Short stroking and Partitioning are 2 very different things.
When you Partition a 1 TB drive say into 2 500GB partitions that means that every track on every platter has the ability to be written to by the Operating System. albeit in 2 different "buckets"
Short Stroking is only using the outer 1/3rdish part of each platter in the drive in total. That's it. so on a 1TB drive you would theoretically have 300GB or so of usable capacity.
Short Stroking a single drive on a PC serves no purpose other than to only use a portion of what you paid for in reality.
The reason to short stroke is for performance purposes. Every drive weather SATA, FC, SAS has a limited number of operations it can perform called IOPS (Input/Output Operations Per second). IOPs is like a highway. If I have a 1 lane road I can expected x amount of cars/hour before we see traffic (higher response time), however If I have a 3 lane highway I can now handle an increased amount of work and have less traffic. Notice I did not say 3x the amount of work because the workload does not increase lineally. Also to complete the analogy if the number of lanes is the IOPS, then the amount of cars I can put through those lanes would be the throughput.
A 7200 RPM drive I can expect ~ 75 iops, for a 15K rpm SAS drive I can expect ~ 175 iops. Now that is just the "Average" iops expected per disk. Block size, sequential vs random access, and response time all come into play as well, but for the sake of simplicity let's just limit the conversation to IOPS.
If I create a RAID group of 3 15K disks, they then act in unison and, again, just keeping it simple, I now have a disk "unit" capable of performing 525 IOPS (175x3).
Now If I only write to 1/3 of those 3 disks platters, because my goals was to have higher iops capability, I have achieved that but at the cost of useable capacity.
So extrapolating that out to high end storage arrays like NetApp, EMC, IBM and others, short stroking is a performance enhancement technique that allows for higher IOPS, and lower response time (because I am only writing to the fastest part of the internal disks) and now my storage arrays are capable of reading/writing data very fast.