The NVIDIA GeForce GTX 465 reference is structured to turn around a GF100 GPUs based on a manufacturing process to 40 nm, clocked at 607 MHz in (just like the GTX 470 and 700 MHz against the GTX 480 ) and provided 352 units of computation, also called CUDA cores, clocked at 1215 MHz (against 448 to the same frequency for the GTX 470 and 480 MHz for the 1401 GTX 480). She also embarks 1GB of GDDR5 memory running at 1603 MHz on a bus with a width of 256 bits (cons respectively 1.25 and 1.5 GB, 1674 and 1848 MHz, 320 and 384 bits for the GTX GTX 470 and 480). It finally shows a TDP of 200 Watts (cons 215 and 250 watts for its predecessors).
One finds also support SLI, allowing to use multiple graphics cards in parallel, CUDA, which relieves the CPU of certain tasks in parallel, acceleration hardware PhysX physics engine and 3D Surround Vision, which offers display in relief on several screens. The map is broken down by a cooling system occupying two slots, requires a power supply with two PCI Express offers two dual-link DVI outputs and a mini HDMI output.
Initial tests of several American colleagues show that, depending on videogames and their settings, the performance of the GeForce GTX 465 evolve into those of an ATI Radeon HD 5830 and a 5870.
The NVIDIA GeForce GTX 465 reference is already available to manufacturers, who will announce their own versions in lockstep, for a suggested retail price of 280 dollars (about 270 euros including VAT).
No comments:
Post a Comment