Re:Samsung Announces 'Shinebolt' HBM3E Memory: HBM Hits 36GB...
by Scene_Cast2 on Thursday October 26, 2023 @01:33PM
(#38028917)
I've heard that running CFD on thr CPU is still quite popular and prevalent, despite GPU implementations being widely available. What is it about the algorithm that makes GPUs less suitable? Model sizes? Branching? Data (non-)locality?
[ Parent
| on HN ]
Re:Samsung Announces 'Shinebolt' HBM3E Memory: HBM Hits 36GB...
by bobim on Thursday October 26, 2023 @05:00PM
(#38031632)
Memory is one limiting factor, needing over 512 Gb is not uncommon.
But complex physics like chemical reactions, particle tracking, any user field functions, is probably what is tricky to implement on a GPU.