The command ".SESSIONS" is typically used to provide Teradata load tools parallelism.
When you are loading data into Teradata, you must provide the used tool with a file to be loaded. Generally speaking, Teradata load utilities opens the provided file and start to read it line by line, in order to send the read lines to the Teradata RDBMS. The lines are sent synchronously, what means that the tool should wait for a OK from Teradata to send a new row. If you define a number of sessions bigger than 1, the opened sessions should be used to receive rows in parallel, what would increase your load performance.
If you use the command ".sessions 8;", for example, you would have 8 sessions opened in Teradata to receive data in parallel, which could be understood as "8 channels" to send data to Teradata in parallel.
The number of available AMPs is a good guessing to initially define the number of sessions, but you may have to increase or decrease this number to tune you process to reach a good load rate.