NVIDIA Introduces GeForce 600M Series

The new series starts off with three separate models, all of which are based upon the GeForce 500M series and thus the Fermi Architecture.

• The NVIDIA GeForce GT 635M which features144 or 96 shader ALUs, otherwise known as NVIDIA "CUDA cores". Despite the higher clock rates on the smaller model, the version with144 Shader ALUs should still be significantly faster since both DDR3 VRAM are on a 192-bit memory interface along with 128-bit connectivity GDDR5 . This card replaces the previous GeForce GT555M and is probably based on the GF116 or GF108 chip, though the older GF106 may be found on some models.

• The NVIDIA GeForce GT 630M also comes in two different versions, featuring the same 144 or 96 shader ALUs. The clock speeds are slightly lower than those of the GT 635M, but the memory interface is 128-bits wide and comes equipped with only DDR3 VRAM. Depending on the model's performance, this card falls in the range of the GeForce GT 540M or GeForce GT 550M. The GF106/116 or the GF108 chips are in use here as well.

• The GeForce 610M is the third and final model and has 48 shader ALUs powered by a 64-bit wide DDR3 memory interface. It's derivative from the 500M series is the GeForce 520MX, which in turn corresponds to a GeForce GT 520M but with slightly higher clock speeds. The 610M is based on the GF119 chip.

For all models, the clock rates quoted are based on maximum values provided by NVIDIA. On the other hand, notebook manufacturers are not bound by these values, and can use lower speeds. 

As for the features of the 40nm Fermi architecture, they are all present including PhysX, CUDA and Optimus; however, 3D Vision Technology appears to have been omitted from the GeForce 610M.

Since the GeForce 500M series was initially released as a rebranding of the previous generation, the same could end up happening with the GeForce 600M series.   Of course, the possibility does exist that new products could be launched along with Intel's new Ivy Bridge CPU generation, which should be arriving sometime in April 2012. The successor to the Fermi line of GPUs is expected to shrink the individual GPUs to 28-nanometer processes, and this could be found on the upcoming Kepler generation.