Well, ATA 3.1 supports queued TRIM, and Linux 3.12 supports it as well, so there's no argument left. The funny thing is that fstrim is far more likely to cause noticeable stuttering than discard.

A lot has been said about enabling TRIM for non-Apple branded SSDs and the issue has become a lot less troublesome since `trimforce` was added to the more recent OS versions. But on 10.9 and earlier ...