News for hackers. Stuff that matters.

A comment on Meta MTIA v2. Parent comment

Re:Meta MTIA v2 by moffkalast on Wednesday April 10, 2024 @02:03PM (#39993843)
It would be interesting if this could be made into a reasonably priced (lmao) card for home inference if they intend to mass produce it.

Can't imagine any other reason other than cost as to why they went with LPDDR5, LPDDR5X has more bandwidth and GDDR6 has even more.

[ Parent | on HN ]
Re:Meta MTIA v2 by chessgecko on Wednesday April 10, 2024 @02:11PM (#39993934)
they didn't use GDDR cause they wanted the memory capacity which is really important for recommendation models. But I totally agree that this is a sort of perfect cost/perf per watt point for a home setup. I really hope they do it, if not for this one at least for v3.
[ Parent | on HN ]