NV12 copies are much slower than similar-sized resources

[Park__James]

System Setup Information:
-----------------------------------------
System Used: ASUS Z170-DELUXE
CPU SKU: i7-6700
GPU SKU: HD Graphics 530
Processor Line: ???
System BIOS Version: ???
CMOS settings: ???
Graphics Driver Version: 26.20.100.6709
GOP/VBIOS Version: ???
Operating System: Windows
OS Version: 10 1903
API: Direct3D 11
Occurs on non-Intel GPUs?: No

Steps to Reproduce:
-------------------------------
1. SetStablePowerState: https://developer.nvidia.com/setstablepowerstateexe-%20disabling%20-gpu-boost-windows-10-getting-more-deterministic-timestamp-queries
2. Open attachment in Visual Studio 2017.
3. Note elapsed time in debug output when running.
4. Resource setup can be altered using the COPY_MODE #define.

Expected Results:
-------------------------------
NV12 copy performance is consistent with formats of similar sizes.

Actual Results:
-------------------------------
NV12 copy performance is slower than formats of similar sizes.

Additional Information:
-------------------------------
I'm not sure if this is the driver, hardware, or bad profiling code, but I'm measuring very slow copy speeds using NV12 formats.

Intel HD Graphics 530 CopyResource, USAGE_DEFAULT -> USAGE_STAGING
NV12(1920x1080): 2355 us
R8(1920x1080), R8G8(960x540): 361 us
R8G8B8A8(1920x405): 711 us

By comparison, times on my NVIDIA card are consistent with similarly sized formats using the same sample.

NVIDIA RTX 2080 Ti CopyResource, USAGE_DEFAULT -> USAGE_STAGING
====
NV12(1920x1080): 273 us
R8(1920x1080), R8G8(960x540): 240 us
R8G8B8A8(1920x405): 240 us