garym@telesoft.UUCP (Gary Morris @telesoft.com) (12/07/89)
I wasn't able to respond immediately to the various comments in INFO-ADA on and around November 21st but I do feel some clarification is required. In "This is getting stupid", Dick Dunn, (ico!vail!rcd) writes: > Bill Wolfe (wtwolfe@hubcap.clemson.edu) writes: > > Doug Schmidt (schmidt@zola.ics.uci.edu) writes: > > > 1. Precisely which Ada compilers produce better code than > > > which C compilers? Not all compilers are created equal... > > The Telesoft TeleGen2 compiler, vs. all available C compilers, > > as of circa October 1988; contact Telesoft for details... > Oh, right, Telesoft is the unbiased judge of the universal superiority of > its own compiler! The facts concerning the benchmark, such as which compilers were used, the benchmark, and so on are available by contacting TeleSoft. You can make your own judgement once you have the facts. How does reporting the results of one benchmark to a users group get interpreted as a claim of "universal superiority"? The consumer is the one who must judge which compiler is right for their application using a variety of benchmarks that reflect the type of work their application will be doing. > > > 2. What are the benchmarks and where can they be obtained?... > > I seem to recall "Dhrystone", among others, being cited. There > > were no limitations on the use of the Ada language. > > You "seem to recall"??? Your grandiose boasting is based on something you > don't quite remember? Even if the one you think you remember is correct, > it's a tiny synthetic benchmark with well-known problems (acknowledged by > the author) and it is NOT suitable for comparisons of compilers. It was dhrystone and it is widely used in spite of its minor flaws. One of the reasons for using it was that numerous C compiler have been benchmarked with it and the results have been published and are readily available for comparison. > But soft! 'Tis fair Amnesia come to stay a while. Isn't it convenient how > little information there is here? It's probably a comparison against pcc. I don't think it's fair to blast Bill for not remembering every detail or to accuse him of "conveniently" leaving some out. A little research, such as a call to TeleSoft, could have answered a lot of your questions. Among the C compilers chosen were those that produced the best times on 16.77 Mhz 68020 based machines out of a list of published dhrystone results for 300 compilers. The TeleSoft optimizing compiler was compared against four other C compilers: Sun/Unix 3.2 C on a Sun 3/160 GreenHills C on ISI optimum V machine Gnu C on a Sun 3/160 MASSCOMP C on RTU 3.1, NC-5700 machine MASSCOMP had the best C compiler, our Ada compiler only beat it by a very small margin. > Bill told us that, thanks to sophisticated compiler technology, Ada > compilers can now give us better object code than C compilers. But the > claim apparently applies only to one Ada compiler; the claim is made by > the vendor of that compiler; the claim is reported at a meeting for users > of the compiler; the benchmarks and hardware platforms used for the com- > parison are conveniently forgotten. > > Pure vapor. Vapor? Bill may not have the details of the benchmarks and platforms handy, but that doesn't make them nonexistant. I've worked on that optimizing code generator and even I had to dig a bit to find the original report with all the details of that comparison. The whole point of the comparison was not to prove that Ada is better than C but that it's possible for Ada compilers to generate code that is just as good or better than C compilers. It takes a lot of time and effort to go from a compiler that meets the minimum language requirements to a highly optimizing compiler. Ada compilers will continue to improve and you can expect to see more optimizing compilers from a variety of sources. --GaryM Gary Morris {uunet,ucsd}!telesoft!garym TeleSoft N6FRT garym@telesoft.com San Diego, CA, USA ASEL garym@nynashamn.telelogic.se +1 619 457 2700 "If something is worth doing, it's worth doing right."