Xbox One will have a high-performance custom chip

IDG News Service - Microsoft's upcoming Xbox One gaming console will contain a custom chip the company designed in conjunction with Advanced Micro Devices with the aim of delivering maximum graphics performance, presenters said Monday at Microsoft's Hot Chips conference.

The high-performance, low-power chip has specialized processors beyond the CPU and GPU to handle tasks such as audio processing, video decode and encode and other small game characteristics. The system has 5 billion transistors, and uses an eight-core AMD CPU code-named Jaguar, which is also being used in the Sony's upcoming PlayStation 4.

The graphics processor is a Radeon GPU that has been customized for the Xbox One, Microsoft presenters said at the conference in Stanford, California.

"Almost every aspect has been customized," said John Sell, a hardware architect at Microsoft.The Xbox One system chip also has 500GB of storage, 8GB of DDR3 memory and 47MB of on-die storage that could act as cache, where information will be stored temporarily. The GPU can provide more than a teraflop of peak performance.AMD's x86 CPUs are based on the Jaguar core, which was introduced for PCs, laptops and servers last year.

The eight CPU cores are broken into clusters of four cores with a total of 4MB of L2 cache.

"We've made some alterations to the CPU clusters to support coherent bandwidth between clusters...and other processors," Sell said.

The graphics processor supports DirectX 11.1, which is Microsoft's graphics engine that will power games. Microsoft calls the engine 1-plus, a reference to unique control processor features for custom graphics and processing in the Xbox graphics core.

"These have been customized to significantly reduce the amount of time, the amount of work, that the CPU has to spend when assembling graphics commands," Sell said.

One unique aspect of the chip is a shared memory pool that can be accessed by CPUs, GPUs and other processors in the system. Typically, GPUs and CPUs have different memory systems, but the new features increase the overall addressable memory in the Xbox One.

The GPUs and CPUs have also been modified to enable shared memory.Shared memory is also part of a specification being pushed by the HSA Foundation, which wants to blur the line between GPU and CPU memory to make programming easier. AMD is one of the founding members of the HSA Foundation, though Sony is also a member, which suggests that shared memory may also be part of PlayStation 4.

"We can let developers have access to direct programming on GPU without a lot of overhead," Sell said.

Overall, the system has been designed to balance bandwidth with the power requirements.

"An important aspect of that is letting CPU and GPU share memory spaces," Sell said.The coherency connections are a derivative of AMD's years-old HyperTransport technology. The bandwidth can sustain data transfers of 30GB per seconds. However, the data transfers vary depending on the component and also the distance between parts.The chip will be made by Taiwan Semiconductor Manufacturing Co. (TSMC) using the 28-nm process, Sell said.

Reprinted with permission from IDG.net. Story copyright 2012 International Data Group. All rights reserved.

View the original article here

Xbox One's massive, custom CPU can pick you out of a lineup

Microsoft’s Xbox One has a problem: you, and your family.

One of the problems that Microsoft had to solve when developing the Xbox One game console was a simple one: how to tell apart the player using the interactive gaming console from his friends, the lamp, the couch, and the coffee table. To solve that issue, Microsoft developed its own light sensor that can figure out what it’s seeing by a variety of calculations.

In talking up the innards of its upcoming gaming console, Microsoft didn’t reveal any new capabilities of the One—due this November—during a presentation at the Hot Chips conference Monday at Stanford University. But company executives discussed the so-called “speeds and feeds” of the main Xbox One processor in detail, as well as the problems that the new Kinect team had to overcome.

MicrosoftThe Xbox One system diagram.

Physically, the system-on-a-chip at the heart of the Xbox One is 363 square millimeters. But the real whopper is the amount of logic integrated within it: 5 billion transistors. Although Wikipedia isn’t necessarily the final arbiter, the Xbox One is possibly the largest chip manufactured to date, with a comparable number to Intel’s 64-core Xeon Phi coprocessor for supercomputers, and easily topping Intel’s 8-core “Poulsen” Itanium that the chip maker launched in 2012.

MicrosoftThe Xbox One system-on-chip.

Unfortunately, Microsoft didn’t disclose how much power that chip will consume. However, the chip will be power gated down to 2.5 percent of its power, said John Sell, a hardware architect at Microsoft who gave the Xbox One silicon presentation at Hot Chips. It will be manufactured in 28-nm process technology by TSMC, Microsoft’s foundry partner.

Nor did Microsoft disclose how fast the chip will run. All told, the Xbox One system contains 500GB of storage, 8GB of DDR3 memory and 47MB of on-die storage that can be used as cache memory.

The system CPU is based on the AMD Jaguar processor architecture; the system uses eight cores with 32 KB of instruction and data cache, with four of the cores connected to a shared 2 MB shared level-2 cache.

MicrosoftThe graphics unit of the Xbox One.

The Xbox One chips are highly customized, Sell said. One significant change from the previous-generation Xbox 360 is that the graphics core is integrated with the CPU, as are most of today’s PCs chips. Sel referred to the GPU chip as a “DirectX 11.1+” chip, as it contains optimized graphics processing unique to the Xbox One. “There’s more than a CPU’s worth of processing in there,” Sell said.

Each Xbox One will contain a Kinect depth camera, designed by Microsoft. (The camera used in the Xbox 360 was designed by PrimeSense, an Israeli developer.)

Patrick O’ Connor, another system designer for Microsoft, explained that the chip can resolve players less than 1 meter away, and players at more than 2 meters away, with a 70-degree field of view. Microsoft also has to take into account a whole range of environments—from players that like to play games in the dark to broad daylight.

The Kinect on the Xbox One actually “illuminates” the room with its own modulated light. What the camera’s 512 x 424 sensor “sees” is the reflected light, which will be out of phase depending on the way it reflects from objects. The Xbox One then adds or subtracts their values to weed out the ambient light, determine the player’s location and motion, and what is the “active” image. Each Xbox One can distinguish up to six different players, O’Connor said, but exactly “who” is playing at any one time takes some figuring.

MicrosoftSidelighting could confuse the Xbox One camera without supplemental illumination, Microsoft said.

What the Xbox One can do, however, is determine, on a per-pixel basis, how best to “see” the object. Since the camera can’t actually lengthen its exposure or widen its aperture, it needs to use different shutter times, before picking the best one. Combining the Kinect’s own illumination with intelligent “reading” of the image eliminates common problems, such as sidelighting that can confuse the camera, O’Connor said.

View the original article here

Intel expands custom processor business

Intel is putting a sharp focus on expanding its custom processor and chip operations in response to a growing trend of companies building servers in-house to meet specific workloads or data center designs.

Last year, Intel has supplied custom processors to 18 companies, most notably eBay and Facebook, said Diane Bryant, senior vice president and general manager of Intel's Datacenter and Connected Systems Group, in an interview.

"That trend is growing. In the last year we've delivered 18 custom silicon processor solutions for the full array of customers -- our direct customers, the OEMS and the end users -- in order to meet their specific needs," said Bryant, who runs Intel's most profitable group.

The growing custom processor business will supplement Intel's bread and butter server chip business, consisting of selling generic Xeon processors available in rack, tower and blade servers sold by companies like Dell and Hewlett-Packard. But server infrastructures are changing with the growing adoption of cloud computing, big data and other applications, which has translated into a growing demand for custom processors, Bryant said.

Companies like Facebook and Google with mega data centers design servers in-house, and get them made from direct server suppliers like Quanta. The barebones servers typically cut excess components and are good enough to handle the growing cloud transactions such as search requests and social networking tasks. Facebook and Google have experimented with ARM processors in servers, while Tilera processors have also been tested.

"When you work with these end users who have technology as their core business, they are very clear on what is required," Bryant said. "They know what their workloads are, what their various applications are, they know what metric they are looking to hit from a performance per total cost of ownership."

The level of processor and chip customization varies with the workload, data center design, and even cooling solutions. Bryant provided an example where a flexible cooling system in a data center would allow customers to run processors at a higher frequency.

"We will have customers that have a very [specific] power target, so we will create versions whether its through changes in frequency, changes in core count, changes to drive down the power," Bryant said.

Customers usually give information about the applications they are running, the accelerators they need, the performance and power consumption levels they are looking to hit. Intel then customizes processors and chips that meets the specifications. Some customers in the technology and the data center business get specific about the server infrastructure.

"It all boils down to scale. I had one cloud service provider who had told me a single application is running across tens of thousands of servers. You can afford to tune that server very targeted against that application and eke out every bit of performance at ever lower cost of operations," Bryant said.

Intel is also now able to build system-on chip (SoC) designs, in which the CPU is combined with other accelerators, I/O, graphics and other processing units. That makes it easier to build custom processors and chips, Bryant said.

"With our SoC capability now, we can actually do rapid turns of our base product with very unique accelerators. Whether it's voice recognition acceleration or encryption or graphics acceleration... all the different types of accelerators that are targeted at different apps. We can deliver unique products there too," Bryant said.

Intel next year will release Xeon server chips based on the Broadwell processor core, which will succeed Haswell. Bryant said that the server SOC will also help optimize the chip to workloads, be it analytics or cloud.

"We have this wonderful Xeon core, and now Intel has a system-on-a-chip capability where we can rapidly turn out grabbing different intellectual property blocks and accelerators. Why not take this Xeon core and marry it with the SoC capability, and come up with... very [specific] processors targeted at unique capabilities," Bryant said.

In some ways, Intel is taking the same route as Advanced Micro Devices, which is creating custom chips based on its CPU and graphics architectures, but largely for non-server products. AMD's custom chips will be used in the upcoming Sony PlayStation 4 and Microsoft Xbox One gaming consoles.

Intel is also investing in software development to tie applications directly to chip development. The chip maker has released its own version of Hadoop, and is also actively contributing an orchestration layer to OpenStack so resources are effectively allocated at server, storage and network levels in distributed computing environments.

Beyond the server, Intel is also looking to change data center design. One of the projects called Rack Scale aims at decoupling the processing, I/O and storage units in data centers with faster throughput mechanisms.

"Instead of a rack being 24 servers slotted in, with each of those servers with compute, memory and I/O, instead break that artificial barrier of the server down and look at it at the rack level. And create pools of compute, memory and I/O so that the application can access and use whatever capacity it needs," Bryant said.

The company is expected to announced a new optical throughput standard called MXC, which will be detailed at the Intel Developer Forum next month. The company is also developing processors for different target markets, Bryant said. Intel will also announce a new Atom processor called Rangeley for embedded networking devices in early September, ahead of IDF.

"We have hundreds of microprocessor products to cover the entire space," Bryant said.

Agam Shah covers PCs, tablets, servers, chips and semiconductors for IDG News Service. Follow Agam on Twitter at @agamsh. Agam's e-mail address is agam_shah@idg.com

View the original article here