to is not incremented because Duff was copying to a single memory-mapped output register. In modern C this would be communicated by the use of the
volatile keyword as a qualifier in the declaration of the pointer
While optimizing this, Duff realized that an unrolled version of his loop could be implemented by interlacing the structures of a switch and a loop.
Based on an algorithm used widely by programmers coding in assembly for minimizing the number of tests and branches during a copy, Duff's device appears out of place when implemented in C. The device is valid, legal C by virtue of the relaxed specification of the switch statement in the language's definition. At the time of the device's invention this was the first edition of The C Programming Language which requires only that the controlled statement of the switch be a syntactically valid (compound) statement within which case labels can appear prefixing any sub-statement. In conjunction with the fact that, in the absence of a break statement, the flow of control will fall-through from a statement controlled by one case label to that controlled by the next, this means that the code specifies a succession of count copies from sequential source addresses to the memory-mapped output port. Note that, as documented in the comment, the code assumes that count is strictly positive.
Many compilers will optimize the switch into a jump table just as would be done in an assembler implementation. C's default fall-through in case statements has long been its most controversial single feature; Duff observed that "This code forms some sort of argument in that debate, but I'm not sure whether it's for or against.
The primary increase in speed versus a simple, straightforward loop comes from loop unwinding, which reduces the number of comparisons performed during a loop. The switch/case statement is used to handle the remainder of the data not evenly divisible by the number of operations unrolled (in this example, 8 byte moves are unrolled, so the switch/case handles an extra 1–7 bytes automatically).
This automatic handling of the remainder may not be the best solution on all systems and compilers — in some cases two loops may actually be faster (one loop, unrolled, to do the main copy, and a second loop to handle the remainder). The problem appears to come down to the ability of the compiler to correctly optimize the device, although at least one person has suggested that it may also interfere with pipelining and branch prediction on some architectures. Therefore, when considering using this code, it may be worth running a few benchmarks to verify that it actually is the fastest code on the target architecture, at the target optimization level, with the target compiler.
*to++ = *from++;
This modified form of the Device appears as a "what does this code do?" exercise in Bjarne Stroustrup's book The C++ Programming Language, presumably because novice programmers cannot be expected to know about memory-mapped output registers. However, the standard C library provides the function memcpy for this purpose; it will not perform worse than this code, and may contain architecture specific optimisations that will make it significantly faster.