bright@dataio.Data-IO.COM (Walter Bright) (12/24/86)
Many people have requested info on how to minimize the memory usage by the parent program when a spawn() is executed. The trick is to include the line (for S and P memory models): int _okbigbuf = 0; somewhere in the source to the program. This will prevent the startup code from allocating a full 64k of data to the program, and will cause the file I/O routines to use only small disk buffers. The effect is to trade off fast disk I/O for reduced memory usage. This is documented in section 6.2 of the manual.
scott@tg.UUCP (02/24/87)
Now that everyone is Flaming C compilers, how about some feedback on
the Datalight C or their Optimal-C. This question comes after reading
February's "Computer Language" magazine and seeing their challenge to
MS's C 4.0.
Any Comments?
Scott Barman
{philabs, pyrnj}!tg!scottdragheb@isis.UUCP (02/26/87)
In article <138@tg.UUCP> scott@tg.UUCP (Scott Barman) writes: >Now that everyone is Flaming C compilers, how about some feedback on >the Datalight C or their Optimal-C. This question comes after reading >February's "Computer Language" magazine and seeing their challenge to >MS's C 4.0. > >Any Comments? To say the least: Walter Bright and Roy Sherril (sp?) (the authors of Datalight, I think, right?) have guts. that is one heck of a challange! I have worked with Datalight C (as an end user) and it is the best one i have used at that price range. I do have one compliant: their libraries could be better (I have the developer's kit (with the source code to the libs), so it is easy to make modifications to it). I really hope that Optimal-C stands up to the test.... -- Do it in C. If you can't do it in C, do it in Assembly. If you can't do it in Assembly, it is not worth doing!! Darius Ragheb isis!dragheb | dragheb@isis.cs.du.edu
dsd@hpsadla.UUCP (02/27/87)
I use Datalight-C on my IBM compatible, 4.77 MHZ, 20Meg HD, V20. At work I use a UN*X System and do quite a bit of programming in C. The dlc routine that came with Datalight-C is pretty much the same as the cc in UN*X. The make utility (that's included) is pretty much the same also. The compiler is fast, and seems to generate tight code. I've got no complaints. In December I got to play with MSC4.0 on an AT, 30Meg HD, 6MHZ? Datalight-C seemed to compiled the version of word-count (UN*X utility wc) that I wrote, faster on my system (above) than MSC on the AT. Sorry I have no real speed comparison numbers for the above test. I really like the Datalight-C compiler. I had no problem installing it, and have had good experiences using it. Sorry no flames... Don St. Denis
kent@ncoast.UUCP (Kent Williams) (05/01/87)
For those of you who are members of the 'Compiler of the Month Club' for PC's, here are the results of the dhrystone tests for a few compilers. I just got Datalight's Optimum C - hence the tests. Test Conditions Aztec C86 3.4B C86 Microsoft C 4.0 Datalight Optimum C 3.05 All compilers set for small data, small code, .exe format, no 'EXEPACK.' Maximum optimization selected for all compilers. Testing done on an IBM AT (bad old model) with 8 MHZ clock. Benchmark Results : aztec 500000 p 308 t 1623 dps elapsed time = 315.65681 microsoft 500000 p 283 t 1766 dps elapsed time = 292.313475 datalight 500000 p 242 t 2066 dps elapsed time = 258.094893 File sizes -rw--a 6128 May 01 07:59 aztdry.exe -rw--a 15232 May 01 07:59 dlcdry.exe -rw--a 9432 May 01 07:56 mscdry.exe Nice Work Mr. Bright. 'You know that the hypnotized never lie' Kent Williams ...!cwruecmp!ncoast