cwr@pnet01.cts.com (Will Rose) (09/10/90)
ghelmer@dsuvax.uucp (Guy Helmer) writes: >In <4284@crash.cts.com> cwr@pnet01.cts.com (Will Rose) writes: stuff omitted here concerning sc 6.8 compiling... >Wow. I just generated a binary on my VAX using gcc, and it was 117K >of text, 21K of data, and 11.5K BSS. Unstripped, its binary is 158K >on the disk. The 3b1 must have an unsual processor or a less-than-optimal >compiler. Dumb looks still free... I must *stop* programming until 0500 - the coffee makes *everything* seem possible, even easy, and fingers float across the keyboard translating thoughts before they are even recognised. Only in the clear light of dawn (well, mid-afternoon) does the horrible truth emerge... Yes, I blew it, forgot to strip the symbol tables. Try these numbers for a PC7300/3b1 running SysV3.50 and compiling sc 6.8 with the standard compiler, 'based on the AT&T portable C compiler': 155,900 bytes of text, 52,640 bytes of data, 18,768 bytes of bss for a total of 227,308 bytes. The actual file is 210,366 bytes. Still won't fit in 64K+64K without unusual cunning; I hope you make it, though, because it would be very nice to have. By the way, the current size record on the 7300 seems to be Kyoto lisp, with a 1.4 *Megabyte* binary. Not knowing anything about lisp, I can't imagine what they put in a file that size - some of the whales they will keep catching perhaps? Thanks for the warning - Will. PS. I've sent you a message and a new man4 - your current version is doubtful. ----------------------------------------------------------------------- "If heaven too had passions | Will Rose even heaven would | UUCP: {nosc ucsd hplabs!hp-sdd}!crash!pnet01!cw grow old." - Li Ho. | ARPA: crash!pnet01!cwr@nosc.mil | INET: cwr@pnet01.cts.com UUCP: {nosc ucsd hplabs!hp-sdd}!crash!pnet01!cwr ARPA: crash!pnet01!cwr@nosc.mil INET: cwr@pnet01.cts.com
jds@mimsy.umd.edu (James da Silva) (09/10/90)
In article <4338@crash.cts.com> cwr@pnet01.cts.com (Will Rose) writes: >155,900 bytes of text, 52,640 bytes of data, 18,768 bytes of bss for a total >of 227,308 bytes. The actual file is 210,366 bytes. > >Still won't fit in 64K+64K without unusual cunning; I hope you make it, >though, because it would be very nice to have. Remember, these sizes are for 32 bit instruction sets. Small model 8086 code for the same program is going to be a *lot* smaller. If the data space contains many pointers, it's going to shrink quite a bit as well. While sc is a *bit* large, it's fair game for a try. >By the way, the current >size record on the 7300 seems to be Kyoto lisp, with a 1.4 *Megabyte* >binary. Not knowing anything about lisp, I can't imagine what they put >in a file that size - some of the whales they will keep catching perhaps? That size includes the compiler, interpreter, dynamic loader, and entire runtime library for a large, rich programming language with lots of library routines. The size isn't too bad, considering. Jaime ........................................................................... : domain: jds@cs.umd.edu James da Silva : path: uunet!mimsy!jds Systems Design & Analysis Group