Right, take one paragraph out of that whole article to try to prove a point.

But let's look at the only paragraph where the 45% percent is actually mentioned in context.
"Curiously, write IOPS [they're referring to random writes] at QD1 again were slightly higher with OPAL than without — perhaps some caching or command batching is happening in the Samsung controller. But the critical aspect is that
software BitLocker dropped random write performance by 45% compared to hardware BitLocker."
So to be clear and fair, and not try to confirm one's existing bias.... they're talking about SOFTWARE encryption, in ONE test mode only, in ONE benchmarking program only. With OPAL, the speed was actually faster, but let's not mention that, right?