GeForce G310 Fermi - Nvidia news! 2010

The appearance of a Geforce G 310 card on nVIDIA’s list has initiall surprised everyone, making people think that the company had managed to finish a Geforce Fermi card before the scheduled time, but the surprise lasted little, as the G310 is nothing but a rebrand of the G210.

The G310 is thus a graphic card featuring a 40 nm chip, 16 cores with the frequency of 589Mhz and featuring 512MB of DDR2 RAM that communicate with the GPU through a 64-bit bus. The card has DVI, VGA and DisplayPort outputs, and an HDMI adapter is included.

It’s not very clear why nVIDIA decided to rename a G210 as the first GeForce 300 series, but the true Geforce 3xx with DirectX 11 will arrive next year, probably in January, but more likely to be launched sometime during the first trimester.

The performance and power of nVIDIA’s graphic processors will be used to optimize and speed the Flash PLayer, not only on the desktop GPU’s but also on the compact devices that use the nVIDIA Tegra platform.

The announcement was made during the last few days, on the occasion of the Computex in Taipei: Adobe and nVIDIA have also said that the collaboration is already going inside the Open Screen Project, an Adobe initiative taht thanks to the collaboration of many important companies throughout the world, will bring Flash contents to any kind of devices, from smartphones to computers and TV.

We’ve talked about the new GT300 few posts ago, and the rumors were that the new chip would be completely revolutionary when compared to the present GT200, and nVIDIA has indeed confirmed that.

The GT300 will not be simply a smaller GT200, thanks to the 40nm fabrication process. It will be compatible with DirectX 11, and nVIDIA has declared that the leap will be similar, if not bigger, than the one taken from the G80/G92 to the GT200.

The new chip, whose codename is still unknown, should be ready by the end of 2009 or the beginning of 2010. It is promised to be a better watt/performance ratio and a bigger number of transistors.

The Ion platform from nVIDIA, developed for netbooks and nettops, is valid alternative in a world were Intel seems to rule.

For the first model of its low-cost platform, Intel has relied on an outdated chipset, certainly worth it when it comes to income but not so adapted to the platform’s philosophy: acceptable performnaces with a low energy consumption. And in this market range, the consumption is a major factor.

Nvidia has definitely done better and aimed also towards other features such as the high-defininition content playing and the possibility of playing games, even though with low quality and “simple” games. In order to answer to that, Intel has decided to speed the process and point on the Pine Trail-M platform, which will see an Atom processor in the PineView-M (both single and dual core) with integrated graphic chips and memory controllers.

Would that be the “checkmate” to nVIDIA? Upon a first glance, some would say yes. But the real thing is that nVIDIA isn’t just looking, and it will fight hard. In fact, the company has the intention of presenting the Ion 2, a 40nm platform, expected to be launched by the end of the year.

We’ll have to wait and see how the things go once both platforms are launched.

During the Computex 2009, nVIDIA has presented the first netbook and MID based on the Tegra platform. THey will have different sized screens and Windows CE at first, later to be replaced by Android.

This architecture is different from the nVIDIA ION, as it puts together an ARM processor with a GeForce graphic card, and not an Intel processor and GeForce 9400M. The lack of a x86 CPU indicates that the devices cannot work with Windows XP, Vista or Windows 7, but will have to feature Windows CE. According to nVIDIA, however, the netbook will migrate towards new distributions, and one of them is Google Android.

The first netbook based on the nVIDIA Tegra won’t be distributed by big computer manufacturers, but mobile companies such as Vodafone and T-Mobile. As for the cost, the small devices with WVGA screens will be sold for $99, as for the classic netbooks will have a cost between $100 and $200.

As we’ve mentioned in a previous article, nVIDIA has decided to abandon the dual-PCB design for their next high-level models, those that feature two GPUs working in parallel. After many rumors, here are the images of the new GeForce GTX295 single-PCB.

As it can be seen from the images, thanks to the new single-PCB design, nVIDIA was able to modify and simplify the heatsink. Although the card still takes two slots, the simplifications have lowered manufacturing costs, which can probably result in a lower final price for the customers as well.

Expected to launch by the end of this month, the new GeForce GTX295 single-PCB will keep all the technical specifications of the dual-PCB model: Two GT200 in parallel, 448bit memory interface x 2, 240 stream processors for each core, 1792MB GDDR3 memory and frequencies at 576/1242/1998MHz for core, shader and memory, respectively.

We’ll leave you to the images. Enjoy.

Click on the thumbnail to view the full image

Nvidia has increased the orders for the 40nm graphic chips from their partner company, Semiconductor Manufacturing Company (TSMC) from Taiwan. The news was leaked through internal sources from Graphic Cards manufacturers and accoding to them, Nvidia is expecting an manufacturing increase of 30% by the end of this year.

Nvidia has recently finished projecting the new GPU generation GT300 with 40nm chips manufactured by TSMC. Always according to Taiwan sources, the company would also have shown interest towards the 28nm technology that is being developed by TSMC, whose production should start at the first trimester of 2010.

The card on this review is the proposal from Asus, that follows all the specifications from nVidia’s reference board: hence we can find the 10 memoryh chips for a total of 320MB (32bit/chip), working at a frequency of 800Mhz GDDR3 (1600Mhz effective), combined to a series 8800 GTS GPU working at a frequency of 500Mhz, and 96 stream processors working at 1200Mhz.

As it was mentioned, the G80 chip was the first to use an unified architecture, integrating thus the Stream Processors that calculate Vertex and Pixel shaders according to the necessity.

The GTX 280 features 1024 MB, GDDR3 RAM (512 bit bus) and supports the DirectX 10. The standard frequences are: 602 MHz for the GPU, 1296 MHz for the shaders and 2214 MHz for the memory. The overclocked version takes those frequences to 650 MHz for the GPU, 1400 MHZ for the memory and 2300 Mhz for the shaders. Overclocking the card brings a 5.8% enhancement, in average.

As for general performance, the GeForce GTX 280 is the fastest card we tested so far. Between the GTX 260 and the GTX 280, the differences are only of 8.7%, which doesn’t exactly justify the price difference. The pressure coming the Radeon HD4870 competition has forced nVidia to lower the initial price for the GXT 280.

Even though the GTX 280 has the same lenght of its lower model, the GTX 269, and higher clock frequences, it doesn’t produce more noise. In 3D mode, the temperature reaches the 53 °C, but the fan only produces 37.7 db(A). If the GPU is kept cooled, the fan remains silent.

The GTX 280 substantially lowers its clock frequency in 2D mode, which makes it better than the HD 4850 on this aspect. As soon as it leaves the 3D mode, it goes on its low-performance 2D mode (400 Mhz on the GPU, 800MHz for shaders and 600Mhz for the memory) and the whole system only uses 130 watt. After a few minutes in idle mode, the clock frequences go even lower (still in 2D mode), to 300 MHz for the GPU, 600 MHz for the shaders and 200 MHz for the memory, and the general power usage goes down to 117 watt. On a maximum usage situation, the GTX 280 uses 352 watt. New Nvidia Cards November 2008.

The GeForce GTX 260 features 896 MB, GDDR3 RAM (on a 448 bit bus) and supports the DirectX 10 technology. The default clock speed is of 576 MHz for the GPU, 1242 MHz for the Shaders and 1998 MHz for the memory. On an overclocked card, these frequencies are enhanced up to 620, 1296 and 2160 MHz respectively.

Considering the 3D performance, the GTX 260 is comparable to the AMD Radeon HD 4870. At the resolution of 1280×1024, the GTX 260 is currently better, while at 1680×1050 (without anti-aliasing) the HD 4870 is superior by around 1.4%. With the anti-aliasing, the GTX 260 is 10% faster.

The noise generated from the GTX 260 while idle is reasonable, reaching the 38.1 dB(A). When it comes to the power consumption, the GTX 260 in 2D mode has a definitely lower consumption than the AMD cards. The consumption while idle is of 111 watt, and when in full mode, a system with the GeForce GTX 260 used 336 watt. New Nvidia Cards November 2008.