NVIDIA has been weaving a nice spell over the past few months, releasing the GeForce GTX 680 and, more recently, the GTX 670, so it is with great bemusement that we report on the company's newest action.To its credit, the company knew it would leave people confused, if not baffled, when it sent out the newest GPU, so it kept things more or less quiet.
Still, this was bound to come out eventually. It just so happens that “eventually” came around faster than anyone had expected this once.
We may as well stop beating around the bush and just say it: NVIDIA has released the GeForce 405M graphics processing unit.
That's right, even though any Advanced Micro Devices APU, Llano or Trinity, can wipe the floor with any entry-level GPUs, NVIDIA released one anyway.
If we were to guess, the Intel GMA HD 4000 might be better than the 405M too, but we haven't conducted any Ivy Bridge tests to see if that is true, so we aren't sure.
GeForce 405M is the weakest GPU in the Santa Clara, California-based company's arsenal, featuring only 16 CUDA cores. To get an idea of what this means, the GTX 670 has 1,344 and GTX 680 has 1,536 CUDA cores.
The shader processor clock is 1,212 MHz, which leads us to believe that the GPU clock is 606 MHz.
Moving on, there are 512 MB of DDR3 available, operating at 800 MHz (1.6 GHz effective) and a 64-bit interface.
As for TMUs and ROPs, NVIDIA hasn't disclosed them, but they are highly likely to be two and four, respectively.
If NVIDIA scores any design wins for the GeForce 405M, it will only be in cheap, low-power systems and consumer electronics. Seeing as how this is the slowest GPU, it is probably the least power-hungry of NVIDIA's products too, so at least there's that.