gistfile1.txt

Table create query:

CREATE TABLE `records` (
  `id` int(11) unsigned NOT NULL AUTO_INCREMENT,
  `number` int(11) DEFAULT NULL,
  PRIMARY KEY (`id`)
) ENGINE=InnoDB AUTO_INCREMENT=1000002 DEFAULT CHARSET=utf8;

Filled the table with one million rows, with `number` being a random integer between 0 and 100.

Both these queries take exactly the same amount of time:

SELECT COUNT(*) FROM (SELECT * FROM records WHERE number > 1) AS records;
SELECT COUNT(*) FROM records WHERE number > 1;

The `EXPLAIN`s for these queries are also the same.

Adding an ORDER BY to the subquery adds another 2-4ms to the query time:

SELECT COUNT(*) FROM (SELECT * FROM records WHERE number > 1 ORDER BY number) AS records;

But adding an ORDER BY to the regular query doesn't add any time at all:

SELECT COUNT(*) FROM records WHERE number > 1 ORDER BY number;