Smokeless powder is primarily cellulose, C6 H10 05, but in modern rifles pretty much burns up in the first few inches in front of the chamber. Or as I noted, all the powder that's going to burn will burn in that distance. How much burns (and how quickly) depends on the peak pressure, since progressive rifle powders are designed to burn most completely at a certain pressure.
IMR4895, for example, was designed to burn best at a slightly lower peak pressure than most newer powders, because it was primarily designed for the M-1 "Garand." Which is partly why a lot of older handloading information firmly stated the most accurate loads occurred at somewhat under "maximum" pressure: Hundreds of tons of the original military-surplus 4895 were used by a LOT of handloaders in the decades after WWII. Most of today's rifle powders, however, are designed to burn best at around 10,000 PSI more than IMR4895 (either the surplus powder, or today's newly produced IMR4895).
But with any of them, the oxygen content of the nitrocellulose is consumed long before the gas produced exits the muzzle. When the still very hot "oxygen-starved" gas exits, it re-ignites due to the fresh supply of oxygen. This produces muzzle flash, NOT still-burning powder granules.
This is exactly why my 1884 trapdoor Springfield .45-70 shoots very accurately with bullets in the 400-grain range and IMR-4895, yet results in a bunch of unburned granules, both in the barrel and out (which can be seen by laying a sheet on the ground in front of the muzzle). The pressure never gets anywhere near ideal for even IMR4895, but enough burns consistently to result in good accuracy. But despite that there's there's almost no muzzle flash, because the remaining powder granules are NOT burning--and the long barrel results in very little hot gas exiting the muzzle.
QL a computer simulation, which may or may not match reality, in varying degrees. This includes the percentage of powder burned.
