Micron announced it’s begun mass production of GDDR6, the next-generation memory standard for video cards that’s expected to take over for GDDR5 in the larger market. Micron’s announcement notes in particular that the company has begun production of its 8Gb (1GB) GDDR6 memory ICs.
Micron recently achieved throughput up to 20GB/s on its GDDR6 solutions, said Andreas Schlapka, director of Micron’s compute networking business unit, in a statement. “In addition to performance increases, Micron has developed a deep partner ecosystem to enable rapid creation of GDDR6 designs, enabling faster time to market for customers looking to leverage this powerful new memory technology,” Schlapka said.
It’s kind of funny to see Micron talking about its deep partner network for GDDR6, not because such alternative use-cases don’t exist — I’m certain that were we to go digging, we’d find plenty of uses for GDDR — but because there’s one use-case that dominates the entire market, and we all know it: video cards. Even claiming new breakthroughs in markets like AI and ML are effectively plays towards video cards because, until custom silicon ships out to handle these tasks, video cards are still doing most of that work. In fact, the only partner mentioned in Micron’s press release is Rambus, and the less said about Rambus the better. (Fun fact: If you light candles in the shape of a pentagram and say “DDR” three times in a mirror, Rambus’ legal counsel appears and sues you for patent infringement.)
Ok, So, GPUs?
GDDR6 is expected to have speeds ranging from 10-14Gbps per pin, at 1.25-1.35v. As we’ve covered in the past, and congruent with other RAM launches, the overall thinking here is “faster clocks” and “lower power” while iterating on the GDDR platform. We’ve discussed before how GDDR6 could prove itself an alternative to HBM2 if it proves to offer similar bandwidth and better costs, and that the shape of next-generation high-end cards would tell us a great deal about which one of these two standards had won the overall memory war — except, of course, there aren’t any next-generation cards to talk about.
SK Hynix and Samsung have made their own GDDR6 memory announcements recently, which means everyone is ramping up production for GPUs that Nvidia CEO Jen-Hsun Huang has told everyone not to expect “for a long time.” Originally, we thought that meant an August launch for new GPUs. New rumors have suggested later dates, and it’s not clear what the situation is right now. AMD has no new GPUs planned for the rest of the year (again, as far as we know), which suggests that rumors of NV delays could hit these volume ramps as well.
Samsung Now Producing 16Gbit GDDR6 for Upcoming GPUs
Samsung is ramping up production of GDDR6 with an 18Gbit clock and a new process node.
With GDDR6 Memory Production Scaling Up, New GPUs Won’t Be Far Behind
SK Hynix expects to ramp up GDDR6 volume production in the near future — and they ain't ramping up production without customers to sell it to.
Micron: Full GDDR6 Production Underway for Nvidia’s RTX Launch
Micron has announced full production of GDDR6 as part of the run-up to the Nvidia RTX launch — but it's done so with some frankly peculiar language.