atschnei@vlsi.waterloo.edu (Todd Schneider) (05/23/91)
I've seen ads in magazines and some articles on the net that indicate Zortech C++ for the Mac is available. I wonder if someone who has used it can answer some questions. 1. How much does it cost and what does the package include? 2. What's it like (a comparision to Think C would be useful)? 3. Can you compile C source files too? 4. Is it System 7.0 compatible? Any other comments you want to make would also be appreciated. Thanks, Todd -- -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- + University of Waterloo | atschnei@vlsi.waterloo.edu + + Department of Electrical Engineering | atschnei@vlsi.uwaterloo.ca + + VLSI Group | Office:(519)885-1211 x3896 +
anders@verity.com (Anders Wallgren) (05/24/91)
In article <1991May23.150712.13602@vlsi.waterloo.edu>, atschnei@vlsi (Todd Schneider) writes: > >I've seen ads in magazines and some articles on the net that >indicate Zortech C++ for the Mac is available. I wonder if >someone who has used it can answer some questions. > >1. How much does it cost and what does the package include? I think it was around $300 >2. What's it like (a comparision to Think C would be useful)? It seems relatively fast, although I had some problems (read my other message in this group). >3. Can you compile C source files too? yes, there's an option that will allow it to do this. >4. Is it System 7.0 compatible? > Seems to be, but again, I haven't really stress tested it. anders
sw@nan.co.uk (Sak Wathanasin) (05/25/91)
In article <1991May23.205933.13024@verity.com>, anders@verity.com (Anders Wallgren) writes: > >1. How much does it cost and what does the package include? > > I think it was around $300 Good grief! I paid 250 pds sterling, which is roughly $430. Now, we're pretty used to being ripped off by UK distributors of US software, but when the parent company is in the UK,... > >2. What's it like (a comparision to Think C would be useful)? > > It seems relatively fast, although I had some problems (read my other > message in this group). Come, come. As you said, it's about the same speed as Apple's CFront/C with load/save. Think C is much faster: on my fx, Think C compiled, linked and launched the "Art Class" demo program (starting from an empty project) in just over 5 mins - nearly 24K lines including headers. Since the hdr files were included several times, TC came up with a line count of about 75K or about 15K lines/min. It took about 3.5 mins for ZTC to build a 2K line MacApp program. The linker, rez etc took nearly 1 min to rebuild the appl from existing .o files, so ZTC took 2.5 mins to compile the 2K lines. The MacApp hdrs total about 10K lines. There were only 2 .cp files, so even if all the MacApp hdr files were pulled in, ZTC would have "seen" 22K lines at most; add some for system hdrs, and we'll call it 25K lines. That's about 10K lines/min *at best*, and I think I'm erring in ZTC's favour. Of course, ZTC is a full C++ compiler, and Think C is not. Sak Wathanasin Network Analysis Limited uucp: ...!ukc!nan!sw other: sw@network-analysis-ltd.co.uk phone: (+44) 203 419996 snail: 178 Wainbody Ave South, Coventry CV3 6BX, UK
anders@verity.com (Anders Wallgren) (05/25/91)
In article <0101000D.ytekob@nan.co.uk>, sw@nan (Sak Wathanasin) writes: >> >2. What's it like (a comparision to Think C would be useful)? >> >> It seems relatively fast, although I had some problems (read my other >> message in this group). > >Come, come. As you said, it's about the same speed as Apple's CFront/C with >load/save. Think C is much faster: on my fx, Think C compiled, linked and >launched the "Art Class" demo program (starting from an empty project) in just >over 5 mins - nearly 24K lines including headers. Since the hdr files were >included several times, TC came up with a line count of about 75K or about 15K >lines/min. > >It took about 3.5 mins for ZTC to build a 2K line MacApp program. The linker, >rez etc took nearly 1 min to rebuild the appl from existing .o files, so ZTC >took 2.5 mins to compile the 2K lines. The MacApp hdrs total about 10K lines. >There were only 2 .cp files, so even if all the MacApp hdr files were pulled >in, ZTC would have "seen" 22K lines at most; add some for system hdrs, and >we'll call it 25K lines. That's about 10K lines/min *at best*, and I think I'm >erring in ZTC's favour. > >Of course, ZTC is a full C++ compiler, and Think C is not. > Indeed - methinks you are comparing apples to oranges... Besides, compiling one 25K line program does not a test make as far as my needs are concerned. anders
Lawson.English@p88.f15.n300.z1.fidonet.org (Lawson English) (05/25/91)
Sak Wathanasin writes in a message to All
SW> That's about 10K lines/min *at best*, and I think I'm erring
SW> in ZTC's favour.
SW> Of course, ZTC is a full C++ compiler, and Think C is not.
I suspect that the main slow-down is in disk I/O. Try this: set up a humongous
RAM disk (about 1.5 megs), drop your ThC project folder into it and do a complete
compile from scratch. Now do the equivalent for Zortech C++. The time difference
should drop dramatically, as Think C pre-loads as much as it can into the project
file, and keeps it handy in memory (if MY memory serves me correctly).
--
Uucp: ...{gatech,ames,rutgers}!ncar!asuvax!stjhmc!300!15.88!Lawson.English
Internet: Lawson.English@p88.f15.n300.z1.fidonet.orgdrs@bnlux1.bnl.gov (David R. Stampf) (05/26/91)
I recieved an add from Zortec about four weeks ago, and ordered it a week later. They said it would ship in a few days, but after waiting a while I called them up. They said that it had not shipped yet - maybe next week. All that I know is that it is an MPW based compiler. The deal they offered was to buy the compiler and get MPW at a small additional charge - $50 if my memory serves. When I get it and try it out I'll post more. < dave
nagle@well.sf.ca.us (John Nagle) (05/27/91)
Are these timings with or without the Zortech global optimizer turned on?
It makes a big difference.
I'm actually rather impressed with the compiler. The hard optimizations
are all there and working. But the present release, V2.1R1, has enough
defects that it's like using a beta version of a new product. Zortech
is good about fixing reported defects, so if you produce a simple test case
for a defect and send it to Zortech, preferably on their BBS, they usually
fix it quickly and send you a compiler update.
So keep those bug reports going out.
And avoid forms like
a = b = c;
in the current release.
John Naglesw@nan.co.uk (Sak Wathanasin) (05/27/91)
In article <12299.283FF27C@stjhmc.fidonet.org>, Lawson.English@p88.f15.n300.z1.fidonet.org (Lawson English) writes: > I suspect that the main slow-down is in disk I/O. Try this:etc.. I'm sorry if I've started a Think vs Zortech war, and I hope we can squash it before it goes any further. As Anders Wallgren pointed out, the timings are irrelevant if you need to use MacApp with C (for whatever reason): CPlus/Zortech are the only choices available now. It's just that the original enquirer asked for a comparison with Think C. Sak Wathanasin Network Analysis Limited uucp: ...!ukc!nan!sw other: sw@network-analysis-ltd.co.uk phone: (+44) 203 419996 snail: 178 Wainbody Ave South, Coventry CV3 6BX, UK
Lawson.English@p88.f15.n300.z1.fidonet.org (Lawson English) (05/29/91)
Sak Wathanasin writes in a message to All
SW> I'm sorry if I've started a Think vs Zortech war, and I hope
SW> we can squash it before it goes any further
Wasn't meant to be a Think C vs anybody war, just an observation (speculation)
about why Think C might be compiling faster.
Lawson
--
Uucp: ...{gatech,ames,rutgers}!ncar!asuvax!stjhmc!300!15.88!Lawson.English
Internet: Lawson.English@p88.f15.n300.z1.fidonet.org