The Santa Clara, California-based GPU maker has finally reached the point where it can start shipping the GeForce GTX 650 Ti. The board has finally been released and, with it, so has a video where Tom Petersen, NVIDIA's director of technical marketing, answers the most pressing questions about the card.
We're sure that many people are already well aware of the capabilities of the new card, but we'll give a short rundown of the specs anyway, just in case anyone needs the record set straight.
Based on the GK106 GPU, the adapter has 768 CUDA cores, a GPU speed of 928 MHz (not 925 MHz as some thought), 1 GB of GDDR5 VRAM (128-bit interface, 5.4 GHz frequency), a TDP of 110W, one 6-pin PCI Express power connector and three video outputs (dual-DVI and HDMI).
All in all, the device has a memory bandwidth of 86.4 GB/s.
The one thing that is absent is support for the GPU Boost technology, meaning that the clock won't go up on its own.
Then again, this is a $149 /149 Euro product, so it stands to reason that some concessions had to be made, and it's not as though OEMs don't have factory overclocked cards up for order.
In addition to the price, the video below has Tom Petersen explaining why there is no GPU or SLI support on the card (they would have added too much cost) and outlining how the newcomer compares with the GTX 650 and 660.
Overall, the 650 Ti is said to be 40% faster than the GTX 650, while the GTX 660 is described as 50% faster than GTX 650 Ti.
As one might expect, comparisons to AMD cards haven’t been made, but we’ve already seen them starting to crop up. One favorable to AMD's Radeon HD 7850 can be found here. We're still waiting on the other camp to speak their mind.
NVIDIA will accept any other questions placed on Facebook or Twitter and will answer the most relevant in a new FAQ video, two or three days from now.