undefined | Better HN

0 pointsHeatrayEnjoyer1y ago0 comments

Only if the inference software doesn't guarantee concurrency, which is CS 101

0 comments

This sort of nondeterministic scheduling of non associative floating point ops is essentially running at the level of GPU firmware, so, I would imagine that in this case, Nvidia is aware.

j / k navigate · click thread line to collapse

0 pointsHeatrayEnjoyer1y ago0 comments

Only if the inference software doesn't guarantee concurrency, which is CS 101

0 comments

pizza1y ago

This sort of nondeterministic scheduling of non associative floating point ops is essentially running at the level of GPU firmware, so, I would imagine that in this case, Nvidia is aware.

j / k navigate · click thread line to collapse