News for hackers. Stuff that matters.
Re:How Meta trains large language models at scale by whazor on Thursday June 13, 2024 @03:07AM (#40666824)
I was imagining that some sys admin has to walk to the server, take out the GPU, blow against the PCI-E pins like a game cartridge, and put it back to try again.
Re:How Meta trains large language models at scale by iszomer on Thursday June 13, 2024 @06:45AM (#40668001)
More to do with bent pins, material obstruction, or something as trivial as cable management (eg: bundles of qsfp weighing down the ports that are press-fitted not soldered).