TECH NEWS – We’ve heard more about the two “60s” Nvidia GPUs for the mid-range.
WCCFTech reported on the two graphics cards, which are based on the Ada Lovelace architecture. The pair will arrive on the market after the RTX 4070, as this model will be available in stores from April 13, while the duo one step down will be available sometime in May. The RTX 4060 Ti will be based on the AD106-350-A1 GPU core and will be a cut-back AD106. According to older rumors, it will have 34 SM, 4352 CUDA cores, 8 GB of GDDR6 memory, a 128-bit bus interface, 288 GB/s bandwidth, and 32 MB of L2 cache (its size will be eight times that of the predecessor 3060 Ti). The PG190 SKU will use 361 PCB, and the Founders Edition cards will also show up with a 16-pin 12VHPWR connector, while the custom models will use an 8/16-pin solution. Its TDP will be between 150-160W, 25% below the 3060 Ti’s claim. 22 TFLOPS of compute capability, which could be as high as 24 on the custom models.
On the RTX 4060, Nvidia may use the AD107-400-A1 GPU, the fully enabled AD107 Ada Lovelace core. This configuration is identical to the laptop RTX 4060. It will have 3072 CUDA cores, 8 GB of GDDR6 memory at 18 Gbps, a 128-bit bus interface with 288 GB/s bandwidth, 24 MB of L2 cache, and a TDP of 110-115W. Compared to its predecessor, the 12 GB RTX 3060, it could be 35% lower. Previously, it was rumored to use a PG190 PCB, but it has been changed to a PG173 SKU 371 PCB. The former can be used by Founders Edition and the latter by custom card manufacturers. You can get power with the standard 8-pin solution, but this card should also have mini-ITX versions.
The RTX 4060 Ti can compete with the RTX 3070 Ti. That was $600, but because of the 128-bit bus, it might only perform reasonably well at 1080p. Pricing will determine its success. It won’t be cheap either, that’s for sure.
Source: WCCFTech
