waynet@kit.CNA.TEK.COM (Wayne Turner) (01/23/91)
In article <1991Jan21.233029.29058@murdoch.acc.Virginia.EDU> randall@Virginia.EDU (Ran Atkinson) writes: > >In article <234@nazgul.UUCP> bright@nazgul.UUCP (Walter Bright) writes: > >(in response to a query about C++ compilers) > >>Zortech's environment is oriented towards large multi-file projects. > >Borland's IDE with TC++ also handles large multi-file projects just fine. >No meaningful difference here. > We are currently migrating our development of a commercial package from TC++ to ZTC++ because (among other reasons) the TC++ *command-line* version cannot handle even *medium* sized projects. The compiler runs out of memory when modules of 200-300 lines are compiled (and > 500 lines of header files are included). This occurs even when 4 Mb of extended memory is present and the appropriate compiler switch is used. The TC++ linker runs out of memory when linking 40-50 OBJS, there is 520K of base memory available and, again, extended memory is present. When the compiler is near to running out of memory it presents meaningless, often garbled error messages to the user. These messages go away when more base memory is freed up (with no mods to the target file required). > >Zortech C++ does seem to compile a bit more quickly than Turbo C++, >but the difference hasn't been significant in my experience. The above mentioned project compiles in 15 min. under TC++ and in 16 min. under ZTC++ (with compile switches as similar as possible). >I believe that Zortech also has C++ for UNIX System V/386 and am surprised >that Walter failed to mention it himself; he is usually fairly aggressive >at promoting his compiler on the net. :-) > >>P.S. I work for Zortech. > My only connection with Zortech is as a customer. Note that I omitted the word "satisfied" pending future results ;-). Wayne Turner Tektronix, Inc. Redmond, Oregon waynet@kit.CNA.TEK.COM
mnrausch@informatik.uni-erlangen.de (Martin Rausche (AKBP WS 1990)) (01/23/91)
jdb@reef.cis.ufl.edu (Brian K. W. Hook) writes: >One note that I would like to make about the Turbo C++ TC compiler is that it >bogs down on my source code every 3500 lines or so. And it is NOT due to >a disk hit. I have a 386-33 with 8MB of memory. 4MB are allocated as a disk >cache using Norton Cache (NCACHE-F EXT=4096) adn 2MB are left as extended >memory for TC itself. I have some stuff loaded in high memory, and this leaves >me 583K free base memory, 2MB EXT, and 4MB of disk cache. However, as I was >saying, on many of my project files the compiler will fly along at Mach 10 >and then it will just hit the brakes. I believe that this is a deficiency in >the compiler, since the cache is large enough to hold ALL my .LIB files, .H >files, and source files. >And this occurs after I've been doing some compiling, so I know everything is >in the cache. Also, when I compile, I watch my hard drive light. It >doesn't even flicker. Not once. And the compiler still bogs down. I tried to compile fractint ver.14 with TC++. In the file fractals.c the compiler stops after about 3500 lines. What's wrong here ????? Thanks in advance for any advice. Martin. >Brian -- ------------------------------------------------------------------------------ | EMail: mnrausch@immd4.informatik.uni-erlangen.de | ------------------------------------------------------------------------------
alex@bilver.uucp (Alex Matulich) (01/25/91)
In article <1991Jan21.233029.29058@murdoch.acc.Virginia.EDU> randall@Virginia.EDU (Ran Atkinson) writes: >I do commercial development and based on the tests I did and >experience with both, I would suggest that optimisation is not >meaningfully different between the two compilers. Library quality and >user-support and platform-support are much more meaningful standards >for comparison. And could someone who knows please say something about the comprehensive- ness of either Zortech's or TC's libraries? I use the Lattice Compiler, but since they stopped developing the MSDOS version I have been half- heartedly looking about for another compiler. I didn't find TC's libraries to be nearly as comprehensive as Lattice's. What about Zortech? Also, Zortech claims 99% ANSI compliance. What, pray tell, is that remaining 1%? I have a sneaky fear that it may be something important to me.... Somebody shed some light please! -- _ |__ Alex Matulich (alex@bilver.UUCP) /(+__> Unicorn Research Corp, 4621 N Landmark Dr, Orlando, FL 32817 //| \ UUCP: ...uunet!tarpit!bilver!alex ///__) bitnet: IN%"bilver!alex@uunet.uu.net"
samuelb@kuling.UUCP (Samuel Bergenstr}hle) (01/29/91)
Borland claims that they will have protected mode versions of their compiler, linker etc soon. Also they will introduce a new concept of dealing with include files that sounds very appealing to me. I do not work for Borland.
draper@buster.cps.msu.edu (Patrick J Draper) (01/30/91)
In article <1891@kuling.UUCP> samuelb@kuling.UUCP (Samuel Bergenstr}hle) writes: > >Borland claims that they will have protected mode versions of their >compiler, linker etc soon. Also they will introduce a new concept of >dealing with include files that sounds very appealing to me. > >I do not work for Borland. Yes Yes! Tell us more about the include file thing, please. ------------------------------------------------------------------------ Patrick Draper In times like these it is helpful to cps.msu.edu remember that there have always draper@cps.msu.edu been times like these. ------------------------------------------------------------------------