More than a year and a half is the period during which the 8800 GTX is still what we may call high-end NVIDIA. Now of course, six months after its release, and happy coincidence, just before the arrival of the R600, we saw land a 8800 Ultra with slightly higher frequencies, but nothing Folichon. There are two and a half months, the arrival of the 9800 GTX foreshadowed a substantial increase in performance but in the end the card offered only limited profit on our good old GTX and was behind the Ultra version. If the owners of these cards could rejoice in their investment, there was still time to propose more than a few MHz scraped here and there or to use combinations of two GPUs on one card.

GeForce GTX 260

NVIDIA finally heard us: the GTX 280 offers the first real reworking of the architecture G8x. We know now the modus operandi of the Californian firm: introduce a new architecture on a proven method of engraving. Indeed, given the often high number of transistors the chip is expensive to produce and using the cards are still expensive but it can occupy the land. During the years following NVIDIA derives its architecture on the whole segment of the range by using an etching finer but less optimized for high frequencies. Finally when the new process is controlled NVIDIA actually enjoy premium sound that becomes more affordable. We saw it with the G70/G71, G80/G92 and the history repeats itself: here is the GT200, a real monster of 1.4 billion transistors engraved at 0.65µm.

With its revolutionary unified architecture, the solution NVIDIA Quadro FX 3700 high-end work stations for mobile offers 1GB of memory, excellent image quality and specifications of display optimized for mobile work stations to 17.” Reference standard for OpenGL 2.1, DirectX 10 and Shader Model 4.0, Quadro FX 3700 enables next generation applications for CAD, DCC and visualization to achieve new levels of interactivity.

Adapted to the needs of mobile work stations to high-end
The 1 GB of GDDR3 memory buffer, 256-bit memory interface and 51.2 GB / s memory bandwidth to allow you to visualize and process large data sets and complex models in real time.

Stable, compatible and optimized for professional applications
Certified Solution for the most current computer aided design (CAD) of digital content creation (SCC) and visualization.

Ability of large parallel calculations
With 128 processor cores parallel, the solution allows to perform floating point calculations sophisticated.

Increased battery life
The NVIDIA ® PowerMizer ® 8.0 intelligently manages battery consumption for autonomy and improved performance.

NVIDIA Quadro FX 3700M

characteristics:

NVIDIA ® CUDA ™
This software provides a C language environment and tool suite that deploy new capabilities to solve complex visualization challenges such as Ray (raytracing) and interactive volume rendering.

NVIDIA ® PowerMizer ® 8.0
The eighth generation of the technology management of material consumption, advanced NVIDIA reduces consumption in adaptive and heat generation of portable computers in the graphics subsystem, offering mobile users a full experience, ultra quiet and without compromise .

Precision sub-pixel 12-bit
3 times higher than the precision obtained with the graphics solutions for workstations competitors, precision sub-pixel 12-bit provides greater geometric accuracy, eliminating sparkles, cracks and other imperfections screening.

Shading language graphic Cg high-level
GC (the “C” for graphics) is a programming language for high-level open standard designed for OpenGL that takes advantage of the power of programmable GPUs. The graphics pipeline programmable solution based on the Quadro shading languages high level to enable the creation and integration of photorealistic effects in real time in the drawings, 3D models and scenes.

128-bit precision throughout the graphics pipeline
Can make mathematical calculations with sophisticated high accuracy, ensuring image quality unrivaled. The full precision IEEE floating point and 32 bits per color component (RGBA) generates millions of color variations with the broadest dynamic range possible.

Cutting 3D windows at the hardware level
Cutting areas accelerated in the hardware level (via a mechanism for transferring data between a window and the graphics buffer) improve overall graphics performance by accelerating the transfer between the color buffer and the buffer graph.