NVIDIA formalizes the GeForce GT 640, the Kepler not expensive

The California skier, NVIDIA, launches its GeForce GT 640, entered graphics card based on the Kepler architecture.

NVIDIA took the Computex currently taking place in Taipei to formalize its graphics card of entry based on the new architecture Kepler, the GeForce GT 640.

However, unlike its counterparts that are the GTX 670, 680 and 690 targeting the high and very high-quality, GK104 on the horizon. The small last hosts actually its PCB a GK107, declination of Kepler to the die of only 118 mm² for 1.3 billion transistors always engraved 28 nm (294 mm² and 3.5 billion transistors for the GK104).

Thus, the GK107 consists of 2 SMX 384 CUDA cores, 32 units of texturing and 16 ROPs. The GPU is running at 900 Mhz and does not benefit from GPU Boost technology, the 1 or 2 GB of DDR3 memory see frequency sets to 891 Mhz and are sent via a 128 bit for a bandwidth bus up to then the 28.5 GB/s. The map takes a format low - profile which monopolisera that a slot and its 65 Watts TDP allows him to escape to power connector, the energy provided by the PCI-Express port, more card is efficient in standby mode it consumes more than 15 Watts.

The outputs GT 640 offer three GeForce video with two Dual-Link DVI ports and a mini-HDMI 1 4A for a fee of $ 99. Many third-party manufacturers and partners of NVIDIA will also offer their own models of references and/or specific with many variants which we reparleront you soon.

Compare prices of Gainward GeForce GT 640 2GoZotac GeForce GT 640 2GoeVGA GeForce GT 640 2 GB Single SloteVGA GeForce GT 640 2 GB Dual Slot

Read More

NVIDIA formalizes the GeForce GT 640, the Kepler not expensive

The California skier, NVIDIA, launches its GeForce GT 640, entered graphics card based on the Kepler architecture.

NVIDIA took the Computex currently taking place in Taipei to formalize its graphics card of entry based on the new architecture Kepler, the GeForce GT 640.

However, unlike its counterparts that are the GTX 670, 680 and 690 targeting the high and very high-quality, GK104 on the horizon. The small last hosts actually its PCB a GK107, declination of Kepler to the die of only 118 mm² for 1.3 billion transistors always engraved 28 nm (294 mm² and 3.5 billion transistors for the GK104).

Thus, the GK107 consists of 2 SMX 384 CUDA cores, 32 units of texturing and 16 ROPs. The GPU is running at 900 Mhz and does not benefit from GPU Boost technology, the 1 or 2 GB of DDR3 memory see frequency sets to 891 Mhz and are sent via a 128 bit for a bandwidth bus up to then the 28.5 GB/s. The map takes a format low - profile which monopolisera that a slot and its 65 Watts TDP allows him to escape to power connector, the energy provided by the PCI-Express port, more card is efficient in standby mode it consumes more than 15 Watts.

The outputs GT 640 offer three GeForce video with two Dual-Link DVI ports and a mini-HDMI 1 4A for a fee of $ 99. Many third-party manufacturers and partners of NVIDIA will also offer their own models of references and/or specific with many variants which we reparleront you soon.

Compare prices of Gainward GeForce GT 640 2GoZotac GeForce GT 640 2GoeVGA GeForce GT 640 2 GB Single SloteVGA GeForce GT 640 2 GB Dual Slot

Read More

NVIDIA: Kepler, it pay?

Small update on NVIDIA finance, in place in its various markets and decided to accelerate its growth with Kepler.

NVIDIA comes from post its financial results for the first quarter of its fiscal year 2013 (i.e. the second calendar quarter ended April 30) and has over these three months a turnover of $ 924,9 million. Net income is down to $ 60.4 million, this being explained by all the investments made now by the company for the market.

Jen-Hsun Huang, President and CEO, explains without detour: "Kepler GPUs accelerate business (...)" Our new range Dekstop is having one of the best notes that we have had in the press. On the Notebook, our GPUs are a record quarter. And Tegra, the growth curve is good, boosted by very good terminals and waiting to see what will give Windows on ARM (...) The graphics are more important than ever. "Stay connected, the GPU Technology Conference, we will introduce new mobile and cloud-computing".

Recent news from NVIDIA includes launches the Kepler GPU GeForce GTX 670, 680 and 690, as well as those of some mobile versions in series GeForce 600 M. The first ultrabook equipped with a NVIDIA GPUS was released March 6, it's the Acer Aspire Timeline M3. The first Tegra 3 phone is available, the One x NVIDIA HTC also announced the launch next GeForce Experience service who, through powerful servers, will be able to optimize the graphics settings of the games based on the configuration of the player.

For the second quarter of its fiscal year, NVIDIA is expected to achieve a turnover between 990 million and $ 1.05 billion, with a margin of about 51.2%. It can be expected that after these launches, this either on the coming months that the range of GPUs Kepler gives its first real results.

Read More

NVIDIA: the CEO sends a mail of congratulations to Kepler

A congratulations email that ended three years of development for this new architecture Kepler...

The manufacture of a microprocessor or a new architecture for graphics cards is not simple. This takes time. Design, development, testing and validation to finish production itself can take several years, between three to five years depending on the complexity.

A view in the long term with intense competition greatly increases the risk of failure. Even with this amount of work accomplished, it may happen that in the end, it does not terminate on a success. The examples are numerous in CPU and GPU over these last ten years.

With the arrival of the Kepler architecture designed by NVIDIA, as at 28 nm lithography, the firm to the Chameleon arrives, with the sale of the GeForce GTX 680 that we spent on our test bench at the end of a long cycle of development. And for the CEO Jen-Hsun Huang, it is a success, so he held to congratulate all employees for this success in an e-mail. The following are translated below into French...

-Original Message-

From: Jensen H Huang
Envoy: Thursday March 22, 2012 9: 48 AM
A: employees
Topic: Kepler rising

Today, the first - GTX 680 - Kepler is available everywhere in the world!

Three years for manufacturing. The effort of thousands of engineers. A single vision - making a revolutionary GPU and make a giant leap in terms of performance and efficiency.

Achieving this level of performance and efficiency, while consuming less energy, required to fully review our approach to design. A very close collaboration between architectures, design, VLSI, software, developers and systems teams, intense audits energy losses, and finds at each level where this was necessary were necessary. The results are fantastic as you'll see in the tests.

Kepler has also cultivated a passion for craft - nothing is lost, everything has been implemented with care - with aims to create an exquisite product that works well. Continue raising the bar and establish extraordinary skills as a mark for our company.

Today is only the beginning of Kepler. For its energy-efficient architecture, we will extend the GPU to the data-centers, bezels portable computers, the superphones. Not to mention the joy and delight of millions of players across the world.

I would like to thank all those who gave their hearts and their souls to create Kepler. You have created something wonderful.

Congratulations to all

Jensen

Read More

NVIDIA: small point on the Roadmap of Kepler

Small overview of the known information about Kepler, the future generation of graphics with Nvidia cards.

A roadmap of Nvidia has appeared on the Net, promising a battle between the two competitors from the market of the GPU on 28 nm. To begin, the roadmap of the Chameleon is called GeForce Kepler Roadmap. Kepler, Yes, it is this gentleman who found the Act of the same name, including the law of the orbits of the planets. But we we righteous just. Although great mathematician of his time, Kepler has demonstrated the calculation of areas and other mathematical glitches in the 1600s, hence the name of the roadmap for Nvidia...

First, the caster will focus on the new version of DirectX, the 11.1, normally integrated natively in the next version of Windows. R & D has also worked on etching at 28 nm to avoid problems related to the manufacture of transistors at this tiny size has just be put in place by TSMC.

The first processor Nvidia of the road map will be the GeForce Kepler 107, the GK107 of his name, which appears to be the entry of the family. Unfortunately, there is not the number of hearts stated on this GK107, but we know that he will be entitled to a 128-bit bus on of the GDDR5 and will be based on the standard PCI Express 2.0. In the aftermath, the GK106 will him the standard PCI Express 3.0 with a 256-bit bus always in GDDR5.

Finally, the GK104, one of the most efficient, will be a bus of 384-bit GDDR5 PCI Express 3.0, it goes without saying. For when? Normally in 2012. We should see cards graphics with code GK110 name, combining two chips GK104, to make the SLI directly on a single card. If the burning in TSMC 28 nm process goes well, it is even planned to see maps for 2013 equipped with the GK112 and can rely on a bus of 512-bit GDDR5. A table found on this site tells us a little more, but this is to take in the conditional.

Read More