Hackernews
new
show
ask
jobs
A Thread-Register Decoupled GPU Execution Model for Efficient Tensor Computation
16 points
posted 7 hours ago
by matt_d
(arxiv.org)
No comments yet