RTX 4060 series are only equipped with PCI-E 4.0 x8 interface
NVIDIA is poised to launch the RTX 4060 Ti this month, possibly unveiling the RTX 4060 simultaneously, though the latter’s embargo lift won’t be as expeditious. Known variants of the RTX 4060 Ti currently include an 8GB and a 16GB version, with the former set for release this month and the latter in July. Furthermore, the RTX 4060 Ti and the RTX 4060 employ distinct GPUs, with the RTX 4060 Ti utilizing the AD106, boasting 4,352 CUDA cores, while the RTX 4060 harnesses the AD107, hosting only 3,072 CUDA cores.
Historically, AMD has tended to manipulate the PCI-E interface bandwidth on graphics cards of this caliber, a move not typically mirrored by NVIDIA. However, commencing with the RTX 40 series, NVIDIA seems to be aligning more closely with AMD. According to the specifications released by VideoCards, both the RTX 4060 Ti and the RTX 4060 employ a PCI-E 4.0 x8 channel. Despite this level of GPU bandwidth being ample, it is identical to PCI-E 3.0 x16. However, performance may be somewhat constrained on certain older platforms, such as Intel’s 10th generation Comet Lake, which only supports PCI-E 3.0, as well as AMD’s Ryzen 5000G series, which similarly has only PCI-E 3.0.
Reducing the number of PCI-E channels can simplify the PCB design for AIC partners, as it lessens the quantity of PCI-E channel-related PCB traces and SMDs by half. Much like memory traces, PCI-E traces are designed and validated by EDA software, ensuring that all channel lengths are equal to maintain signal integrity. Naturally, the difficulty is significantly diminished by the direct halving.