Nvidia Tegra K1: new details on energy efficiency consumption

When last March Nvidia started to talk and show the first details of its new Tegra K1 , professionals and above all the users in the process of changing tablet, laptop or smartphone, they began to desire this solution given the excellent performance that this chip should offer (especially in 3D).

What is interesting, however, is even more consumption and energy efficiency, an aspect on which Nvidia worked great.

Tegra K1 will use not only on mobile devices, but also on laptops, convertible, arriving at the server world, where, albeit with lower performance for example a Nvidia Tesla K40 itself, will be able to undoubtedly be more efficient.

Although the comparison may seem risky, the site wccftech.com publishes a slide that sees in front of the absolute performance of Tegra Tesla K40 with K1; the result in terms of raw power is not a surprise and of course in favor of Tesla K40, while when it comes to efficiency, we note that Tegra K1 is much better: 

WCCFTech.com	Tesla K40 + CPU	Nvidia Tegra K21
Single Precision Peak	4.2 TeraFlops	326 GFlops
Single Precision SGEMM	3.8 TeraFlops	290 GFlops
Memory	@ 288GB 12GB / s	2GB@14.9GB / s
Power (CPU + GPU)	~ 385Watt	<11Watts
Performance Per Watt	10SP GFlops Per Watt	26SP GFlops Per Watt

With these numbers Tegra K1 opens the road to become one of the most viable platforms for the coming years, especially given the versatility that should ensure consideration of the agreements with the automakers that take advantage of the Tegra solution for future infotainment systems.

Samsung introduced the 8-core smartphone

Samsung at CES 2013 announced a new processor Exynos 5 Octa for smartphones and tablets, which has eight cores. Developing a processor, the company has focused for multi-tasking and graphics performance boost, but tried to reduce power consumption chip.

Nvidia Tegra systems within single chip which is material form like 4plus1. Unlike the latter, where 1 core, low-power switches with 4 normal, when their performance is not claimed, the processor Samsung combined 4 high on the architecture of the core Cortex-A15 and four energy-efficient architecture for core Cortex-A7. All are made on a common substrate.

The first group of cores used for processor-intensive tasks, and the second - for tasks that do not require high processing power. Exynos 5 Octa made ​​on 28-nanometer process technology, as well as a new generation of Nvidia chip - Tegra 4 .

Samsung sets the processor Exynos, based on the architecture of ARM, in its smartphones. In addition, she produces for communicators memory and displays, as well as other components. According to the head of the unit, which supplies large-scale integrated circuits, in this sense, the company is unique.

According to estimates of Samsung, the chip, the development of which is Exynos 5 Octa, is already used in 53 million devices. Processors on the market for smart phones Samsung is the second largest in the world with a share of 27%. Qualcomm is in the first with a 42% market share.

NVIDIA Project Shield: 5-inch gaming console to Android

Apart from the world's quickest mobile processor Tegra 4 NVIDIA presented the public a new product in the field of gaming entertainment - mobile console called Project shield, which goes on sale in the second quarter.

Game console is based on the new NVIDIA chip with a clock speed of 1.9 GHz and running the operating system Android. The device has display with a diagonal of 5 inches, the resolution is 1280x720 pixels (294ppi). Of interfaces - HDMI connector to connect the device to an external monitor or TV, USB-port, audio jack and a memory card slot micro SD. The battery provides an autonomy of games from 5 to 10 hours and video playback in high resolution - up to 24 hours.

According to the manufacturer, the quality of built-in speakers Project Shield is on par with the top laptops, and from the communications capabilities is support for Wi-Fi and Miracast. Available standard analog sticks and all controls, familiar to players, including the side buttons.

As stated in NVIDIA, the console is created in order to "really enjoy" games for Android and Windows. If the first OS is more or less clear, from the Windows-based applications, things are slightly more complicated. It is assumed that the games for Windows, you can run only through the Big Picture Mode for Steam online streaming. Native games will be available from the online store TegraZone.

It is known that the system will be pre-installed mobile customers service Hulu, Netflix and Slacker Radio to view audio and video content over the Internet. Cost NVIDIA Project Shield has not been announced.

Nvidia has announced the world's fastest processor for smartphones and tablets

Nvidia announced a new generation of mobile processors - Tegra 4. The new chip is 6 times the speed of visual computing its predecessor, Tegra 3, and is "the fastest mobile processor "today.

Chipset Tegra 4, codenamed Wayne has 72 cores and Nvidia GeForce four basic and one additional cores. According to the manufacturer, the processor for smartphones and tablets faster than any current mobile chips on the market, and up to six times faster than Tegra 3 speed visual computing.

Processor, manufactured on 28-nm process technology, is based on the architecture of ARM Cortex-A15, which has also received a number of improvements, including nearly three-fold increased speed of surfing the Internet. Optional Nvidia chipset allows you to provide your mobile device supporting network LTE.

Another feature of the Tegra 4 has enhanced opportunities for photographers. The new processor is capable of doing HDR-pictures (images with high dynamic range), which will provide them with a better quality.

In addition, Nvidia claims that Tegra 4 consumes 45% less energy compared to previous generation chip, under standard operating scenarios. This is partly achieved through the Prism 3 technology, which reduces power consumption of the display illumination while maintaining high visual quality.

Saves battery power helps an additional fifth core - a decision the same processor Tegra 3, which also has five cores. At low load, such as when the user starts the application for social networks or mail checks, basic core idle, and tasks are passed on to the extra core with less power consumption.

According to the specs, Tegra 4 can provide up to 14 hours of video playback in high resolution on a smartphone.

Nvidia embraced Open Source (Android Tegra 3) and so it was that?

It was June when Linus Torvalds say that Nvidia was the worst company we have worked with. From these statements are only a few months and the Tegra 3 has not only been chosen for the Nexus 7, device peak in the Tablet of Google, but the same company graphics cards and Mobile SoC seems ready to enter actively into the world of Open Source. Nvidia has announced intentions to release the documentation that should allow developers to work with the Tegra 3 platform, consisting of a core ARM Multicore and graphics subsystems owners.

Tegra 3 being very popular on Android, this statement puts in turmoil community developers. It really is not all that glitters is gold and when the intention is to only release the documentation for the 2D graphics engine, while the 3D graphics have to wait again.

Although it is still early to say that the code of the Tegra 3 will be Open, the next steps of the manufacturer will be critical to understand in which direction it will move Nvidia considering the arrival of Windows 8 Tablet and with the same processor. In short, many new and further developments may be expected for the present and future devices are likely even if we still wait.

Note: obviously the news is related to Android and Tegra 3 is not generally to Nvidia and Open Source.

Quadro K5000: the most powerful graphics from NVIDIA for the Mac

To meet the most sophisticated needs of users 'Macs', using computers for their professional activities, NVIDIA suggested graphics accelerator Quadro K5000. This is the most powerful solution in its class to serve on the desktop Apple.

Video card, made the architecture Kepler, performance twice as product Quadro 4000 for Mac generation Fermi. Support Cinema 4K technology provides the ability to process and view video resolution of 4096 x 2160 pixels.

Quadro K5000 for Mac has 1536 stream processors and comes with 4 GB of GDDR5 memory with 256-bit bus, providing unsurpassed performance in video editing software, visualization, and three-dimensional rendering. A powerful processor allows you to work simultaneously with four displays, while the maximum power consumption of the card is 122 Tues

To connect the display units provided by two DVI and DisplayPort. Also, the support OpenGL, OpenCL, CUDA, to install two Quadro K5000 in one Mac Pro.

Quadro K5000 for Mac will be available before the end of the year at a price of $ 2250.

Galaxy GeForce GTX 670 hours with 4 GB of memory GDDR5

Galaxy has released, currently only for the Chinese market, a new variant of the GeForce GTX 670 equipped with 4 GB of GDDR5 video memory , a product that is in addition to the three models already in the catalog. The Galaxy GTX 670 4GB is proposed with a heat sink custom dual-fan, characterized by a massive aluminum body crossed by four copper heatpipes.

 

The GPU of this card operates at 1008 MHz (1084 MHz in Boost mode) while the 4 GB of GDDR5 memory clock is identical to the reference Nvidia 6008 MHz with a 320 bit bus. The power supply is standard with two connectors 6 +8 pin PCI-E, the TPD equal to 170W. Onboard will also find two connectors for SLI and DVI video output (x2), HDMI, Display Port. The price is not its content, and is around $ 520.

by

EVGA Precision X 3.0.3 Download Free for Windows

Convenient tool for managing video cards NVIDIA, the only one which supports Frame Rate Target (dynamic overclocking GeForce or decrease the frequency of 600 to achieve the desired quantities FPS).

List of changes EVGA Precision X 3.0.3 overlaps with MSI Afterburner 2.2.2 :

• Increased the limit of power for some extreme versions of the GeForce 600.
• In the separated graphs can now specify the minimum and maximum values.
• Corrected LED fan speed on systems with passive cooling.
• Updated to version 4.3.4 OSD-module (On-Screen Display):
• Fixed calculation of frame rates in some applications, DirectX 10/11, including the «Max Payne 3."
• Fixed bugs:
• incorrect frequency on GeForce GT 640;
• incorrect frequency when switching to a GeForce 600 video card prior to «Kepler» generation;
• incorrect display of frequency and voltage after the change of skin. 

EVGA Precision X – A utility for overclocking and tweaking NVIDIA video cards with a simple, convenient and colorful interface. RivaTuner is based on a video card with support from the manufacturer EVGA (works with any video cards from GeForce 6000 to 600 except for mobile). Allows you to control the frequencies and voltages manually set the fan chart, depending on the temperature, up to 10 profiles with switching on the hot keys. Knows how to save screenshots of the games (BMP, PNG, JPG with a given quality). Keeps track of monitoring data (temperature and frequency, fan speed, load the GPU and video memory), can output indicators in its box and in the tray, the LCD screen keyboards Logitech, as well as the top of the screen in games (including the value of FPS). Download EVGA Precision 3.0.3 X (Windows, Freeware)

Mike Rayfield: "We will build in other tablets like the Nexus 7"

How has Google in partnership with Asus to produce a tablet on the market and to launch Tegra 3 for only $ 199? According to Mike Rayfield , Chief Engineer Nvidia Tegra 3, the greatest effort that brought down the price covered the cost of components. Nvidia Kai , unveiled a few months ago, using memories DDR3L relatively less expensive than other solutions and if We combine all of a less complex chip Tegra 3 classic, here is that the price is lowered.

 

Nvidia is trying to tighten acccordi business with several partners to bring to market low-cost solutions such as the Nexus 7 and try to further expand this market in an attempt to take off this area and create a greater competition, which in turn will lead to a degradation of the component prices and then a final cost that will further lower.

Rayfield has confirmed that a version of ' Acer Iconia Tab A110 will arrive on the market with chipsets Nvidia Kai , as well as a kid-friendly tablet called Nabi 2.  Not all devices can take advantage of Nvidia but Kai proprietary technologies. Nvidia refers to the fact that Nexus 7 is powered by Nvidia Prism to extend battery life by reducing the screen brightness, compensating for the chromatic and easing the transition of data from the touchscreen to the processor , but this was only possible using memories DDR3L then in the presence of diverse hardware could also change the optimization software.

According to Mike in the near future will see more and more solutions of this type identified by a price particularly aggressive but able to guarantee optimal performance.  by

Preview: Jelly Bean Kernel 3.1, Tegra 3 thanks!

Great news for all users and especially the owners of Android Tegra devices. We can officially confirm and preview, Jelly Bean, which uses the new 3.1 kernel and this is also used by Nvidia Tegra 3with SoC. So news that opens the doors of hope for many users and developers. Nvidia has finally integrated support for the new kernel and not only is the new 3.1.

The image is compelling and clearly shows the suspects had during the presentation. All that remains is to await the movements of the community hoping to soon see major developments for device Tegra 3, Tegra 2 and hopefully in the coming months. We leave you with a new video of Jelly Bean:

GeForce GTX 680 was overclocked to 1442 MHz, using air-cooled

Overclocking enthusiast k | ngp | n explore the potential of increasing the clock frequency on GeForce GTX 680 sample 3D-card EVGA GTX 680 SC. With the use of "extreme cooling" graphics processor was overclocked to 1957 MHz.

More interesting is the result of dispersal from the air cooling system, accessible to more overclockers.

In this case, k | ngp | n and the EVGA GTX 680 SC was subjected to 1442 MHz mark. Recall from the reference sample base clock frequency is 1058 MHz, high - 1124 MHz.

 

As can be seen in the illustration, with the GPU voltage had to increase to 1.212, which corresponds to the upper limit stipulated by NVIDIA. Memory clock was 1812 MHz.

The system used to disperse, was part of the processor Intel Core i7-3960X, overclocked to about 5 GHz, and the memory running at 1245 MHz (effective 2490 MHz). In 3DMark 11 the result was recorded P12745.

With the NVIDIA GeForce Experience will be set optimum settings in games

The event GeForce LAN Shanghai NVIDIA presented its new development. This development does not apply to single-chip graphics processors, or systems for mobile devices - categories in recent years are the most active generators News NVIDIA with the word in the title. Although a definite relation to the first of these categories has a new product. It is called the GeForce Experience is a technology and optimal tuning graphics performance PCs for gaming, implemented by means of the "cloud" computing.

 

The principle of the GeForce Experience seems pretty simple. The system sends to the server details of your configuration, and in return receives the optimum settings. The interesting part is that, allegedly involved in the elaboration of the optimal settings supercomputers that can take into account all the nuances of possible combinations of GPU, CPU, motherboard and device drivers.

The need for such an instrument, according to NVIDIA, is very high: statistics show that 80% of gamers do not change the settings, not resulting in the highest quality graphics, which are capable of their PCs.

The work of the service in a test mode to start in June.

We present 3D-card NVIDIA GeForce GTX 690, the expected price - $ 999

As expected , in the event GEFORCE LAN / NVIDIA Game Festival 2012 in Shanghai, which brought together about 6,000 Chinese gamers, NVIDIA introduced its new product. According to the manufacturer, GeForce GTX 690 is the "world's fastest consumer graphics card" .

 

Among the special features of the skeleton of the manufacturer notes, "covered with aluminum trivalent chromium" and the fan shroud cast magnesium alloy.

The design of the cooling system includes two evaporation chamber and plated radiator with a fan in the center.

 

At the heart of the video card - two 28-nanometer GPU, built on the architecture of Kepler. In the configuration of the GTX 690 is a total of 3072 nuclei CUDA. The base frequency is 915 MHz, higher - 1019 MHz. The total volume of GDDR5 memory is 4 GB. Each GPU is connected with his half of 256-bit bus.

Supply Chain products, collected on a 10-layer PCB, based on a 10-phase scheme. The additional power is supplied by two Octal sockets. The declared value of TDP - 300W.

 

On the mounting bracket slots are DL-DVI (three pieces) and the mini-DisplayPort 1.2.

The card supports PCI Express 3.0 and takes two in the PC expansion slot.

 

In limited quantities NVIDIA GeForce GTX 690 will be available starting on May 3, and more widely - from 7 May. According to NVIDIA, GeForce GTX 690 will deliver the company's ASUS, EVGA, Gainward, Galaxy, Gigabyte, Inno3D, MSI, Palit and Zotac. The expected price of 3D-card - $ 999.

Dark Meadow: The Pact apk v1.1.20 Download Free for Android

With at least a couple of months of delay comes on the Google Store Play "Dark Meadow: The Pact ", a survival horror first-person view with the support of the ' Unreal Engine 3 , currently available only for devices Tegra 3 .

 

The reason for the delay is easily explained : the software company initially wanted to release the standard version of the game , already present for many months on the IOS platform, device Tegra 2 and Tegra 3, the game had to be sold just like on the Apple platform. Then the developers have created a sort of expansion for the title , changing just the name of the game by adding " The Pact ".

Not to be left behind Android have therefore decided to also apply this expansion to version dedicated to our own devices , taking also the mode free for sale on the Play Store.

But let's get to the point: as we said earlier Dark Meadow is a survival horror first-person view, with great graphics support the Unreal 3 engine.

The story sees our hero waking up in a typical abandoned hospital , an elderly man who begins to dispense advice warns us against the dangers that lurk in the building of course. Without revealing anything more we will tell you only that the controls will be based on gesture on the screen .

A dark tale of intrigue and mystery, where you pay for the sins of your past ... Here's the scenario develops in which The Dark Meadow: a world made of visually stunning exploration in depth, and a story that captures the heart-pounding combat, created with the unique power of Unreal Engine 3.

You wake up in a hospital without knowing how you got there and why. An old man warns you of a beautiful witch whose minions roam the corridors looking for living beings. Only when the witch dies, you will free from this endless torment.

With intuitive controls and fun, completely based on gestures, explore the crumbling ruins of the Montclair Hospital, armed with a crossbow for ranged fighting and sword to those bodies and body.

The best elements of role-playing games serve as a stimulus for exploration and a wide range of items are hidden throughout the world, from paper cutouts that reveal the hidden mystery of your weapons and items to catch up to the money can be used to purchase weapons and better equipment.

The beauty of the game weighs 538 MB , with no need (and we needed more) of additional data after the installation. At the time the Play Store   returns only compatible with the ASUS Prime Transformer  (or other devices with Tegra 3) but not necessarily in the next few hours will not be extended to our support for Tegra 2 . We do not know with certainty whether the compatibility will be further extended to devices not Tegra .

As we said Dark Meadow Bride free policy and is available free of charge. 

Name: Dark Meadow: The Pact Download from PlayStore

Qualcomm and Nvidia at loggerheads with TMSC, next to 28Nm chips entrusted to Samsung?

Two of the largest U.S. chipmaker, Qualcomm and Nvidia , are beginning to reckon with a poor yield of the production 28Nm carried out by TMSC , Taiwanese semiconductor company to which they refer many organizations in the field. Already last month came forward signals determined by Nvidia CEO Jen-Hsun Huang, who said in Forbes hope that establishments Intel could "accommodate" the production of the next (SoC Tegra 4) and Kepler GPU to 28Nm. Vain hope given the response of the lapidary home in Santa Clara:

Thanks, but no thanks

TMSC has therefore serious problems in the manufacturing shift to a more sophisticated architecture and is the right partner with which to face future challenges in the mobile industry. For this reason the search started for an alternative that may provide greater safety and the names circulating are the United Microelectronics, Global Foundries and even Samsung , which relies on the structures of Austin, Texas, in the further modernization and expansion.

Be noted that Nvidia stands to lose more with Tegra SoC and 4 GPUs Kepler just waiting to start production for a market introduction in early next year as announced on the roadmap . A delay would be really catastrophic and would leave the field open to emerging reality with better means as Intel , colossus that may rely on its facilities and is already producing in large volumes to 22Nm. No small advantage.

TMSC, meanwhile, is investing huge sums to adapt to the situation and, if it has to deal with the discontent of Nvidia and Qualcomm, can also enjoy the positive approach of recent Apple and AMD obviously aim at different targets. The chipset market will be even more busy in the months and Intel, which is emerging only now seriously in the mobile sector, could be a real thorn in the side.

Tegra 4 specification first exposure: the A15 architecture is 4 + core

The NVIDIA Tegra processor has just become very popular in the tablet machine and the smart phone , the VR-Zone today received some specifications about the next generation of the Tegra. The Tegra 4 codename Wayne (Batman Wayne) on the official road map is scheduled for 2012, but actually to 2013 will be launched, and most models have to wait until after a year and a half.

NVIDIA Tegra4

According to the latest intelligence, the Tegra will have four different models, but only three of which will be based on the latest ARM Cortex-A15 processor architecture, the other a will continue to follow the Tegra A9 architecture, while the core configuration or "4 - the PLUS-1 ", that is, the four main core plus a secondary core.

The first quarter of 2013, starting flagship model clocked at 1.8GHz, high-end positioning, for the 10-inch flat-panel machine equipment to the third quarter will be upgraded to 2.0GHz.

Is the same in the third quarter of next year, the two mainstream models will be available, one of the A15 architecture, clocked at 1.2-1.8GHz, still for the 10-inch flat-panel machine, market positioning will be more mainstream and even can bring cheap products The second is the A9 architecture, clocked at 1.2-2.0GHz, 7 inch tablet and smartphone-oriented, and in particular, by adding support on LTE100/HSPA42 network.

This will be the first Tegra to support LTE networks, can make up for the Tegra 3 is unable to support LTE, leading to a large number of 4G mobile phone to switch to the shortcomings of the Qualcomm platform.

As to the exact specifications of the GeForce GPU graphics core integrated in the Tegra 4, does not yet have a more accurate statement.

Obviously, until the second half of 2013 is the Tegra 4 real heat and light for the beginning of the year and a half before that will still see the Tegra 3, so come back to look at the Tegra 3 of the six models of frequency specifications, put into operation in progress.

TSMC is having difficulty with the release of 28-nanometer production

According to industry sources, the world's largest contract manufacturer of semiconductor products TSMC is having difficulty meeting demand for 28-nanometer products. The main customers are companies TSMC Qualcomm, AMD and NVIDIA. It is believed that the situation will improve in the third quarter of this year.

However, Qualcomm agreed to transfer part of their orders for ARM processors for the company United Microelectronics (UMC), because the volume of products manufactured by TSMC, is not enough for the rapid market of smartphones and tablets.

AMD has introduced the first graphics accelerator based on a 28-nanometer GPU at the end of last year, but shipments family of video cards Radeon HD 7000 is still quite small.

NVIDIA has managed to release two cards based on 28-nanometer GPU Kepler: GeForce GTX 680 desktop PCs and GeForce GT 640M mobile, but, apparently, will also be forced to postpone plans.

For this reason, TSMC is considering the expansion of production . Construction of a new line in the factory Fab 14, which works with 12-inch (300 mm) plates, will begin April 9. Total for the year 2012 the company plans to increase production by 10%.

There are new details about the 3D-card NVIDIA GeForce GTX 670 Ti, including price

The information that NVIDIA is going to release 3D-card GeForce GTX 670 Ti , which will serve as the basis for the graphics processor GK104, appeared in February. This week, a source has published some details about this model.

The number of clusters of graphics processing (GPC) will be equal to four, but in contrast to the GeForce GTX 680, one of eight streaming multiprocessors (SMX) GK104 will be blocked, so that the total number of CUDA cores decrease from 1536 to 1344. In addition, with 128 to 112 reduced the number of texture units, so the number of ROP units will be equal to 32. Do not suffer, and the memory bus - its width will be equal to 256 charges.

As for clock speeds, allegedly, the kernel will run at a frequency of "900 MHz", and memory - "1250 MHz". Will the GeForce GTX 670 Ti used in the dynamic increase in the frequency and, if so, that which is equal to the increased value - is still unknown.

Tentatively, the new 3D-card will cost $ 349-399, but the market is due in May. Source: TechPowerUp

From a Galaxy GTX 680 with single-slot liquid cooling

The site has published weibo.com images that portray the new solution for the Galaxy Geforce GTX 680. The card is characterized by a heat sink in the liquid that, unlike the reference solution, occupies only a PCI slot.

The company has not released details on the frequencies that will almost certainly higher than the standard version (1006 MHz GPU and 6008 MHz Memory), with regard to other features include 1536 CUDA cores, 2GB of GDDR5 memory bus to 256 GB, PCI interface E-3.0.

We will keep you updated on any news or statements by the manufacturer.

GTX, the 670Ti smoke bomb: to refresh the BIOS turned 680 performance surge

670Ti refreshed version of the BIOS that morph into GTX, 680, in fact , GTX, 670Ti and GTX 680 NVIDIA's named not completely sure GK104 of the two names, graphics core and only one .

This new broke the news and other channels is similar to the GTX, 670Ti just a smoke bomb. The previous test results show that lagged behind the GTX, 670Ti HD 7970 5%, while flashing the BIOS, performance jumped 40% after the replacement of the new driver version . I guess perhaps the 670Ti core frequency only 7xxMHz 680 of the core frequency of more than 1GHz. According to various sources, GTX, 680, the actual power consumption is about 200W, 3DMark 11 X-mode score between 3200-3300.

AMD's low default rate of the Radeon HD 7970, and perhaps will launch overclocked versions GHz Edition, "to close the gap. ALU efficiency of Tahiti as good as Pitcairn and Cape Verde also have to consider the issue.