SO IT LOOKS LIKE Nvidia is going to take its Turing architecture and rework it into a whole range of GeForce GTX cards that drop the ray-tracing and DLSS capabilities of their costly RTX siblings.
The so-called GeForce GTX 1650 will be something of an entry-level Nvidia graphics card, akin to the Pascal-based GeForce GTX 1050.
According to the online murmurs, the GTX 1650 will pack some 4GB of GDDR5 or GDDR6 video memory and a 128-bit memory bus, with the number of CUDA cores expected to be slightly down on the 1,280 the RTX 1660 is said to have.
Base clock speed is suspected to be 1,485MHz, while the graphics card’s chip die is thought to be the TU117 architecture, which is different to the TU116 architecture of the GTX 1660 and GTX 1660 Ti.
Given its rumoured specs, the GTX 1650 isn’t likely to be a powerhouse of a graphics card, but it should represent a step up from Nvidia’s previous lower-end graphics.
Such a graphics card would likely find a good home in laptops like the Dell XPS 15, if it’s configured into a mobile form factor, giving such machines some graphical grunt without needing to opt for laptop variants of the RTX GeForce cards, which currently come at quite a premium price.
PC fans looking at building more powerful desktops will probably want to opt for the GTX 1660 cards or push up to the RTX cards with something like the GeForce RTX 2060.
The most interesting part of Nvidia expanding a range cut-back Turing cards is that it can offer some of the gains in performance and power- efficiency to GeForce cards without asking PC builders to fork out for ray-tracing features that aren’t being widely implemented at the moment.
That should allow for Nvidia to keep challenging AMD in the mid-range graphics card arena, at least until Team Red comes out with its Navi GPUs. µ
Source : Inquirer