nak: Add MemOrder::Constant
to reduce the coherency of global loads.
The series brings the Sascha Willems computecloth example from 6 fps to 240 fps on my ampere machine, for a cool 40x improvement
Edited by Mel Henning
Due to an influx of spam, we have had to impose restrictions on new accounts. Please see this wiki page for instructions on how to get full permissions. Sorry for the inconvenience.
to reduce the coherency of global loads.
The series brings the Sascha Willems computecloth example from 6 fps to 240 fps on my ampere machine, for a cool 40x improvement