utoddl@uncecs.edu (Todd M. Lewis) (03/06/90)
I've been working with Manx C v5.0a lately and I'm wondering why small programs tend to be larger than they are when compiled with the version 3.6a compiler. My first thought was 32-bit ints, but going with 16-bit ints didn't help much. No floating point. With or without prototypes doesn't seem to matter. Various combinations of optimizations can help some, but I still can't seem to get my small programs back down to size. "Small" means <10k. Case in point: one program comes out to 8488 bytes under 3.6a. The best I've been able to do with v5.0a is about 10730 bytes. I'm guessing the difference has to be in the library if the claims that the generated code is smaller for v5.0a are true. This program does NO file I/O, it's all intuition messages and IDCMP stuff. I wouldn't be too worried about it for large projects because the smaller code reaches a break-even point somewhere in the 10k-20k range, but I've got lots of LITTLE programs that are gonna grow by 10%-40%, and I'd like to keep 'em as small as possible. So, what's the deal here? Is there some way to get these small programs back down to size? _____ | Todd M. Lewis Disclaimer: If you want my employer's ||\/| utoddl@ecsvax.uncecs.edu ideas, you'll have to || || utoddl@ecsvax.bitnet, @unc.bitnet _buy_ them. | || utoddl@next1.mscre.unc.edu |___ ("Prgrms wtht cmmnts r lk sntncs wtht vwls." --TML)