FWIIW:
I think you are probably correct.
I have 2 Windows 10 virtual systems that are several years old. As time went on, I saw more incidents of the VM "pausing" for a few seconds. Presumably for housekeeping? When I updated my host a while back, I put the VM's on an NVME drive (1tb) and that took care of that. I think the move from 32G to 64G RAM made a difference although I had not seen swapping.
The systems perform well although I have not run benchmarks. I think the VM overhead has increased as the OSes put more and more protection crap in the game, especially Windows. Unavoidable in today's malware environment. It takes some effort to setup AV's to avoid slowdowns.
Apples to oranges. I don't have a Windows host.
That said, here is a comparison of disk read times from a Linux VM running on my Linux host. Same NVME drive:
Guest (sda is on /dev/nvme1n1):
sudo hdparm -t /dev/sda
/dev/sda:
Timing buffered disk reads: 982 MB in 3.00 seconds = 326.90 MB/sec
Host:
sudo hdparm -t /dev/nvme1n1
/dev/nvme1n1:
Timing buffered disk reads: 7522 MB in 3.00 seconds = 2507.08 MB/sec
7:1 seems rather large. I have made no attempt to understand.
We run production stuff on ESXi, I use WS for programming and test. I think I still have WS 4.0 on the shelf around here somewhere. 