Thanks Alex, About why I am trying it - it's primarily just a learning exercise for me. What I am trying is expanding the cell to 4 words instead of two and use the extra words as the meta data and not require tagged pointers. This way I could even build it with TCC that does not seem to align functions at 4 byte boundary.
Sorry if I am missing something, but wouldn't x-1 result in shifting the pointer by 8 bytes instead of 4 (since size of cell is 8). Which is probably why cdr is used instead of car? regards, Kashyap On Wed, Aug 26, 2020 at 10:56 PM Alexander Burger <a...@software-lab.de> wrote: > Hi Kashyap, > > > I am still working on the idea of using non-tagged pointers in > miniPicoLisp > > - in order to get rid of the requirement that function pointers be word > > aligned. > > Why would you want to do that? The concept of encoding the tags in the > lowest > bits of a pointer is the core idea of PicoLisp. Pointer alignment to 8 > bytes (on > 32 bits) gives three tag bits for free. > > Three tag bits are needed (gc mark, number and symbol tag). If you use > other > bits (e.g. the highest ones) you lose 87.5 percent of the address space. > And > runtime checks for higher bits are more expensive. Similar for other > concepts > like using separate heaps. > > The alignment comes at low cost, only once when the heap is allocated. > After > that all cells are guaranteed to be properly aligned. Most hardware has > that > requirement anyway. If not, word accesses become intolerably slow. > > > > > #define tail(x) (((x)-1)->cdr) > > > > Does this not result in going outside the 0th cell? Would this be more > > appropriate? > > No, because 'x' points to an offset of four (because of the tag bit!) to > the > symbol > > Symbol > | > V > +-----+-----+ > | | | VAL | > +--+--+-----+ > | tail > | > V name > > so that 'tail' points to the start address of that cell. > > ☺/ A!ex > > -- > UNSUBSCRIBE: mailto:picolisp@software-lab.de?subject=Unsubscribe >
