Notes on this library¶

This was imported into MBED specifically to support the GP9002 VFD, as a result it has some "hacks" to optimize it for the GP9002.

Due to the GP9002 internal organization it draws vertical lines much faster than horizontal (assuming you orient it "landscape"). This is likely to be true of other displays that have bits in a byte arranged vertically, which seems a common theme in small GLCDs. Some types may have a more CGA-like memory organization and will draw faster horizontally.

On a vertical-organised display the graphics functions are often substantially faster if X and Y are exchanged, especially with the dot-write optimization. This is because vertical lines can be written byte-at-a-time with no need for read-modify-write, and even when individual bits are written a significant number will "land" in the byte previously written. In contrast a horizontal line would require changing one bit of each byte in turn, requiring a sequence of set-address,read,write operations for each dot in turn.

I've hacked this in the library. I forget exactly how but I believe I simply exchanged X with W in the code for drawing filled shapes.

I would like to come up with a more generic way to do this, such as having internal coordinates that are not defined as X and Y, then leaving it up to the display library to "wrap" them in the way that is best for that display.

I would like to apologize for abandoning this project, but the test harness still exists and I might return to it, though I'm more interested in TFT systems like the STM discovery now.