我有一个长期运行的查询，我想加快速度。查询的结果是一个新表。
这些表都是MYISAM，它运行在一个大型EC2实例上（m2.4xlarge，64GB RAM）。
系统使用情况如下：
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
17438 mysql 20 0 32.7g 7.1g 7420 S 2 10.6 1:35.82 mysqld
我的cnf的相关部分：
key_buffer = 32768M
max_allowed_packet = 96M
thread_stack = 192K
thread_cache_size = 8
sort_buffer_size = 2M
read_buffer_size = 2M
read_rnd_buffer_size = 8M
table_cache = 512
thread_concurrency = 8
bulk_insert_buffer_size = 2048M
max_write_lock_count = 1
# ~1/4 of memory of machine
max_heap_table_size = 16384M
tmp_table_size = 16384M
# ~1/4 of memory
myisam_sort_buffer_size = 17179869184
当我运行这个简单的查询时，它需要比我想象的要长得多，并且机器上的内存和CPU使用率很低。
查询解释计划如下所示：
mysql> explain SELECT encounter_id
-> FROM encounters e, sampled_patients sp
-> WHERE e.patient_id = sp.patient_id;
+----+-------------+-------+-------+---------------+------------+---------+--------------------+---------+-------------+
| id | select_type | table | type | possible_keys | key | key_len | ref | rows | Extra |
+----+-------------+-------+-------+---------------+------------+---------+--------------------+---------+-------------+
| 1 | SIMPLE | sp | index | patient_id | patient_id | 4 | NULL | 1537954 | Using index |
| 1 | SIMPLE | e | ref | patient_id | patient_id | 5 | noah.sp.patient_id | 1 | Using where |
+----+-------------+-------+-------+---------------+------------+---------+--------------------+---------+-------------+
2 rows in set (0.00 sec)
它必须查看~1.5M行，但它已编入索引。我怎样才能加快速度呢？