Efficient Parallel Implementation of Matrix Multiplication for Lattice-Based Cryptography on Modern ARM Processor

Table 2

Matrix transpose performance (Unit: ms).

N	M	L	C version	Proposed (NEON)
			(Auto-Vectorization)
536	1024	256	364.2304	0.446443
663	1024	256	630.0066	0.707373
816	1024	384	970.4782	1.78282
952	1024	384	1172.607	2.078113