From: MagerValp (MagerValp_at_cling.gu.se)
Date: 2002-11-16 14:02:59
Yes, Ullrich, I do agree with what you're saying (and that /. article
was interesting but didn't say anything that hadn't been said before),
but I still think that most people will be better off with cbm.h. The
disk drive will always be the bottleneck when you're doing I/O, not
the code. But there are still a few reasons why stdio code will be
quite a bit slower than the kernal wrappers, and a big one is the lack
of a load command. If you need to load a resource into memory, the
kernal LOAD call is quite a bit faster than OPEN/CHKIN/GETIN/CLOSE,
especially when you take fastloaders into account: that 10-30x speedup
is only for the LOAD call. And even if you use stdio, you still have
to remember to treat the the c64 as a c64. Any call to fseek will have
the system grind to a halt while the drive chews through that linked
sector list.
But don't misinterpret me, I think it'd be a great thing to have stdio
support for the CBM machines, I'm just saying that most apps don't
actually need it -- its absence is not an excuse to not use cc65.
Anyway, my train leaves in 21 minutes, but I'll gladly flog this dead
horse a bit more when I come back :)
--
___ . . . . . + . . o
_|___|_ + . + . + . Per Olofsson, arkadspelare
o-o . . . o + MagerValp_at_cling.gu.se
- + + . http://www.cling.gu.se/~cl3polof/
----------------------------------------------------------------------
To unsubscribe from the list send mail to majordomo_at_musoftware.de with
the string "unsubscribe cc65" in the body(!) of the mail.
This archive was generated by hypermail 2.1.3 : 2002-11-16 14:03:38 CET