A shared-memory CUDA port speeds up geodesic shooting landmark matching by nearly two orders of magnitude over naive CPU code.
Benchmarking the cost of thread divergence in CUDA
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2019 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Fast geodesic shooting for landmark matching using CUDA
A shared-memory CUDA port speeds up geodesic shooting landmark matching by nearly two orders of magnitude over naive CPU code.