Bakul Shah <firstname.lastname@example.org> wrote:
| Rob Warnock wrote:
| > Bakul Shah <email@example.com> wrote:
| > You don't want to have to force conses to 16-byte boundaries,
| > since on a 32-bit machine that would *double* the space taken
| > by them [though on 64-bit it's a no-brainer],
| I am lost. What does consing have to do with loops? I am only
| talking about aligning loops (as part of some compiled code).
And I was talking about needing to *maintain* that alignment
when a copying GC copies such an aligned code block.
But as Thomas Burdick said in a parallel reply, I may have
been trying too hard to save a few bytes when I suggested
partitioning the heap. In fact, when copying a code object,
if you simply allow the GC to waste 8 bytes if it needs to
for maintaining alignment, that should be "good enough".
Rob Warnock <firstname.lastname@example.org>
627 26th Avenue <URL:http://rpw3.org/>
San Mateo, CA 94403 (650)572-2607