GeForce GTX 465

  The video card is equipped with 352 CUDA cores - a little more than two-thirds of the computing resources of the crystal. Already by these characteristics, you can guess what the performance of the GeForce GTX 465 will be. But the CUDA cores are not the only resources that have been cut from the new product. The GeForce GTX 465 also provides 44 texture units and four ROPs capable of thirty-two 32-bit integer pixels per clock. The memory interface has also been cut down, from 320-bit for the GeForce GTX 470 to 256-bit, that is, the video card is equipped with 1 GB of GDDR5 memory. 

The claimed TDP is 200W - 15W less than the GeForce GTX 470 and 50W less than the GeForce GTX 480. In our original review of the GTX 470 and 480, we already noted a significant disparity between the declared power consumption levels of the graphics cards based on Nvidia GF100 and competitors from AMD. Indeed, the two companies seem to define power consumption levels differently, much like AMD and Intel CPU ratings are not comparable. As far as Nvidia is concerned, the thermal output (TDP) is the maximum power draw in real-world applications, and does not match the maximum power draw in benchmarks such as FurMark. This helps explain why the GeForce GTX 480 draws 133W more than the GTX 465 under load in our tests, even though the official TDP difference is 50W.

The GeForce GTX 465 uses three GPC clusters - that is, one cluster is completely disabled. Of the 12 remaining SMs (recall that there are four SMs in one cluster), Nvidia has disabled one more, leaving 11 SMs, each with 32 CUDA compute cores. Actually, from here we get the number of cores 352.

Also remember that in the Fermi architecture texture units are tied to streaming multiprocessors (SMs). Four texture units are allocated for each multiprocessor. You can do the math yourself - 11 SMs of four blocks each result in 44 texture units. Moreover, the geometric performance also sags, since there are now 11 PolyMorph engines in the GPU.

Nvidia decided not to change the frequency of the GPU and shader unit compared to the GeForce GTX 470 - they still run at 607 and 1215 MHz, respectively. But the memory frequencies have slightly decreased - from 837 to 802 MHz. Of course, all these cuts have led to a change in performance.

The video card proved to be very inefficient in terms of power consumption and too slow in terms of performance. It appeared as a temporary player in the graphics market and quickly gave way to the GeForce GTX 460.