Reading from /dev/zero and writing out to RAM with a more optimal block size of 256K yields a throughput of 1,536 MB/s Actually, in this article, Linux writing

Reading from /dev/zero and writing out to RAM with a more optimal block size of 256K yields a throughput of 1,536 MB/s Actually, in this article, Linux writing to RAM is slower than MacOS writing to SSD. Hard to conclude that "it's just OS X being slow". I suspect the hardware used is wildly different from the one in the original article.