October 5, 2022

Micron’s ultra-fast 24 Gbps GDDR6X memory is entering mass production and will soon be available on next-generation NVIDIA GPUs!

Well, it looks like Micron has officially started mass production of its next-generation GDDR6X memory chips offering speeds of up to 24 Gbps.

Micron’s 24Gbps GDDR6X memory goes into mass production and will be available on NVIDIA’s next-gen GPUs later this year

Micron announced earlier this year that it was readying its all-new 24 Gbps GDDR6X memory modules for future graphics cards. The announcement came at a time when NVIDIA’s GeForce RTX 3090 Ti had just entered the market for which Micron had produced 16GB DRAM modules rated at speeds of up to 21Gbps.

Now Micron is going to offer even faster 24Gbps spindle speeds in 16GB density, which means that once again we’ll be looking at 2GB VRAM capacities and up to 24GB on an interface 384-bit bus. Much like the GeForce RTX 3090 Ti which featured all its memory modules on the front of the PCB, the next-gen GeForce RTX 4090 would feature its 12 GDDR6X modules on the front, allowing for more cooling compared to solutions that feature modules of memory on the back of the PCB such as the RTX 3090 (Non-Ti).

While full 24Gbps speeds aren’t meant to be utilized by the NVIDIA GeForce RTX 4090 (rumored to offer die speeds of up to 21Gbps), 24Gbps GDDR6X memory modules can deliver up to 1.152 TB/s of bandwidth, which is a 14% increase over the 1.008 TB/s offered on current 21 Gb/s GDDR6 memory modules.

Micron has begun mass production of its 24 Gbps GDDR6X memory modules for NVIDIA Next-Gen GPUs. (Image credits: Harukaze5719)

Here are the bandwidth figures you can expect with a 24 Gbps DRAM solution:

  • 512 bit solution – 1.5 TB/s
  • 384 bit solution – 1.1 TB/s
  • 320 bit solution – 960 GB/s
  • 256 bit solution – 768 GB/s
  • 192 bit solution – 576 GB/s
  • 128 bit solution – 384 GB/s
  • 92 bit solution – 276 GB/s
  • 64 bit solution – 192 GB/s

The NVIDIA GeForce RTX 4090 will be among the first next-gen graphics cards to use the latest GDDR6X memory modules when it launches later this year. Since these chips are likely to be downclocked to meet power targets, there will be huge overclocking potential for enthusiasts to take advantage of.

AMD, on the other hand, will likely depend on its partner, Samsung, for its next-gen RDNA 3 lineup. Samsung is also working on a 24Gbps GDDR6 memory module that is expected to go into mass production soon.

GRAPHICS MEMORY GDDR5X GDDR6 GDDR6X
Workload Chart GraphicsAI Inference Accelerator GraphicsAI Inference Accelerator
Platform (Example) TitanX Titan RTX
RX5700XT
GeForce® RTX™ 3090Ti
GeForce® RTX™ 3080Ti
Number of slots 12 12 12
GB/s/pin 11.4 14-16 19-24
GB/s/slot 45 56-64 76-96
GB/s/system 547 672-768 912-1152
Settings (Example) 384 IO
(12pcs x 32 IO package)
384 IO
(12pcs x 32 IO package)
384 IO
(12pcs x 32 IO package)
Frame buffer of a typical system 12 GB 12 GB 24 GB
Average device power (pJ/bit) 8.0 7.5 7.25
Typical IO channel pcb
(SM P2P)
pcb
(SM P2P)
pcb
(SM P2P)

News source: Harukaze5719

Products mentioned in this article