AMD Radeon RX 6000 “Big Navi” PCB Reportedly Leaked

Earlier today, AMD teased its Zen 3 CPUs and the RDNA 2 GPUs with an announcement slated to arrive on the 8th and 28th of October, respectively. While Big Navi is still more than a month away, the PCB of one of the RX 6000 GPUs has been leaked on the Chinese Bilibili forums.

This PCB is supposed to feature the Big Navi GPU, and the XT label on it further leads credence to that claim. While the GPU core is hidden behind the heatsink, you can just make out the positions for the GDDR6 memory stacks.

As per earlier rumors, it appears that this will be a 16GB variant, paired with a rather tame 256-bit bus. This means that we’ll get eight memory dies (32bit x 8) powering the supposed 4K capable GPU. It will be a shame if the top-end Navi 2x card ends up featuring a 256-bit bus as that will hamper its ability to compete with the RTX 3080 which features a massive 760GB/s bandwidth, thanks to the use GDDR6X memory and a 320-bit bus.

Like 1st Gen Navi, we’ll likely see 8GB and 16GB variants for the budget-oriented cards, and a single 16GB configuration for the flagship with as many as 80 CUs or 5120 stream processors

This image has an empty alt attribute; its file name is Eg1PIRuXsAEiiQ_-scaled.jpeg
Navi 21 (Big Navi) *Credits to Locuza

The most likely approach to achieve this configuration would be to opt for 16Gb dies, resulting in a 512 GB/s of bandwidth. While that’s still much lower than that of the RTX 3080, faster clock speeds and higher core counts may just help alleviate the gap. Keep in mind that NVIDIA is using a hybrid approach with the Ampere SM, meaning you’ve got two data-paths and as such, with mixed compute-graphics workloads, the FP32 performance will be much lower than the marketed 2x.



Computer hardware enthusiast, PC gamer, and almost an engineer. Former co-founder of Techquila (2017-2019), a fairly successful tech outlet. Been working on Hardware Times since 2019, an outlet dedicated to computer hardware and its applications.