NVIDIA has announced its RTX 4060 family, including the 4060 and the 4060 Ti. The latter is slated to launch on the 24th of May at an MSRP of $399 for the 8GB variant and $499 for the 16GB model. Both GPUs will leverage the AD106 graphics core with 4,352 FP32 cores spread across 34 SMs with a peak boost clock of 2.54GHz. The GPU will be paired with 8GB of GDDR6 memory clocked at 18Gbps via a 128-bit bus and 32MB of L2 cache.
Despite its slimmer 128-bit bus, it’ll boast an effective bandwidth of 554GB/s, courtesy of the wider L2 cache, placing it ahead of the RTX 3070. Additionally, the 4060 Ti has a TBP of 160W, 60W less than the 3070, and 40W less than the 3060 Ti. Regarding theoretical throughput, it boasts 22 TFLOPs of FP32, up from 20 TFLOPs and 16 TFLOPs on the 3070 and 3060 Ti, respectively.
The performance deltas between the 8GB and 16GB variants of the RTX 4070 Ti are minimal. As per NVIDIA’s benchmarks, it is 10% to 25% faster than its predecessor, not exactly an impressive number. On average, however, we are looking at a gain of only 10% to 15%.