| TIP: Click on subject to list as thread! | ANSI |
| echo: | |
|---|---|
| to: | |
| from: | |
| date: | |
| subject: | math functions |
Hi, Paul. PE> PE> Does anyone have the ability and tenacity to implement some PE> PE> or all of the following functions marked "unimplemented" PE> PE> (except the last 3, as I can do them myself? I will post the PE> PE> descriptions of the functions in the next message, quoting PE> PE> from the ANSI C draft... PE> FM> Why don't you just look at the code generated from one of your PE> compilers PE> FM> when you compile for 'copro present', then duplicate that? PE> I thought that was a good idea at first, but then I realised that PE> these things, like sin(), are library functions, and will be coded PE> already. Are you sure? It might be worth checking to see if inline code is generated rather than a function call. If the numeric representation is compatible with what the copro uses, it might just be one or two instructions. I've just checked Turbo Pascal, it does use library calls unfortunately. And the whole thing is complicated by the possibility of emulation or using the copro, which is checked at run-time. I seem to recall when I looked at this once that it sets up every function as a call to the emulation code, then the first time each call is done, if there's a copro present, it patches the call with a NOP. Anyway, the code doesn't make much sense. PE> I could disassemble the library function of one of my PE> compilers, or even just look at the library source code which I PE> have for 2 of my compilers (at least), but that is basically PE> cheating, as I could be accused of copyright violation doing PE> something like that. It would be OK if I looked at the library PE> functions, then understood both the formulae and the copro PE> instructions, and then tackled the problem a month later, using PE> my knowledge/understanding. It's cleaner to get someone who already PE> has the understanding to do it from scratch though. Sure, and it's almost certainly been done. PE> FM> If C can't do asm instructions PE> The language does not include support for assembler. PE> FM> (Turbo Pascal can) PE> However, specific compilers do, including Turbo C. "The language" is as implemented by its compiler, IMHO. Of course then there is ISO C, a different animal. :-) PE> FM> you might have to make external objs with MASM or TASM. PE> This is what I want to do. Basically I was looking for a PE> matha.asm. BFN. Paul. Understand. If I ever come across one I'll let you know. Regards, FIM. * * If it's stupid and it works - its not stupid @EOT: ---* Origin: Pedants Inc. (3:711/934.24) SEEN-BY: 690/718 711/809 934 |
|
| SOURCE: echomail via fidonet.ozzmosis.com | |
Email questions or comments to sysop@ipingthereforeiam.com
All parts of this website painstakingly hand-crafted in the U.S.A.!
IPTIA BBS/MUD/Terminal/Game Server List, © 2025 IPTIA Consulting™.