Re:Use SQL subqueries to count distinct 50x faster
by corresation on Friday January 24, 2014 @05:40PM
(#7118681)
The default is configured for magnetic disks, where random IO is brutally slow (on the order of 30-100 per second) and sequential reads are reasonably fast. But what if you're running it on a Micron P420m with 700,000 IOPS per second?
[ Parent
| on HN ]
Re:Use SQL subqueries to count distinct 50x faster
by jdub on Saturday January 25, 2014 @02:37AM
(#7120188)
700,000 I/Os per second per second? That is some serious storage acceleration you've got there.
Re:Use SQL subqueries to count distinct 50x faster
by Sami_Lehtinen on Sunday January 26, 2014 @08:34AM
(#7125013)
With proper SSD drives you can get seriously high IOPS.
https://www.fusionio.com/products/iodrive2-duo/
Re:Use SQL subqueries to count distinct 50x faster
by corresation on Sunday January 26, 2014 @01:46PM
(#7126230)
I think he was humorously noting that I erroneously said IOPS per second, akin to saying ATM machine, CD disc, etc.