How to (and should I) change local parallel profile core number to enable hyper threading from command line

I currently have a set of analysis functions which make use of webread to pull data from a web service, and then process this data. The sleep time while webread is waiting for a response from the server is greater than the analysis time within matlab. A couple of questions.

1) Is it possible to enable hyper threading within matlab to allow for more webreads to be running at once, along with calculations? If so to what extent (i.e. limited by number of virtual cores assigned? or can the number of threads be greater?)

2) Would this provide any appreciable time reduction in the runtime of the script under the assumption that the majority of the time is waiting for the server to respond as opposed to calculations?

3) The computing environment I have access to requires that queries be submitted in bash script format. My current script is as follows.

#!/bin/bash ./matlab -nojvm -nosplash -r "function($SGE_TASK_ID)"

where $SGE_TASK_ID is an environment variable that allows for batch jobs on various machines. The question is, is there any way to change the number of threads allotted to the local parpool/create a new parpool with more threads and make it the default from the command line?

Thanks.