Nvidia Uses Expensive GA106 for Cheap RTX 3050, But What’s The Point?

Being in keep watch over of a whopping 79% of the desktop discrete graphics playing cards (a lot of which can be some of the easiest graphics playing cards round) has its perks. Amid provide glut, Nvidia can play with silicon portions it makes use of. It is going to have its benefits with the supply of quite reasonably priced parts, and we can check out to give an explanation for why.

Nvidia offered its GeForce RTX 3050 graphics card (the record of forums is right here) at a proper MSRP of $249 previous this week. Many observers idea that the corporate was once the use of its quite reasonably priced and quite low-power GA107 graphics processing unit (GPU) for the answer. Sadly, this isn’t the case. In step with Inno3D, the Inexperienced corporate makes use of a much more dear GA106 silicon for this entry-level product. But this generally is a excellent signal. Let’s attempt to analyze.


(Symbol credit score: Inno3D)

Inno3D despatched a high-resolution render of its GeForce RTX 3050 as a part of a press liberate and not using a cooling device, revealing that the PCB makes use of the GA106 silicon. Sadly, we can not say anything else about different graphics playing cards with the GeForce RTX 3050 badge on them for evident causes. Nonetheless, our humble enjoy with GPU corporations tells us that everybody makes use of the similar silicon at release, however it could alternate through the years. Anyway, Nvidia’s GA106 and GA107 are completely other, so let’s read about those variations.

Comparability of Nvidia’s GPUs

ChipChip Configuration (CUDA)Reminiscence InterfaceDie DimensionMSRP
GeForce RTX 3050 desktopGA1062560128-bit276mm2$249
GeForce RTX 3060 desktopGA1063584192-bit276mm2$329
GeForce RTX 3050 pocket bookGA1072048128-bitSmaller than 276 mm2?
GeForce RTX 3060 Ti pocket bookGA1072560192-bitSmaller than 276 mm2?

The GA106 is a fairly large (272 mm2) chip with as much as 3,584 CUDA cores enabled and an as much as 170W TDP (the desktop model). The cut-down model has 2,560 CUDA cores, which overlaps with the fully-enabled GA107 graphics processor with 2,560 CUDA cores used for laptops and has a TDP vary between 35W and 80W.


(Symbol credit score: Jon Peddie Analysis/Compiled by means of Tom’s {Hardware})

This can be a query for Nvidia, which it’s not prone to reply to publicly. Then again, giant OEMs need one device driving force, and as soon as an organization wins a freelance, it lasts for a cycle, which occurs to be a yr in our fashionable global. So we will speculate that amid profitable the high-performance early 2022 pocket book cycle, Nvidia has causes to have two GPUs that overlap relatively than having no GPUs to send. 

The excellent news for PC players is that Nvidia has two entry-level GPUs in manufacturing. The dangerous information, there are miners, scalpers, and some other bunch of folks that won’t will let you get a GeForce RTX 3050 at its MSRP.

(Visited 1 times, 1 visits today)