TECH NEWS – The AD107-400 will appear on Nvidia’s GeForce RTX 4060 graphics card.
Nvidia has finally started bringing down the price of its new graphics cards. These include the announced but not yet available RTX 4060, a card based on TSMC’s 4-nanometer process node with the AD107-400 GPU. It is the smallest class 107 chip in years. At just 346 square millimeters, it’s 27% smaller than both the GA107 (Ampere architecture, 200 mm2, 8.7 billion transistors) and the TU117 (Turing architecture, also 200 mm2, 4.7 billion transistors). We don’t know how many transistors it has yet, but the chip has already appeared in the laptop RTX 4060 and RTX 4050, and Nvidia will also use it in the desktop versions.
The image was leaked by MEGAsizeGPU, and based on that, it’s a safe bet that Nvidia will use the AD107-400-A1 GPU, which is the entire chip so that it won’t be a cut back release. (The company will cut it back for the RTX 4050, but that will be entry-level… until they make an RTX 4030, but we don’t see much chance of that.) The RTX 4060 will have 3072 CUDA cores and 8 GB of GDDR6 VRAM. It has 17 Gbps with a 128-bit interface, so it has a bandwidth of 272 (effectively 453) GB/s, a speed 26% higher than the RTX 3060. It will also have 24 MB of L2 cache. The RTX 4060 will have 35% lower TGP than its predecessor, the 12 GB (!?) RTX 3060. It will consume 115W, about 110 in gaming, and about 7W in sleep mode.
There will be no Founders Edition version of the Nvidia GeForce RTX 4060, released in July for $300. The company will launch the 16GB version of the RTX 4060 Ti at the same time. So the card’s pricing is starting to reach the mainstream consumer segment. It’s not as expensive as a PlayStation 5, but with a decent processor, 16 GB of RAM, and a suitable power supply, it’s already good for 1080p performance.
Source: WCCFTech
