FS: MK2a Memory Upgrade chips

The empeg design uses PIO (not DMA) for disk data transfers. Doing a lot of consecutive PIO prevents other stuff from running at the same time (there's only one CPU..).

When the player does readahead, it tries to fill nearly all available memory with data before stopping. The more memory. The more available memory, the longer this takes to complete. While the disk reads are happening, the display updates get stalled, and can appear to stutter.

Adding more RAM helps with large databases, extra apps, etc.. but the player software is pretty clueless about it all, and will just try to hog the CPU while filling up the extra RAM with more tunes..

Cheers

Hi,

Is the RAM configured as cache, or an elastic buffer with dynamic pointer(s)?

I use a lot of .WAV files which certainly exceed the buffer size.

I have noticed that if I am listening to a selection and change to another selection it does cause a short delay. The same happens with some long selections. There is always the readahead after the spin-up of the drive. I realize that most of this time is disk spin-up and I will try to temporarilly disable spin-down to verify it.

My questions are:

1) With additional RAM, if I change selections, will it flush the cache, or is it elastic and re-point to a new index and start reading there?

2) Will more RAM worsen the delay with a deeper buffer and cache flush?

3) Will more RAM reduce the readahead delay (read the complete music file) or spend more time filling the buffer?

4) Does the laptop disk drive cache come into play or is it a hinderance with the Empeg?

Thanks,

Ross