Hi
I have a quite large table storing words contained in email messages

mysql> explain t_message_words;
+----------------+---------+------+-----+---------+----------------+
| Field          | Type    | Null | Key | Default | Extra          |
+----------------+---------+------+-----+---------+----------------+
| mwr_key        | int(11) | NO   | PRI | NULL    | auto_increment |
| mwr_message_id | int(11) | NO   | MUL | NULL    |                |
| mwr_word_id    | int(11) | NO   | MUL | NULL    |                |
| mwr_count      | int(11) | NO   |     | 0       |                |
+----------------+---------+------+-----+---------+----------------+

table contains about 100M rows
mwr_message_id is a FK to messages table
mwr_word_id is a FK to words table
mwr_count is the number of occurrencies of word mwr_word_id in message mwr_message_id

To calculate most used words, I use the following query

SELECT SUM(mwr_count) AS word_count, mwr_word_id
FROM t_message_words
GROUP BY mwr_word_id
ORDER BY word_count DESC
LIMIT 100;

that runs almost forever (more than half an hour on the test server)

mysql> show processlist;
+----+------+----------------+--------+---------+------+----------------------+-----------------------------------------------------
| Id | User | Host           | db     | Command | Time | State                | Info
+----+------+----------------+--------+---------+------+----------------------+-----------------------------------------------------
processlist
| 41 | root | localhost:3148 | tst_db | Query   | 1955 | Copying to tmp table | SELECT SUM(mwr_count) AS word_count, mwr_word_id
    FROM t_message_words
    GROUP BY mwr_word_id |
+----+------+----------------+--------+---------+------+----------------------+-----------------------------------------------------
3 rows in set (0.00 sec)

Is there anything I can do to "speed up" the query (apart from adding more ram, more cpu, faster disks)?

thank you in advance
stefano