One of many reasons for unexpected (inferior) performance may be that on your UAT machine QVS isn't claiming any substantial amount of memory, leaving a larger chunk to the QVB process. While on your production machine, many users have many documents open (causing QVS to theoretically run off with as much as 90% of system memory), leaving not enough memory to do a reload entirely in RAM... As soon as the QVB process starts trashing, performance goes south...
You will probably understand that there is no one-fits-all answer to this question. As Rob already indicated, it depends on machinery and circumstances: available cores, available RAM, memory speed, processor speed, I/O channel bandwidth, number of processes, I/O load, CPU load, etc. Task manager may offer you your first view on possible reasons for sub-optimal performance.
I have made changes to the task to run the first thing in the morning when nothing else is running. Hopefully I will know if it finishes or runs for long time.
You can compare the server document log with local document log and see if the time differences are on particular statement (which ones?) or all (in this case it should be a problem of server resources).
