NVIDIA’s upcoming entry-level RTX in GPU-Z database
It looks like NVIDIA’s slowest Ampere GPU has finally been confirmed in a leak.
NVIDIA RTX 3080 Laptop GPU (GA104), Source: VideoCardz
Somebody submitted anonymously and publicly the RTX 3050 Ti laptop GPU to the GPU-Z validation database, ultimately confirming that RTX 3050 series use GA107 graphics processor. Both RTX 3050 Ti and RTX 3050 non-Ti are to be based on the GPU that has not yet even been used by RTX 30 desktop series. This GPU will top at 2560 CUDA cores, which also be the specs of the RTX 3050 Ti Laptop GPU, judging from this leak.
The validation confirms that GA107 GPU has a PCI Device ID of 25A0. The subvendor ID confirms it’s a Clevo/Kapok system and since those laptops are not yet availble on the market, we should assume that the GPU has been validated by Clevo engineers themselves.
The leak confirms the boost clock of RTX 3050 Ti of 1485 MHz and the memory clock of 1500 MHz. Since the graphics card features a 128-bit bus and uses 4GB memory in GDDR6 technology, the bandwidth has been automatically calculated at 192 GB/s. According to the validation, the GPU offers 32 ROPs. The validation also suffers from its typical error of showing the wrong TMU count. GPU-Z has been known to show incorrect values here for unreleased GPUs. In any case, it should show 80 TMUs.
NVIDIA GeForce RTX 3050 Ti specifications, Source: TechPowerUP
NVIDIA GeForce RTX 3050 series are expected to make a debut alongside Intel 11th Gen Core H-series (Tiger Lake-H), which is now scheduled to launch on May 11th. The RTX 3050 series will be the slowest Ampere gaming GPUs that should power mainly 1080p screens. Since NVIDIA has never added RT cores and Tensor cores to its XX50 series, we are not quite sure what to expect in terms of raytracing or DLSS performance yet.
| NVIDIA GeForce 30 Mobile | |||||
|---|---|---|---|---|---|
| VideoCardz.com | GeForce RTX 3080 | GeForce RTX 3070 | GeForce RTX 3060 | GeForce RTX 3050 Ti | GeForce RTX 3050 |
| Codename | GN20-E7 | GN20-E5 | GN20-E3 | GN20-P1 | GN20-P0 |
| GPU | GA104-775 | GA104-770 | GA106 | GA107 | GA107 |
| CUDA Cores | 6144 | 5120 | 3840 | 2560 | 2048 |
| GPU Clock | 1245 – 1710 MHz | 1290 – 1620 MHz | 1283 – 1703 MHz | 1030 – 1485 MHz | 1060 – 1500 MHz |
| Memory Bus | 256-bit | 256-bit | 192-bit | 128-bit | 128-bit |
| Memory | 8GB / 16GB GDDR6 | 8GB GDDR6 | 6GB GDDR6 | 4GB GDDR6 | 4GB GDDR6 |
| TGP | 80W – 150W | 80W – 125W | 60W – 115W | 60W – 80W | 60W – 80W |
Source: TechPowerUP
