From mboxrd@z Thu Jan 1 00:00:00 1970 X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on polar.synack.me X-Spam-Level: X-Spam-Status: No, score=-0.9 required=5.0 tests=BAYES_00,FREEMAIL_FROM, FREEMAIL_REPLY autolearn=no autolearn_force=no version=3.4.4 X-Google-Language: ENGLISH,ASCII X-Google-Thread: 103376,7bcba1db9ed24fa7 X-Google-Attributes: gid103376,public X-Google-ArrivalTime: 2001-07-09 06:24:02 PST Path: archiver1.google.com!newsfeed.google.com!newsfeed.stanford.edu!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!upp1.onvoy!msc1.onvoy!onvoy.com!tethys.csu.net!nntp.csufresno.edu!sn-xit-01!sn-post-02!sn-post-01!supernews.com!corp.supernews.com!not-for-mail From: Vinzent Hoefler Newsgroups: comp.lang.ada Subject: RE: is ada dead? Date: Mon, 9 Jul 2001 09:23:59 -0400 Organization: Posted via Supernews, http://www.supernews.com Message-ID: <3B635921@MailAndNews.com> X-InterChange-Posted-By: vinzent@MailAndNews.com Sender: Vinzent Hoefler X-EXP32-SerialNo: 50000000 Mime-Version: 1.0 Content-Type: text/plain; charset="ISO-8859-1" Content-Transfer-Encoding: quoted-printable X-Newsreader: InterChange (Hydra) News v3.61.08 X-Complaints-To: newsabuse@supernews.com Xref: archiver1.google.com comp.lang.ada:9671 Date: 2001-07-09T09:23:59-04:00 List-Id: Original Message From "Jean-Pierre Rosen" >"Vinzent Hoefler" a =E9crit dans le message news: 3B5F53D1@MailAndNews.com... >> future). I'm still doing ancient, but f* fast assembly code and to tell you >> about some success with that "consider it to be dead, too"-language, I got a >> 50-60 _times_ (5,000%) speed improvement last week compared to some old >> code. >> >Well... Recently, I improved some code from one of my clients 50 times too - by >replacing convoluted Ada code by better Ada code. That's cool. And probably you did a much better job then. >Not to diminish the value of what you are doing - Oh no, writing it wasn't that hard. The hard work was still done with pencil and paper to figure out the algorithm. The fact is that Assembly just does a much better job in simple bit shuffling with floating point variables than a dumb compiler that always decides to use floating point instructions with such ones[*] and does not even really know of 64-bit mantissas or of the existance of quite strange 24-bit floating point formats at all. Not to mention the little/big endian problem. Expressing those in HLL would have become more unreadable than the less than 40 lines of well commented assembly that probably no-one is ever looking at again once it is working. [*]Bad side effect of strong typing, I guess. ;-))) >just another example that you can get tremendous speed-ups by doing the right >thing, irrespective of language. Yes. First improve the algorithm. That's what I really did in the end. Just chose a better suited language. :-) Vin"Writing too much, should work on my device driver."zent.