I'm the author of gdisk. I modelled it closely after fdisk, so it works in almost exactly the same way. The biggest single difference is the fact that gdisk uses two-byte partition type codes, but mostly those are the same as the MBR/fdisk codes with "00" tagged on, as in "8300" for Linux or "8200" for Linux swap. These codes get translated into the GUIDs that the GPT data structures use, so you don't need to deal with GUIDs directly unless you want to.
Modern versions of fdisk, like gdisk, deal in sector numbers rather than cylinders. The truth is that "cylinders" have evolved from being a usable reality (in the 1980s) to a convenient fiction (in the 1990s) to an inconvenient fiction (in the 2000s) to a serious problem (in the 2010s). Thus, if you've been thinking about disk layout in terms of cylinders, it's long past time that you begin thinking about it in terms of sectors (or blocks; the two terms are effectively synonymous in this context). Note that both fdisk and gdisk accept partition sizes in terms of mebibytes, gibibytes, and tebibytes, as well as in sector values. For instance, you can enter "+200G" as the partition size to create a 200GiB partition.
You might find cgdisk a little less confusing than gdisk; cgdisk is modelled after cfdisk, so it offers fewer options but has a somewhat friendlier user interface. The two programs (gdisk and cgdisk) use exactly the same underlying partition-manipulation code, though; they've just got different user interfaces.
For more on using gdisk, see its man page or
Web-based documentation, and particularly its
gdisk walkthrough and
cgdisk walkthrough pages.