This week on TechPowerUp News: Jim Keller joins Tesla Motors as head of the team working on its self-driving tech; and prototype desktops with dual-GPU "Fiji" graphics cards.

Elon Musk handed over the reins of one of Tesla Motors' most important research and development divisions, autopilot, to chip whiz Jim Keller. Keller joined Tesla Motors as Vice President of Autopilot Hardware Engineering. With Tesla being at the very frontier of automobile development, and self-driving cars being the next big thing for the industry, Keller holds an enviable, albeit challenging position.

Jim Keller led teams that designed some of AMD's most commercially successful processors, before a stint at Apple, where he helped it gain hardware self-reliance with the company's Ax series SoCs; and returning to AMD, and leading the team that designed the company's upcoming "Zen" and K-12 micro-architectures. Tesla cars are currently driven by electronics powered by NVIDIA Tegra SoCs. With NVIDIA's immeasurable investments in deep-learning tech that forms the foundation of self-driving car hardware, and Tesla Motors yet choosing a CPU designer to lead its autopilot division, it's easy to speculate that Musk's company is seeking the same kind of hardware self-reliance that Apple did, as its iOS devices were taking off.

NVIDIA revealed more details of its upcoming "Pascal" GPU architecture at the Japanese edition of the Graphics Technology Conference. The architecture will be designed to nearly double performance/Watt over the current "Maxwell" architecture, by implementing the latest tech. This begins with stacked HBM2 (high-bandwidth memory 2). The top "Pascal" based product will feature four 4-gigabyte HBM2 stacks, totaling 16 GB of memory. The combined memory bandwidth for the chip will be 1 TB/s. Internally, bandwidths can touch as high as 2 TB/s. The chip itself will support up to 32 GB of memory, and so enterprise variants (Quadro, Tesla), could max out the capacity. The consumer GeForce variant is expected to serve up 16 GB.

It's also becoming clear that NVIDIA will build its "Pascal" chips on the 16 nanometer FinFET process (AMD will build its next-gen chips on more advanced 14 nm process). NVIDIA is innovating a new interconnect called NVLink, which will change the way the company has been building dual-GPU graphics cards. Currently, dual-GPU cards are essentially two graphics cards on a common PCB, with PCIe bandwidth from the slot shared by a bridge-chip, and an internal SLI bridge connecting the two GPUs. With NVLink, the two GPUs will be interconnected with an 80 GB/s bi-directional data path, letting each GPU directly address memory controlled by the other. This should greatly improve memory management in games that take advantage of newer APIs such as DirectX 12 and Vulkan; and prime the graphics card for higher display resolutions. NVIDIA is expected to launch its first "Pascal" based products in the first half of 2016.

NVIDIA today announced that Microsoft will offer NVIDIA GPU-enabled professional graphics applications and accelerated computing capabilities to customers worldwide through its cloud platform, Microsoft Azure. Deploying the latest version of NVIDIA GRID in its new N-Series virtual machine offering, Azure is the first cloud computing platform to provide NVIDIA GRID 2.0 virtualized graphics for enterprise customers.

For the first time, businesses will have the ability to deploy NVIDIA Quadro-grade professional graphics applications and accelerated computing on-premises, in the cloud through Azure, or via a hybrid of the two using both Windows and Linux virtual machines. Azure will also offer customers supercomputing-class performance, with the addition of the NVIDIA Tesla Accelerated Computing Platform's flagship Tesla K80 GPU accelerators, for the most computationally demanding data center and high performance computing (HPC) applications.

IBM, in collaboration with NVIDIA and Mellanox, today announced the establishment of a POWER Acceleration and Design Center in Montpellier, France to advance the development of data-intensive research, industrial, and commercial applications. Born out of the collaborative spirit fostered by the OpenPOWER Foundation - a community co-founded in part by IBM, NVIDIA and Mellanox supporting open development on top of the POWER architecture - the new Center provides commercial and open-source software developers with technical assistance to enable them to develop high performance computing (HPC) applications.

Technical experts from IBM, NVIDIA and Mellanox will help developers take advantage of OpenPOWER systems leveraging IBM's open and licensable POWER architecture with the NVIDIA Tesla Accelerated Computing Platform and Mellanox InfiniBand networking solutions. These are the class of systems developed collaboratively with the U.S. Department of Energy for the next generation Sierra and Summit supercomputers and to be used by the United Kingdom's Science and Technology Facilities Council's Hartree Centre for big data research.

NVIDIA today unveiled a new addition to the NVIDIA Tesla Accelerated Computing Platform: the Tesla K80 dual-GPU accelerator, the world's highest performance accelerator designed for a wide range of machine learning, data analytics, scientific, and high performance computing (HPC) applications.

The Tesla K80 dual-GPU is the new flagship offering of the Tesla Accelerated Computing Platform, the leading platform for accelerating data analytics and scientific computing. It combines the world's fastest GPU accelerators, the widely used CUDA parallel computing model, and a comprehensive ecosystem of software developers, software vendors, and datacenter system OEMs.

GIGABYTE Technology, a leading creator of high performance server hardware, is happy to announce today the release of its G210-H4G. Developed in partnership with Carri Systems, a French specialist of GPU computing & HPC solutions, this product is a 2U rackmount housing 4 blades capable of receiving one NVIDIA GRID, Tesla or AMD FirePro card each.

In the context of an unfolding GPU virtualization market and the vast amount of new possibilities that it opens, the G210-H4G is an elegant solution designed with simplicity and flexibility in mind to let organizations, big and small, enjoy the benefits of this fast-growing computing trend.

Global supercomputer leader Cray Inc. today announced the launch of the Cray CS-Storm -- a high-density accelerator compute system based on the Cray CS300 cluster supercomputer. Featuring up to eight NVIDIA Tesla GPU accelerators and a peak performance of more than 11 teraflops per node, the Cray CS-Storm system is one of the most powerful single-node cluster architectures available today.

Designed to support highly scalable applications in areas such as energy, life sciences, financial services, and geospatial intelligence, the Cray CS-Storm provides exceptional performance, energy efficiency and reliability within a small footprint. The system leverages the supercomputing architecture of the air-cooled Cray CS300 system, and includes the Cray Advanced Cluster Engine cluster management software, the complete Cray Programming Environment on CS, and NVIDIA Tesla K40 GPU accelerators. The Cray CS-Storm system includes Intel Xeon E5 2600 v2 processors.

Eurotech, a leading provider of embedded and supercomputing technologies, has teamed up with Applied Micro Circuits Corporation and NVIDIA to develop a new, original high performance computing (HPC) system architecture that combines extreme density and best-in-class energy efficiency. The new architecture is based on an innovative highly modular and scalable packaging concept.

Eurotech, which has years of significant experience in designing and manufacturing original HPC systems, has successfully developed an HPC systems architecture that optimizes the benefits of greater density, as well as the energy efficiency of ARM processors and high-performance GPU accelerators.

NVIDIA today announced that multiple server vendors are leveraging the performance of NVIDIA GPU accelerators to launch the world's first 64-bit ARM development systems for high performance computing (HPC).

ARM64 server processors were primarily designed for micro-servers and web servers because of their extreme energy efficiency. Now, they can tackle HPC-class workloads when paired with GPU accelerators using the NVIDIA CUDA 6.5 parallel programming platform, which supports 64-bit ARM processors.

NVIDIA's new Tegra K1 mobile processor will help self-driving cars advance from the realm of research into the mass market with its automotive-grade version of the same GPU that powers the world's 10 most energy-efficient supercomputers. The first mobile processor to bring advanced computational capabilities to the car, the NVIDIA Tegra K1 runs a variety of auto applications that had previously not been possible with such low power consumption.

Tegra K1 features a quad-core CPU and a 192-core GPU using the NVIDIA Kepler architecture, the basis for NVIDIA's range of powerful GPUs -- including the processors that are used in the top 10 systems featured in the latest Green500 list of the world's most energy-efficient supercomputers. Tegra K1 will drive camera-based, advanced driver assistance systems (ADAS) -- such as pedestrian detection, blind-spot monitoring, lane-departure warning and street sign recognition -- and can also monitor driver alertness via a dashboard-mounted camera.

ASUS today announced that its world-leading ESC4000 G2 series server range now supports the next-generation NVIDIA Tesla K40 (Kepler) GPU accelerators.

ESC4000 G2 series ranks very highly in the world's high-performance computing (HPC) tables, claiming the number 59 spot in the November 2013 TOP500 - the definitive list of the world's most powerful commercially available computers. The ESC4000 G2-powered SANAM cluster put in an incredible performance to deliver total processing power of 532.6TFLOPS.

Super Micro Computer, Inc., a global leader in high-performance, high-efficiency server, storage technology and green computing, exhibits its latest high-performance computing (HPC) solutions at the Supercomputing 2013 (SC13) conference this week in Denver, Colorado. In sync with the launch of the NVIDIA Tesla K40 GPU accelerator, Supermicro debuts new 4U 8x GPU SuperServer that supports the new and existing active or passive GPUs (up to 300 W) with an advanced cooling architecture that splits the CPU (up to 150 W x2) and GPU (up to 300 W x8) cooling zones on separate levels for maximum performance and reliability.

In addition, Supermicro has 1U, 2U, 3U SuperServers, FatTwin, SuperWorkstations and SuperBlade platforms ready to support the new K40 GPU accelerator. These high performance, high density servers support up to twenty GPU accelerators per system and in scaled out Super Clusters provide massive parallel processing power to accelerate the most demanding compute intensive applications. Supermicro's new platforms extend the industry's most comprehensive line of servers, storage, networking and server management solutions optimized for Engineering and Scientific Research, Modeling, Simulation and HPC supercomputing applications.

NVIDIA and IBM today announced plans to collaborate on GPU-accelerated versions of IBM's wide portfolio of enterprise software applications on IBM Power Systems. The move marks the first time that GPU accelerator technology will move beyond the realm of supercomputing and into the heart of enterprise-scale data centers. The collaboration aims to enable IBM customers to more rapidly process, secure and analyze massive volumes of streaming data.

"Companies are looking for new and more efficient ways to drive business value from Big Data and analytics," said Tom Rosamilia, senior vice president, IBM Systems & Technology Group and Integrated Supply Chain. "The combination of IBM and NVIDIA processor technologies can provide clients with an advanced and efficient foundation to achieve this goal."

Global supercomputer leader Cray Inc. today announced the Cray CS300 line of cluster supercomputers and the Cray XC30 supercomputers are now available with the NVIDIA Tesla K40 GPU accelerators. Designed to solve the most demanding supercomputing challenges, the NVIDIA Tesla K40 provides 40 percent higher peak performance than its predecessor, the Tesla K20X GPU.

"The addition of the NVIDIA K40 GPUs furthers our vision for Adaptive Supercomputing, which provides outstanding performance with a computing architecture that accommodates powerful CPUs and highly-advanced accelerators from leading technology companies like NVIDIA," said Barry Bolding, vice president of marketing at Cray. "We have proven that acceleration can be productive at high scalability with Cray systems such as 'Titan', 'Blue Waters', and most recently with the delivery of a Cray XC30 system at the Swiss National Supercomputing Centre (CSCS). Together with Cray's latest OpenACC 2.0 compiler, the new NVIDIA K40 GPUs can process larger datasets, reach higher levels of acceleration and provide more efficient compute performance, and we are pleased these features are now available to customers across our complete portfolio of supercomputing solutions."

NVIDIA today unveiled the NVIDIA Tesla K40 GPU accelerator, the world's highest performance accelerator ever built, delivering extreme performance to a widening range of scientific, engineering, high performance computing (HPC) and enterprise applications.

Providing double the memory and up to 40 percent higher performance than its predecessor, the Tesla K20X GPU accelerator, and 10 times higher performance than today's fastest CPU, the Tesla K40 GPU is the world's first and highest-performance accelerator optimized for big data analytics and large-scale scientific workloads.

NVIDIA is readying its next big single-GPU compute accelerator, the Tesla K40, codenamed "Atlas." A company slide that's leaked to the web by Chinese publication ByCare reveals its specifications. The card is based on the new GK180 silicon. We've never heard of this one before, but looking at whatever limited specifications that are at hand, it doesn't look too different on paper from the GK110. It features 2,880 CUDA cores.

The card itself offers over 4 TFLOP/s of maximum single-precision floating point performance, with over 1.4 TFLOP/s double-precision. It ships with 12 GB of GDDR5 memory, double that of the Tesla K20X, with a memory bandwidth of 288 GB/s. The card appears to feature a dynamic overclocking feature, which works on ANSYS and AMBER workloads. The chip is configured to take advantage of PCI-Express gen 3.0 system bus. The card will be available in two form-factors, add-on card, and SXM, depending on which the maximum power draw is rated at 235W or 245W, respectively.

Global supercomputer leader Cray Inc. today announced the Company has broadened its support for accelerators and coprocessors, and is now selling the Cray XC30 series of supercomputers with NVIDIA Tesla K20X GPU accelerators and Intel Xeon Phi coprocessors. This marks the latest step in Cray's Adaptive Supercomputing vision, which is focused on delivering innovative systems that integrate diverse technologies like multi-core and many-core processing into a unified architecture.

"Our first experience with climate and materials science applications showed that replacing one of the multi-core processors in the XC30 with an NVIDIA Tesla GPU boosts application performance and disproportionally reduced energy to solution," said Thomas Schulthess, professor at ETH Zurich and director of the Swiss National Supercomputing Center, which was one of the first Cray customers to order a hybrid Cray XC30 system. "This provides necessary proof of principle in favor of hybrid compute nodes as a promising solution to the energy challenges we face in supercomputing."

TechPowerUp released GPU-Z 0.7.2, the latest version of the popular graphics hardware information, diagnostic, and monitoring utility that enthusiasts and overclockers can't leave home without. Version 0.7.2 adds support for new GPUs, notably NVIDIA's upcoming GeForce GTX 760, and the new Radeon HD 8970M; improves support for Intel HD 4xxx series graphics embedded into Core "Haswell" processors, and a few more user-interface feature additions.

To begin with, GPU-Z 0.7.2 adds support for NVIDIA's GeForce GTX 760, GeForce GT 740M (based on the new GK208 silicon), Tesla K10 compute accelerator; AMD's Radeon HD 8970M, HD 8490, and HD 7400D. Detection and information of Core "Haswell" integrated GPUs, are improved. A new AMD Radeon logo was added, and will show up for AMD-branded Radeon GPUs. Tooltip translations were added for Greek, French (improved), and thanks to our friends at Clube do Hardware, Brazilian Portuguese. A rare crash during DirectCompute detection, is fixed.DOWNLOAD: TechPowerUp GPU-Z 0.7.2, GPU-Z 0.7.2 ASUS ROG-themed

The change-log follows.

Researchers at the University of Illinois at Urbana-Champaign (UIUC) have achieved a major breakthrough in the battle to fight the spread of the human immunodeficiency virus (HIV) using NVIDIA Tesla GPU accelerators, NVIDIA today announced.

Featured on the cover of the latest issue of Nature, the world's most-cited interdisciplinary science journal, a new paper details how UIUC researchers collaborating with researchers at the University of Pittsburgh School of Medicine have, for the first time, determined the precise chemical structure of the HIV "capsid," a protein shell that protects the virus's genetic material and is a key to its virulence. Understanding this structure may hold the key to the development of new and more effective antiretroviral drugs to combat a virus that has killed an estimated 25 million people and infected 34 million more.

NVIDIA updated its GPU micro-architecture roadmap at the 2013 GPU Technology Conference (GTC). Currently spearheaded by the "Kepler" micro-architecture, which drives its GeForce, Quadro, and Tesla product lines, and which will drive Tegra mobile SoCs in 2014; NVIDIA's next-generation "Maxwell" could make its debut some time in 2014. Going by NVIDIA's graph that puts performance-per-Watt against time, "Maxwell" should nearly double performance. Maxwell GPUs feature unified virtual memory, which lets CPUs treat graphics card memory as system memory, for faster general-purpose performance.

Although not backed by a tentative launch year, and assuming "Maxwell" has its run for another two years, 2016 could see the launch of NVIDIA's "Volta" GPU micro-architecture. In addition to advancements by its predecessors, "Volta" could introduce stacked DRAM technology. It would enable GPU memory bandwidths as high as 1 TB/s. Current high-end graphics cards such as GeForce Titan and Radeon HD 7970 GHz Edition, are capable of breaching the 300 GB/s mark, so NVIDIA's claims don't sound far-fetched.

EK Water Blocks, Ljubljana based premium water cooling gear manufacturer, is proud to introduce the EK-FCTK20, world's first water cooling solution for NVIDIA Tesla K20 series professional computing card, the key part of NVIDIA Maximus Technology.

EK Water Blocks now offers complete water cooling solution for both NVIDIA Quadro K5000 series graphics- as well as NVIDIA Tesla K20 series computing cards thus completing the offer for this high-end workstation platform. EK Water Blocks' versatile and fully flexible solution allows for virtually any NVIDIA Maximus configuration to be used and expanded when necessary.

TYAN, an industry-leading server platform design manufacturer and subsidiary of MiTAC International Corp, will present its cutting-edge GPU supporting platform during the NVIDIA GTC 2013. Ranging from 2U to 4U and compatible with1 to 8 GPUs. TYAN offers a full product portfolio to customers which are compatible with NVIDIA Tesla K Series GPU Accelerators. The TYAN S7055 and S7056 motherboard will be showcased as well as TYAN TA77-B7061 and FT48-B7055 barbones at TYAN's booth (#503) during the GTC 2013.

Targeting from mainstream applications to the most complicated high performance computing, TYAN develops a rich product portfolio of GPU based platforms. These products offer flexible and scalable performance as well as maximum efficiency in order to help customers deploying an ideal GPU cluster solutions. All the TYAN GPU platforms are under-validation with industry-leading companies' GPU products. The TYAN TA77-B7061, FT48-B7059 and FT77-B7059 GPU platforms passed the validation and are fully compatible with NVIDIA Tesla K Series GPU Accelerators. The bundled product portfolios of TYAN GPU platforms and NVIDIA Tesla K Series will be widely available in Q2.

TYAN, an industry-leading server platform design manufacturer and subsidiary of MiTAC International Corp, launched the TYAN TA77-B7061, its latest and flexible 2U GPU supported platform; the S7042, entry-level dual socket motherboard; GT62A-B5512 and the GT20A-B7040, the cost-effective 1U server during the CeBIT. TYAN welcomes customers to visit the TYAN website for more detail product information during CeBIT or experience the latest products at ISS EMEA in Dublin, Ireland from April, 10-12.

TYAN TA77-B7061 is the ideal platform designed for users who are looking for accelerated data-processing and efficient computing performance. Supporting up to four GPUs in a 2U server chassis, it is a high performance option that also offers outstanding computing density, saving both time and server/rack space in conjunction with industry-leading companies' GPU products such as Intel Xeon Phi processors, NVIDIA Tesla K20 Series and ATI FirePro. The non-compromised TA77-B7061 will be displayed at TYAN booth at ISS EMEA (Intel Solution Summit) in Dublin, Ireland from April, 10-12. TYAN TA77-B7061 supports (2) Intel Xeon E5-2600 Series Processors, (8+8) DDR-III R/U/LR-DIMM, (4) PCI-E x16 G3 slots, (1) PCI-E x8 G3 slots, (2) GbE ports and (8) 2.5" HDDs. Customers could enjoy more flexible and upgradeable computing performance to tailor to their own computing environment with TYAN's TA77-B7061 server platform.

NVIDIA (NASDAQ: NVDA) today reported revenue for fiscal 2013 ended Jan. 27, 2013, of a record $4.28 billion, up 7.1 percent from $4.00 billion in fiscal 2012. GAAP earnings per share for the year were $0.90 per diluted share, a decrease of 4.3 percent from $0.94 in fiscal 2012. Non-GAAP earnings per diluted share were $1.17, down 1.7 percent from $1.19 in fiscal 2012.

During the quarter, NVIDIA repurchased $100.0 million of stock and paid a dividend of $0.075 per share, equivalent to $46.9 million. "This year we did the best work in our company's history," said Jen-Hsun Huang, president and chief executive officer of NVIDIA. "We achieved record revenues, margins and cash, despite significant market headwinds.