A few more optimizations, about 1% overall speedup.
Unroll horizontal pass, no more intermediate buffer, faster special transpose. Change-Id: I05df75be4e5f01420066cdf3c61a2edf35bedb64
Loading
Please register or sign in to comment
Unroll horizontal pass, no more intermediate buffer, faster special transpose. Change-Id: I05df75be4e5f01420066cdf3c61a2edf35bedb64