How Micron’s GDDR6X memory is the secret to unlocking 4K on Nvidia’s RTX 30-series card


Even with consoles running at a regular pace, PC hardware has found ways to stay ahead. This is usually due to innovations that affect how computer graphics work, and that’s happening right now with Nvidia’s next wave of GeForce RTX GPUs. RTX 3070, 3080, and 3090 microns use the new GDDR6X memory. This is a step above the traditional GDR6 VRAM, and may represent a generational leap forward in system-bandwidth performance for Micron’s PAM4 technology.

Pulse amplitude modulation (PAM4) is the name for pushing 2 bits of data over a connection instead of 1 bit. This is a technical concept, but essentially, it enables the system to call from 4 possible states instead of 2 with each burst of information. With PAM4, Micron and Nvidia are offering a significant boost to I / O data rates and bandwidth. The GDDR6X is also more power efficient. At 21Gb / s, the GDDR6X uses 15% less power per bit than GDDR6 at 14Gb / s.

Micron claims that the GDDR6X has a per pin data rate of 21Gb / s, when you combine all those pins and memory modules on something like the RTX 3090, the GPU has a total memory bandwidth of 936GB / s. He is at large. For comparison, the Xbox Series X’s memory bandwidth tops out at 560GB / s and PS5 448GB / s.

Simply put, this means that the GPU can pull massive amounts of data from its memory in an instant. And this will lead to better sports.

“[Gamers] Benefit from an increase in system-data rates, ”Micron Graphics owner Ralph Ebert said in a media presentation. “It enhances the data being transferred between the GPU and the memory interface. We believe that we are enabling 8K proposals and getting closer to tracing over real time. ”

Does this actually mean that GPUs will finally have enough bandwidth to handle 4K comfortably. In a world where the $ 1,200 RTX 2080 Ti struggled to hit 60 frames per second at 4K, the GDDR6X is crucial in helping us overcome that hurdle.

Leave a Reply