GeForce G310 Fermi - Nvidia news! 2010

The choice of adpoting a memory bus of 512bit has required the use of 16 memory chips per card, and each chip is connected to the memory controller through a 32bit bus. Each of the memory chips used on this card has a 64MB capacity, which results on 1GB memory for this particular card.

The memories use the GDDR3 technology, with a frequence of 2.214 MHz. In this review we’ll go through the main features, the performance and the overclocking capacities that the Gainward graphic card is able to perform. nVidia has implemented its own second-generation video processing engine on its GTX 280 cards, called “Pure Video HD”. The “Pure Video HD” allows for a better high-definition video flux management, especially when compared to the VP1 that was featured on the previous G80 GPUs.

The nVidia Quadro FX 1700 is a medium-range product that offers a good balance between performance and energy consumption. For those who want a graphic card that can handle less complex renderings, the FX 1700 offers sufficient performance for around €400.

Considering the memory capacity, it’s clear to see that this model wasn’t made to manage complex shading and texture operations, but it’s more than enough for simple projects. Despite having 512MB RAM, they’re DDR2 and with a 128bit interface. The maximum bandwidth for this card is, thus, 12.8 GB/sec. It uses the G84 chip with 80nm, present also in the GeForce 8600 cards. Also with this board, customers can find an additional connector for 3D virtual glasses.

Here are the more detailed stats on this board:

The Quadro FX 5600 is considered an OpenGL “monster”. Equipped with 1536MB GDDR3 RAM, the Quadro FX 5600 offers enough power for higher resolutions and bigger textures. The car dis based on the G80 graphic chip, present also in many GeForce 8800 gaming card series. The Quadro FX 5600 is sold for an average price of €2500, and usually gets the first position in many benchmark categories.

The “Maxtreme 11″ driver from nVidia is very interesting. It’s a plugin, developed especially for 3D Studio Max, that offers a significant performance increase with that particular software. In comparison with the previous versions, Maxtreme 11 supports the OpenGL and DirectX API. The 3DSM shader operations also benefit from this plugin.

The 90nm chip and the bigger amount of memory require more energy, which is supplied by two 6-pin Molex connectors.

The GPU-Z’s screenshot shows the main technical information about the FX 5600.

Designed mainly for CAD, DTP and visual simulators, the professional graphic cards include optimizations for professional applications such as AutoCAD, SolidWorks, ProEngineer and similar. Both nVidia and AMD have noticed that the difference between these cards and the classical ones are mainly connected to the drivers, the other features being very similar to their FireGL and Quadro FX (for videogames) models. The high price on these products, therefore, is due to the little amount of units manufactured, and the need to develop specific drivers for them (which includes the certification process and the support).

In order for this market to succeed, producers must make sure that a classic graphic card, such as a GeForce or a Radeon, cannot be transformed into a Quadro FX or a FireGL mode. Many hackers have, indeed, repeteadly created soft-mods or modified BIOS to transform gaming cards into professional workstation cards; they succeded exactly for the fact that the two products are, indeed, 99% identical to each other. That means the manufactures do not have an easy time.

When it comes to the performance, the differences are clear: comparing a gaming card with its (almost identical) workstation version, users will find that the drivers allow the professional version to work a lot faster.

In these next few articles, we will be reviewing the nVidia Quadro FX 5600, 170, 570 and 370.

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.

The biggest motherboard manufacturers are working on having a SLI system on an Intel X58 chipset motherboard. The first ones will be ASUS, EVGA, MSI, Gigabyte and DFI, which, according to nVidia’s announcements, have intentions of using the NVIDIA SLI technology on their own motherboards with the new Intel X58 chipset, designed to work with the Intel Core i7 and Core i7 Extreme CPU’s.

The motherboard will be featuring the NVidia Nforce 200 SLI chipset, which will allow triple-SLI or “normal” SLI configurations with two graphic cards. It doesn’t seem possible for now to think of a quad-SLI. It’s easy to assume that the chipset will allow the SLI configuration normally (avoiding the need to insert the nForce 200 chipset) in a near future, thanks to an agreement with Intel.

NVIDIA just released another set of beta ForceWare drivers (version 180.43), adding a variety of features and enabling NVIDIA SLI on Intel X58-based motherboards. They also offer performance improvements for Ubisoft’s shooter Far Cry 2. The drivers are being released with a quote from Ubisoft’s Louis-Pierre Pharand:

“Since we first began working on Far Cry 2 in early 2005, we have used GeForce GPU’s to develop the game and worked closely with NVIDIA’s excellent engineers. While Far Cry 2 represents an outstanding PC gaming experience for a wide range of systems, GeForce users can expect to enjoy the game quite literally ‘The Way It Was Meant To Be Played’,” says Pharand.

The latest 180.43 drivers are available for Windows XP 32-bit, Windows Vista 32-bit, Windows XP 64-bit, and Windows XP 64 bit.

With the GPU GT200, Nvidia establishes their own second generation of graphic cards, with unified architecture and shading. The GeForce GTX280 and GTX 260, the first ones to use this new chip, establish the new point of reference for the top graphic cards designed for the most demanding video-gamers.

After a few weeks of rumors about the new Nvidia graphic card generation that would use a new GPU architecture to replace the G80 and G92 used on the GeForce 8800 and 9800, the confirmation came today, at the official launching of the new NVIDIA solutions. The new GT200 has in fact doubled the number of stream processors integrated in the new architecture, going from the previous 128 to 240 in the new top version, the GeForce GTX280. The number is lower for the GTX260, going to 192, following NVIDIA’s tradition of creating a cheaper version of their top card, with reduced features.

The GeForce GTX 280 card is visually similar to the previous 9800 GTX solutions: 2-slot building, with a plastic case that covers all the components, in order to allow a better air flow towards the outside. For the GTX260, the design is almost the same: it keeps the same external size and design, the only difference is the power supply connectors: two 6-pin PCI Express for the GTX 260 while the top card, the GTX 280, presents one 6-pin PCI-express connector combined with an 8-pin one.

Besides the increase in the stream processors, we also see an increase in the memory controller, that was taken to 512bit in the GeForce GTX 280. The 512bit memory bus applied on the top card has required a 16 chip memory for each card, while the 448bit of the GTX 260 features a 14 chip memory.

As it was done previously on the G92 architecture cards, NVIDIA has implemented their own second generation Pure Video HD video processing engine, featuring a better high-definition video streaming than the predecessors.