The idea is that calculation of each chunk takes 0.5 second. It will start from 1000 rows, if this query takes more than 0.5 seconds, it next chunk will be with less rows, if the chunk takes less than 0.5 second, the next chunk it will take more rows and so on.
This is one of the safeguards that pt-table-checksum have and you can always modify this values (–chunk-time, --chunk-size, etc), also it will monitor the server load, replication lag, etc.
Please take a look to the documentation for more information. http://www.percona.com/doc/percona-toolkit/2.1/pt-table-chec ksum.html