How to explain dramatically different MongoDB read performance?

I made a PHP script to benchmark some basic reads & writes in MongoDB and ran it on 2 different virtual servers. One server had about 3x the RAM & CPU power of the other. This strong server was able to insert about twice as many entries per second as the weaker server; however, the strong server was only able to read from the MongoDB database at about half the speed of the server with less resources. Any ideas as to what could cause that (I’m aware that it quite possibly could be due to these being virtual servers, but I’m wondering if there are settings that could be affecting this)? Both are running CentOS 5.6 and MongoDB 1.8.1.

I hate it when this happens. It turns out I made a typo when I created the indexes, so there wasn’t an index on the larger server.