[RC5] LIFO? Is this true?
David_Taylor at msn.com
Tue May 12 23:05:37 EDT 1998
> > Ryan Anderson wrote:
> > > Wouldn't it be smart to have a separate buff-out for each
> > > block?
> > No! I sure don't want 1000 separate files in any directory.
> Okay, for each active block? (Store non worked upon blocks in the
> standard buff-in, as each block is worked on, break it out into a separate
> file, when it's finished drop it into a buff-out. This gives you much
> faster block turn around time, and lets each local collection of computers
> work in much the same fashion as the overall project..)
Well, definitely not for every block - 1000 (ok 1024 to be exact, but
anyway) blocks each in
in their own file would take up, with a FAT system (FAT16 would usually have
32Kb clusters with decent sized drives) (FAT32 would have 4Kb - 8Kb with
4Kb [What I now have] - 4MB
8Kb - 8MB
16Kb - 16MB
32Kb [What I had til recently] - 32MB!!!!!!!!!!
And, as for each active block, umm, what's the point? Infact what's the
point for 1 for every block? We already have checkpoint files for blocks
that are affected by a crash, and read by the specific client that was
working on it, Blocks which are saved in-progress back into the buff-in are
(I *think*) ignored (or just processed) by other clients..... I cant see how
it would do anything except increase disk usage, and make it utterly
impossible to manually move buff files from comp. - comp.
Just my opinion.. if I've missed anything..... Basically, I don't care :o)
To unsubscribe, send 'unsubscribe rc5' to majordomo at lists.distributed.net
rc5-digest subscribers replace rc5 with rc5-digest
More information about the rc5