-
sech1preliminary results for software AES on Orange Pi RV2: 75.28 ns/iteration (scalar code) vs 48.82 ns/iteration (vector code)
-
sech11.5x speedup, less than expected but still good
-
sech1I guess it gets bottlenecked by the random table lookups
-
sech1correction: the number above is for 2 AES rounds per iteration, so 1 AES round takes half of this time
-
sech1CPU speed is 1.6 GHz, so it's 60 clock cycles per round for scalar and 39 cycles per round for vector code
-
sech1lol, AES got 2x faster in XMRig, but the other parts got slower, the end result is almost negligible :D
-
sech1
-
sech1
-
sech1bottlenecked by memory (this CPU doesn't have enough cache for the scratchpad)
-
sech1still ~1% faster with vectorized soft aes
-
sech1I need to test hashrate with 512 KB scratchpad and a single thread, to see the pure performance
-
sech1
-
sech1soft aes itself is 2.2x faster
-
sech1
