So as a few people have shared their allocated RAM for A+ 512M up to 8GB and beyond.... is there any benefit beyond caching/readying of the tracklist after the song currently playing if you are not up-sampling ?
Increasing allocated RAM is useful when you need to store temporary a large amount of computed data ready to be consume in another computation cycle. In our case, we're dealing with piece of data (tracks), which have to be loaded in RAM, treated to be ready to be sent to audio interface, and finally consumed by it. Preparing a whole track helps to prevent input buffer from interface getting empty. Preparing the next track is even better, so you will move from one track to another without any latency. Preparing the whole tracklist could be interesting if you want to play suddenly a track from your list, or if your player would compute the next track index at the very end of the current one when playing randomly. Also, reserving larger amount of memory let the player deals with several tracks in once, resulting in less multiple hard disk accesses so it (may) reduce interferences.
BUT, this is a very theoretical point of view, because we need to consider that :
1) Dealing with digital playback audio stream nowadays is kind of a piece of cake for processors from desktop & notebook, even with a need of up-sampling. So unless your listening your music on a production server whose procs are fully loaded, most of time, your will be able to do that pretty much on the fly. This argument is of course balanced by the fact that doing very high quality up-sampling is not that simple, and can lead to cpu cycle starvation.
2) Reserving a large amount of memory means your processor has reached its limits for the audio processing you'd like to do. So I guess, if you have an average hardware configuration, that you also don't have that much memory to give to A+. If you don't let enough memory for other processes to run (especially the hungry ones like a good-old safari opened with a tons of javascript running on and some not-as-good-old flash plugins opened), your system will start swapping (yes it's not an exclusive Ms Windows feature), meaning it will get the extra memory from...the hard disk, which will be the worst thing you could expect if your system has been set up to do the swap on one mechanical hard disk.
3) Talking about interferences, if you do have a SSD disk, you probably won't get into that problem.
So actually it makes a little sens to reserve larger amount of memory, but if you're reading your audio files from a SSD, without any up-sampling on a decent computer you wouldn't need it at all. However, if you're having some trouble while reading your music like latency from playing one track to another, you can try to increase the shared memory of A+, but just don't forget to let some for the others