Archive for the ‘Tech News’ Category

AMD Next Gen HD 8000M Series Details

HD8000M-5

With the New Year quickly approaching, AMD’s HD 7000-series and its GCN architecture is about to celebrate its first birthday. After a year on the market, it’s time for the architecture to undergo a refresh and the notebook segment is going to be the first to benefit with AMD’s upcoming HD 8000M series of mobile graphics processors.

At this point, not much is actually known about these new chips but after an all to quick presentation by AMD today, a clearer picture is beginning to emerge. First and foremost, the initial HD 8000M entrants will use a slightly revised GCN architecture that boasts increased overall efficiency, higher memory clocks and optimized on-die resource distribution. Boost capability will also be incorporated onto some dies instead of only being reserved for higher end desktop parts. The will also implement a “Solar” naming scheme so expect product designations (ie: code makes) based on planetary bodies like Venus and Mars. Additional concrete information isn’t available at this point since AMD’s cards are being played close to their chests but we do know the HD 8000M series will be officially announced on January 7th, right before CES. Notebooks supporting these new discrete graphics will be launched around that time as well (even though some have already been leaked) so expect reviews on Hardware Canucks soon.

HD8000M4

For the time being, AMD is solely focused on providing the mid-level and mainstream gaming segments with refreshed products. In order to accomplish this, the first HD 8000M graphics processors to become available will be the HD 8500M, HD 8600M, HD 8700M and HD 8800M series, each of which will incorporate full DX11.1 support. These refreshed parts are being billed as evolutionary rather than being revolution and these products’ specifications bear this out.

The HD 8500M and HD 8600M are particularly important since they replace HD 7000M GPUs which were based off of AMD’s older Northern Islands architecture (ie: the HD 6000 series) so we could see vastly improved performance per watt in one of today’s most sought-after niches. Another important aspect is the additional capabilities these new cores bring to the table; their 28nm manufacturing process, an upgraded UVD decoding engine, better compute throughput and a number of other aspects are all hallmarks that were sorely missing from this segment.

AMD’s HD 8700M and HD 8800M are actually the real question marks here. With the admittedly small number of differences between the first and second generation GCN-based mobile architectures, finding points of distinction between these new GPUs and Cape Verde-based parts may not be all that clear cut. Nonetheless, AMD assures us their designs have moved forward and end users will see benefits, especially since the new HD 8000M series will bring GDDR5 memory to lower price points, thus improving performance in certain applications.

HD8000M-5

From a high level standpoint AMD’s new product stack is undergoing a broad scale rationalization which could mean cutting down on the number of models within each product category. This in effect could improve AMD’s overall offering by cutting down on consumer confusion and will allow clearer juxtaposition with NVIDIA’s offerings. But will this be successful against NVIDIA lineup that seems to be dominating the notebook market? We certainly hope so.

While the entry level is currently receiving the lion’s share of AMD’s attention, several other cores will be introduced in Q2 2013. This will include a pair of enthusiast-level GPUs which will plunge into the space vacated by the HD 7900M and compete against NVIDIA’s GTX 670M and GTX 680M. Meanwhile, yet another HD 8000M offshoot will be parachuted into a gap between the mainstream and “performance” segments.

For the time being, AMD’s lower-end HD 7400M series will remain as is, and the reason for that should be self-evident. There’s a need to retain strict separation between APU and GPU so neither will step on the other’s toes. So with the Kaveri and Kabini successors to Trinity already in the pipeline, AMD is likely keeping this section free and open in preparation for next generation APUs. With that being said, the HD 8000M series won’t be compatible with Trinity’s Dual Graphics technology which leverages a discrete GPU alongside the APU’s onboard unit for higher in-game performance.

HD8000M-3

AMD’s Enduro technology was originally conceived as a way to compete with NVIDIA’s Optimus GPU switching. With a claimed ability to seamlessly switch between a discrete GPU and the on-die integrated graphics subsystem, it could increase battery life and dynamically control TDP of two hot-running components.

Unfortunately, AMD’s first time at bat didn’t go quite as planned since Enduro went through its fair share of growing pains. The situation has supposedly improved on the HD 8000M series through the use of an improved user interface and revised application detection.

HD8000M-2

HD8000M-1HD8000M-0

With so much weight being put upon the Mobile and Notebook markets, AMD needs to hit this one out of the park from a number of perspectives. The initial performance numbers given to us do show some promise, particularly against NVIDIA’s ultra-popular GT 650M, but there are still some challenges ahead for the HD 8000M. The first area of concern is a lack of information about their partners on this launch. While we have been assured that several products will be available at launch, AMD has to significantly build upon the HD 7000M’s somewhat limited market penetration.

With this bit of initial information, some of you are likely wondering where the rumored desktop parts figure into this equation. At this point, AMD is focused upon refreshing their mobile parts but the desktop HD 8000 series products will follow sometime afterwards, possibly in Q2 2013 so we’re still a ways off from seeing AMD’s entire 2013 GPU product stack. Until then, we’re excited to see what these new GPUs have to offer for notebooks.

Samsung Annoused 14 nm FinFET Logic Process and Design Infrastructure Details

Samsung Electronics Co., Ltd., a global leader in advanced semiconductor solutions, today announced that it reached another milestone in the development of 14-nanometer (nm) FinFET process technology with the successful tape-out of multiple development vehicles in collaboration with its key design and IP partners. In addition, Samsung has signed an agreement with ARM for 14 nm physical IP and libraries. This agreement is the latest in a series from Samsung and ARM that has delivered production proven SoC enablement. Samsung, together with its ecosystem partners, is in a position to offer leading edge customers a robust design infrastructure to drive an ever expanding advanced mobile SoC market.

“As we move closer to true mobile computing, chip designers are eager to take advantage of the gains in performance and significantly lower power of 14 nm FinFET to deliver PC like user experience on a mobile device,” said Dr. Kyu-Myung Choi, senior vice president of System LSI infrastructure design center, Device Solutions Division, Samsung Electronics. “The design complexities at 14 nm require complete harmony between the process technology, design methodology, tools and IPs. We are synchronizing all the key elements so our customers can deliver their newest chips to market quickly and efficiently.”
As part of its 14 nm FinFET development process, Samsung, and its ecosystem partners – ARM, Cadence, Mentor and Synopsys – taped out multiple test chips ranging from a full ARM Cortex-A7 processor implementation to a SRAM-based chip capable of operation near threshold voltage levels as well as an array of analog IP.

The full ARM Cortex-A7 processor test chip tape-out represents a significant milestone for silicon manufacturing for the fabless ecosystem. The Cortex-A7 implementation on FinFET demonstrates the low-power component of the ARM big.LITTLE processor configuration/technology strategy for mobile computing platforms. The Samsung 14 nm FinFET enablement for SoC design provides improved leakage and dynamic power advantages to the expanding mobile computing market. This collaboration builds on the long-standing partnership between Samsung and ARM including SoC design enablement for the production proven 32/28 nm High-K Metal Gate (HKMG) technology. Enabling SoC design on FinFET allows the continued fast pace of innovation that is the hallmark of the mobile market segment.

The Cortex-A7 processor test chip was implemented by Cadence in collaboration with ARM and Samsung. Cadence delivered a full RTL-to-signoff flow, building upon a tool set that has been thoroughly tested on 20 nm designs requiring automated double patterning. The tight collaboration with Samsung and ARM enabled Cadence to hone its technology for 14 nm FinFET designs, paving the way for 14 nm market readiness. ARM used Cadence tools to develop the 14 nm FinFET libraries, and Cadence tools were also used for a full-flow RTL-to-signoff tapeout of the processor core on Samsung’s 14 nm FinFET process, as well as chip-level integration and verification.

Samsung used Synopsys tools optimized for FinFET devices to implement additional IP on this vehicle, including low power SRAMs intended to operate with the power supply close to threshold voltage levels. The move from two-dimensional transistors to three-dimensional transistors introduces several new IP and EDA tool challenges including modeling. The multi-year collaboration between Samsung and Synopsys has delivered foundational modeling technologies for 3D parasitic extraction, circuit simulation and physical design-rule support of FinFET devices.

Samsung is also extending their work with Mentor to enable a complete solution at 14 nm FinFET that addresses customer challenges in design, validation, manufacturing co-optimization, and post-design production ramps. The collaborative efforts leverage the unique capabilities of Samsung’s processes, while helping designers deal with the complexities of multi-patterning lithography, FinFET transistors, and more complex reliability requirements.

Silicon-based PDK Availability

With its process design kit available to customers today, customers can start designing with models, design rule manuals and technology files that have been developed based on silicon results from previous 14 nm FinFET test chips run in Samsung’s R&D facilities. This PDK includes design flows, routers and other design enablement features to support new device structures, local interconnects, and advanced routing rules. The investments that Samsung is making into the entire ecosystem at 14 nm will give customers early access to all elements of the design infrastructure to accelerate their chip development.

Sapphire Annoused 7870 XT With Boost

SAPPHIRE Technology has just announced a new graphics card in its HD 7000 series – the SAPPHIRE HD 7870 XT with Boost. Like the other cards in the HD 7000 series, the new card is based on the highly acclaimed Graphics Core Next architecture from AMD.

The SAPPHIRE HD 7870 XT with Boost delivers a new price : performance point to the series. It is based on AMD’s Tahiti architecture with its 256-bit memory interface, and 1280 stream processors and 80 Texture units, unlike the remainder of the HD 7800 series that uses the Pitcairn architecture. Configured with 2 GB of high speed GDDR5 memory running at 1500 MHz (6 GHz effective) the SAPPPHIRE HD 7870 XT has a core clock of 925 MHz which dynamically rises to 975 MHz with PowerTune Boost, AMDs dynamic performance enhancement for games.

This new model also features SAPPHIRE’s latest Dual-X cooler technology, with its graduated heatpipes and two aerofoil fans with dust repelling bearings for efficient cooling and low noise operation. Output configuration is Dual Link DVI-I (supporting VGA with dongle supplied), HDMI and two mini-DisplayPorts, allowing a wide range of monitor types to be connected. AMD’s multi-monitor SLS technology, Eyefinity is supported.

For enthusiasts wishing to maximise performance of this graphics card, the latest version of the SAPPHIRE overclocking tool, TriXX supports this technology and is available free to download from the SAPPHIRE website. SAPPHIRE TriXX allows tuning of GPU voltage as well as core and memory clocks, whilst continuously displaying temperature. Manual control of fan speed is supported, as well as user created fan profiles and the ability to save up to four different performance settings.

The HD 7000 series is the third generation of DirectX 11 compatible solutions from the graphics division of AMD – formerly ATI. The SAPPHIRE HD 7870 XT is ready for DirectX 11.1 and the new Windows 8 operating system as well as delivering all the features of the current DirectX 11 such as DirectCompute11 and multi threaded communications with the CPU. Communication with the host PC is optimised with the implementation of the latest high speed PCI-Express Gen3.0 interface, and multiple cards can be used to further enhance performance in CrossFireX mode on a suitably specified mainboard.

HDMI specifications have also increased in the HD 7000 series. Fast HDMI 1.4a supports Stereoscopic 3D with enhanced frame rates of 60 Hz per eye – 120 Hz total. It is also ready for the next generation of 4K displays that can be driven from a single high speed input of 3GHz HDMI 1.4a or DisplayPort 1.2 HBR2.

APP Acceleration is the name given to the enhanced performance of a wide range of applications using the stream processing capabilities and specific hardware features of the graphics architecture. The HD 7000 series has been optimised for this type of task. Typical applications include Video encoding and decoding, compression and multiplexing for media distribution, gaming or video conferencing. APP acceleration can also enhance everyday tasks such as Internet browsing, Office applications and file compression with supported software suites.

ASUS : Intel Will Find a Way to Keep Socketed CPUs

ASUS Motherboard Business General Manager  Joe Hsieh commented on reports of Intel abandoning CPU Image is Loading.....sockets in favor of processors being hardwired to motherboards in BGA packages,In an interview with DigiTimes. Hsieh said that the issue will not be as bad as people think and Intel could find a strategy that allows both soldered and socketed processors to be sold, which is much like today, except that hardwired processors are limited to notebooks (Core i3 and i5 processors in the BGA1224 package) and low-end Atom-driven desktop motherboards.

What lends Hsieh’s statement weight, apart from the fact that he leads the biggest PC motherboard design team, is that Intel recently denied those reports, saying it would provide socketed CPUs for “the foreseeable future.” Last month, Japanese publication PC Watch, credited for generally accurate tech predictions based on information at hand, reported that following its 22 nm Core “Haswell” CPU family, Intel could transform its entry-, mainstream-, and performance-segment client CPUs to hardwired BGA packages, probably leaving socketed CPUs only to HEDT (high-end desktop) and enterprise Xeon processor lines. Other PC motherboard vendors DigiTimes spoke with echoed ASUS’ opinion, they don’t believe Intel could “suddenly” completely change the way processors are sold to consumers.

Intel Haswell Based Core i7 M-Series Notebook CPU Lineup Details

Intel Haswell

Intel’s Core i7 “Haswell” line of high-performance mobile processors will launch around roughly the same time as its first desktop counterparts, in April 2013. The April launch will consist of three models, which occupy conventional TDP ratings, making them fit for performance/gaming classes of notebooks, they’re not quite Ultrabook-material.

Leading the pack is the Core i7-4930MX Extreme Edition, which enables nearly every component on the silicon. The quad-core chip features HyperThreading, which enables 8 logical CPUs, 3.00 GHz clock speed with 3.90 GHz maximum Turbo and 3.70 GHz all-core Turbo, new Intel HD 4600 Graphics clocked at 400 MHz with 1350 MHz maximum boost, 8 MB L3 cache, and 57W TDP.

Next up, is the Core i7-4900MQ, another quad-core chip featuring HyperThreading, 2.80 GHz clock speed with 3.80 GHz maximum Turbo and 3.60 GHz all-core Turbo, the same HD 4600 graphics as its bigger sibling, 8 MB L3 cache, and 47W TDP. Lastly, there’s the Core i7-4800MQ, which features just a notch-lower clock speeds than the i7-4900MQ, at 2.70 GHz with 3.70 GHz maximum Turbo and 3.50 GHz all-core Turbo; but lower L3 cache amount, at 6 MB. The TDP is unaffected at 47W.