Manli has unveiled the industry’s first Nvidia GeForce RTX 3060 graphics card with just 8GB of memory. The new board features the same computing performance as his full-blown GeForce RTX 3060 12GB variant, but with reduced memory capacity and bus width, resulting in reduced memory bandwidth. All of these changes should have a definite impact on performance.
Manlis GeForce RTX 3060 8GB GDDR6 graphics card (Via VideoCardz (opens in new tab)) probably uses the same GA106 graphics processing unit as the 12GB card. It comes with 3584 CUDA cores with a boost clock of 1777 MHz, the same clock as the reference RTX 3060 12GB.
Unlike the popular GeForce RTX 3060 12GB board with a 192-bit memory bus with a peak memory bandwidth of 360 GB/s, the GeForce RTX 3060 8GB only features a 128-bit memory interface. Similarly, peak memory is reduced by 33%, down to just 240 GB/s. If memory bandwidth matters, the new board will be significantly slower than the existing GeForce RTX 3060 with 12GB of GDDR6 memory.
Aside from the narrower memory bus, the GeForce RTX 3060 8GB has the same specs as the 12GB card. It comes with a 170W TBP (Total Board Power) rating, a dual-slot cooling system, and four display outputs (three DisplayPort 1.4, one HDMI 2.0).
Neither Manli nor Nvidia have announced an MSRP for the new RTX 3060 8GB model, but it should fall between the GeForce RTX 3050 8GB ($249) and the GeForce RTX 3060 12GB ($329), so it will be priced accordingly. It’s possible. Other graphics card vendors will soon follow Manli’s lead with their own variants of the RTX 3060 8GB.
The driving force behind this card seems to be a desire to offer something between the rather meager GeForce RTX 3050 and the existing RTX 3060 12GB. The 3050 is beaten by AMD’s RX 6600 and Intel’s Arc A750 on both price and performance. However, reducing the memory capacity may lower the price compared to the 12GB card, but it also reduces the performance.
Our best guess is that Nvidia probably has a bunch of GA106 graphics processors, some of which may not have three fully functional 64-bit GDDR6 controllers, but potentially Almost all 30 streaming multiprocessors (SM) are available. This looks like a way to try to increase the price compared to the 3050 while still providing a modest performance boost.
Nvidia’s goal is to take every opportunity to sell off as many Ampere GPUs as possible in the coming months to pave the way for next-generation GPUs based on the Ada Lovelace architecture. I can’t help but think that lowering the price would be more successful than trying to bridge the gap between his existing 3050 8GB and 3060 12GB cards.
