As I read it, the article assumes that the 20% performance penalty measured when writing large blocks is almost entirely due to APFS performance. This seems reasonable, since the number of syscalls will be low.

However, it is then also assumed that this 20% performance penalty between HFS+ and APFS is going to be flat no matter which block size is used. I don't think this is a reasonable assumption.

You are comparing the maximum write data rate of HFS+ and APFS with the write rate of mkfile with small blocks. I do not think you can assume that the write rate in the filesystem stays the same if you decrease the block size. The expensive operation might be the block allocation that is easy for larger blocks, but for small blocks it might try to squeeze them in somewhere to avoid fragmentation.

Yes, syscalls are expensive and most likely became very expensive with the Meltdown patch. Even more so on OS X, where filesystem drivers are running in their own process.

I am not arguing against performance drops due to both Meltdown path or APFS, but I do not think your data shows the conclusion you are using in the headline.