of Linaro, making it the perfect forum to drive the enablement of open-source ARM server software support for industry standards.

 

One such example is ACPI, which despite some of the controversy around it in the Linux community, is the right standard to move forward for Hyperscale because it reduces the disruption in the data center where all the tooling is based on these industry standards. Linaro is actively driving the enablement of ACPI both in UEFI (the firmware stack) and the Linux kernel. This kind of progress was very exciting to experience first-hand.

 

As a result of enabling these industry standards in ARM servers software, implementing them becomes literally as simple as flipping a (compile-time) switch and with all key server partners being Linaro members, there is little excuse for them not to adopt these standards. The benefits of this teamwork extends to our customers as well as ARM server vendors.

 

So what does it mean to adhere to recognized server standards? Take for example AMD’s “Seattle” SOC, it consists of industry standard ARM Cortex A57 cores (with SIMD, cryptography and TrustZone® extensions) together with industry standard interfaces such as SATA, PCIe, and NICs. This allows us to be interoperable at the hardware level with existing RAID controllers, hard disks, and a host of other data center peripherals. Similarly, at the software level we are adhering to industry standards as well such as UEFI for the firmware, ACPI, SMBIOS and PSCI to interface to the operating system and IPMI for out of band system management. Conforming to these standards allows us to seamlessly plug into existing infrastructure, thereby reducing the barrier for adoption and, more important, reducing the cost of entry for our customers.

 

One of my next stops is at Red Hat Summit in April in San Francisco. Red Hat and AMD have a long history of close collaboration – dating back to efforts around bringing 64-bit CPU technology to the x86 world. Our current efforts with Red Hat include our collaboration around Hyperscale computing along with the support for our 64-bit ARM server technology.

 

In fact, I will be participating on Hyperscale computing panel hosted by Jon Masters, Chief ARM Architect at Red Hat. The panel will include folks from Google, IBM, and ARM and will discuss challenges, opportunities, and emerging trends in Hyperscale computing. If you are attending the Red Hat Summit I encourage you to join what looks to be a lively discussion. The Session is on Wednesday, April 16, 15:40 to 16:40PST.

 

This is an exciting time for AMD…play to win!

 

Leendert

 

Leendert van Doorn is a corporate fellow at AMD. His postings are his own opinions and may not represent AMD’s positions, strategies, or opinions. Links to third-party sites and references to third-party trademarks are provided for convenience and illustrative purposes only. Unless explicitly stated, AMD is not responsible for the contents of such links, and no third-party endorsement of AMD or any of its products is implied.

AMD and Leap Computing took to the sky at the Game Developers Conference (GDC) 2014 (Booth# 1024) in San Francisco, showcasing the evolution of cloud gaming.

 

AMD Radeon™ Sky cloud gaming cards and AMD RapidFire technology enable Leap Computing to deliver a fully-realized, turnkey cloud solution for uncompromising performance on games like Battlefield 4, Dirt 3 and Tomb Raider. As a bonus, it’s also a cost-effective option for video game developers.

 

 

 

While at the show I presented a session entitled “RapidFire: The Easy Route To Low Latency Cloud Gaming Solutions”. This was an opportunity for attendees to learn more about how AMD's RapidFire SDK simplifies the delivery of multi-game streaming from a single GPU while working to minimize latency to ensure one of the best cloud gaming experiences.

 

At the AMD exhibit, the AMD RapidFire and Leap Computing streaming gameplay demonstration generated a great deal of interest and inquiries by showcasing the next major step in cloud gaming that allowed attendees to play popular titles via streamed content on laptops.

photo2.JPG.jpg

 

“Although these technologies have existed for the past several years, the play experience has been too resource-heavy to actually make cloud deployment possible, until now,” said Alexander Nataros, chief executive officer at Leap Computing. “The performance and precision of the AMD Radeon Sky cloud gaming GPUs have been instrumental in optimizing hardware for our service. AMD has positioned itself as an industry leader in the emerging cloud graphics space through both the AMD RapidFire technology API and its Direct Output Post Processing (D.O.P.P.) support. With these tools, Leap Computing has been able to provide a graphically intense, low-latency environment that would not have been otherwise possible.”

 

AMD Radeon™ Sky Series products support up to six HD game streams at once(1). Furthermore, service providers can maximize existing infrastructure and resources to support even more simultaneous streams, from casual games up to the most demanding AAA titles. AMD Radeon Sky Series products are designed to support a wide range of servers and systems.

 

Additionally, AMD RapidFire technology enables cloud gaming partners to benefit from an open API, simplifying manipulation of key hardware controls to provide HD visual quality, minimal latency and optimal network bandwidth, resulting in a compelling and responsive experience on virtually any device. In line with AMD’s commitment to industry standard APIs, like OpenCL™, DirectX® and OpenGL, an open standard API for cloud gaming can help align the industry around one platform and drive continued innovations.


For more information on AMD FirePro graphics solutions, visit: www.fireprographics.com

 

Allen Bourgoyne is the director of ISV Alliances for Professional Graphics at AMD. His postings are his own opinions and may not represent AMD’s positions, strategies or opinions. Links to third party sites are provided for convenience and unless explicitly stated, AMD is not responsible for the contents of such linked sites and no endorsement is implied.

 

(1)  Test conducted at AMD measuring the ability of a Colfax CX 1250-N4 1U rack mount server with Ciinow Cumulus Cloud Services version 2.0 running on an AMD Opteron™ 6380 16 core Server processor with one AMD Radeon Sky Series model 700, 32GB RAM, video driver 12.10.17.1 to stream to games simultaneously. At 60 fps and 720p resolution, 3 streams were achieved; at 30fps and 720p, 6 streams were achieved. 3 games: LEGO® Batman™, Harry Potter™ Years 1-4, and Devil May Cry; 6 games: Trine, LEGO® Batman™, LEGO® Harry Potter™ Years 1-4 and Years 507, Far Cry 3, CardBoard Castle. FP-77

 

OpenCL and the OpenCL logo are trademarks of Apple Inc. used by permission by Khronos.

Filter Blog

By date:
By tag: