Performance comparison in sample alignedTypes not significant

[vipcxj]

This is my result:

Starting...
GPU Device 0: "Turing" with compute capability 7.5

[NVIDIA GeForce RTX 2080 Ti] has 68 MP(s) x 64 (Cores/MP) = 4352 (Cores)
> Compute scaling value = 1.00
> Memory Size = 49999872
Allocating memory...
Generating host input data array...
Uploading input data to GPU memory...
Testing misaligned types...
uint8...
Avg. time: 5.870122 ms / Copy throughput: 7.932716 GB/s.
        TEST OK
uint16...
Avg. time: 0.653178 ms / Copy throughput: 71.291444 GB/s.
        TEST OK
RGBA8_misaligned...
Avg. time: 0.411919 ms / Copy throughput: 113.046586 GB/s.
        TEST OK
LA32_misaligned...
Avg. time: 0.221544 ms / Copy throughput: 210.188781 GB/s.
        TEST OK
RGB32_misaligned...
Avg. time: 0.203884 ms / Copy throughput: 228.394200 GB/s.
        TEST OK
RGBA32_misaligned...
Avg. time: 0.192203 ms / Copy throughput: 242.274994 GB/s.
        TEST OK
Testing aligned types...
RGBA8...
Avg. time: 0.372953 ms / Copy throughput: 124.857542 GB/s.
        TEST OK
I32...
Avg. time: 0.372953 ms / Copy throughput: 124.857542 GB/s.
        TEST OK
LA32...
Avg. time: 0.217597 ms / Copy throughput: 214.001280 GB/s.
        TEST OK
RGB32...
Avg. time: 0.193803 ms / Copy throughput: 240.274804 GB/s.
        TEST OK
RGBA32...
Avg. time: 0.194241 ms / Copy throughput: 239.733621 GB/s.
        TEST OK
RGBA32_2...
Avg. time: 0.191466 ms / Copy throughput: 243.208190 GB/s.
        TEST OK

[alignedTypes] -> Test Results: 0 Failures
Shutting down...
Test passed

I found almost no difference between the aligned and unaligned versions.
uint8 and uint16 version is so slow, but there are no aligned version compared to them.