In a tech brief regarding its latest memory solutions, Micron mentioned an “RTX 3090,” which seems to suggest that NVIDIA will, in fact, release its flagship Ampere graphics card under that oft-rumored name. Micron even detailed its memory configuration thoroughly: we’re looking at a GPU with 12 GB of cutting-edge GDDR6X RAM, a type of memory that hasn’t even been officially announced yet. Memory clock is 19 to 21 Gbps, memory bus width is 384-bit, and memory bandwidth 912 to 1008 GB/sec.
Obviously, this is quite an improvement over the current flagship, the GeForce RTX 2080 Ti. That GPU only features 11 GB of GDDR6, whose memory speed and bandwidth maxes out at 14 Gbps and 616 GB/s, respectively. Its interface width is also limited to 352-bit.
You can check out Micron’s table below, which shows how the RTX 3090 compares to its other siblings, such as the TITAN RTX.
“In Summer of 2020, Micron announced the next evolution of Ultra-Bandwidth Solutions in GDDR6X,” an excerpt reads. “Working closely with NVIDIA on their Ampere generation of graphics cards, Micron’s 8Gb GDDR6X will deliver up to 21 Gb/s (data rate per pin) in 2020. At 21 Gb/s, a graphic card with 12pcs of GDDR6X will be able to break the ITB/s of system bandwidth barrier! Micron’s roadmap also highlights the potential for a 16Gb GDDR6X in 2021 with the ability to reach up to 24Gb/s. GDDR6X is powered by a revolutionary new PAM4 modulation technology for Ultra-Bandwidth Solutions. PAM4 has the potential to drive even more improvements in data rate.”