NVIDIA GeForce RTX 3050 Ti Features 2560 & RTX 3050 Features 2048 CUDA Cores With Ampere GA107 GPU

     

Specifications of NVIDIA's upcoming GeForce RTX 3050 Ti and GeForce RTX 3050 graphics cards have leaked out along with their first benchmarks. The specifications & benchmarks were leaked within the Geekbench database and tested on unreleased laptops featuring Intel's Tiger Lake-H processors.

NVIDIA GeForce RTX 3050 Ti & GeForce RTX 3050 GPUs Spotted - Feature GA107 GPU With 2560 & 2048 CUDA Cores, Respectively

The NVIDIA GeForce RTX 3050 series would include both the Ti and Non-Ti variants. The two GPUs that were spotted are laptop variants but we have seen that laptop Ampere variants feature the same core specifications as the desktop variants. The GeForce RTX 3050 series will also be utilizing the Ampere GA107 GPU which has yet to make an appearance on the desktop or laptop segment. The GPUs have shown up in Samsung laptops featuring the Tiger Lake-H (45W) CPUs, namely the Core i7-11800H and the Core i5-11400H.

Coming to the specifications, the NVIDIA GeForce RTX 3050 Ti features 20 SMs or 2560 CUDA Cores. This should be the full-fat GA107 GPU configuration. The GPU was running at a maximum clock of 1.03 GHz and also features 4 GB of GDDR6 memory running across a 128-bit bus interface. NVIDIA is likely to feature 12 Gbps memory modules so we can expect a total bandwidth of 192 GB/s. The GPU should have a TGP of sub-100W. The NVIDIA GeForce RTX 3050 Non-Ti variant features 16 SM units or 2048 CUDA cores. The GPU was running at a maximum frequency of 1.06 GHz. These are most likely the laptop-tuned clocks and we can expect higher in the desktop parts which will hit the retail segment in the coming months.

NVIDIA GeForce RTX 30 Mobility GPU Lineup:

GPU Name	NVIDIA GeForce RTX 3050 Ti	NVIDIA GeForce RTX 3060	NVIDIA GeForce RTX 3070	NVIDIA GeForce RTX 3080
Process Node	Samsung 8nm	Samsung 8nm	Samsung 8nm	Samsung 8nm
GPU SKU	GA107?	GA106	GA104-770	GA104-775
SMs	TBA	30	40	48
CUDA Cores	TBA	3840	5120	6144
Base Clock	TBA	1283 MHz	1290 MHz	1245 MHz
Boost Clock	TBA	1703 MHz	1620 MHz	1710 MHz
Memory Clock	12 Gbps	12 Gbps	12 Gbps	12 Gbps
Memory Type	GDDR6	GDDR6	GDDR6	GDDR6
Memory Size	4 GB	6 GB	8 GB	8/16 GB
Memory Bus	128-bit	192-bit	256-bit	256-bit
Bandwidth	192 GB/s	288 GB/s	384 GB/s	384 GB/s
TGP	60W?	60-115W	80-125W	80-150W+
Configurations	Max-Q Max-P	Max-Q Max-P	Max-Q Max-P	Max-Q Max-P
Launch	1H 2021?	Q1 2021	Q1 2021	Q1 2021

As for performance, in OpenCL, the NVIDIA GeForce RTX 3050 Ti is about as fast as the Radeon RX 5600 XT and the GeForce GTX 1080 Ti. The GeForce RTX 3050 is about as fast as the GTX 1080. The Non-Ti variant is also slightly faster than the GeForce GTX 1650 SUPER which is what it replaces by the time of launch. Also, this is just OpenCL performance and the Ampere architecture offers a huge increase in performance in OpenCL & Vulkan APIs so average performance in DirectX titles would be slightly lower.

The other key feature is that the GeForce RTX 3050 Ti is listed as an "RTX" variant. Previously, ray tracing was kept exclusive to **60 series graphics cards and above but this changes with Ampere and now we can get ray tracing on entry-level graphics cards too. However, the pricing of the GPU will be higher with the RTX 3050 Ti expected to be set at around $229-$279 US while the RTX 3050 will cost just slightly below the sub-$200 US range.

It seems like neither NVIDIA nor AMD will have a decent sub-$150 US solution this generation but that remains to be seen. As for TGP, the GeForce RTX 3050 Ti is expected to feature a 60W design and is expected to make use of the GA107 GPU. Intel's Tiger Lake-H and NVIDIA GeForce RTX 30 series mobility GPU-based laptops are going to start at around $999 US pricing so expect it to be quite popular amongst mainstream gamers who require portable and budget-friendly gaming solutions.

News Sources: TUM_APISAK #1 , #2