Path: csiph.com!v102.xanadu-bbs.net!xanadu-bbs.net!news.glorb.com!npeer02.iad.highwinds-media.com!news.highwinds-media.com!feed-me.highwinds-media.com!border3.nntp.dca.giganews.com!border1.nntp.dca.giganews.com!nntp.giganews.com!news.iecc.com!nerds-end From: mev Newsgroups: comp.compilers Subject: Re: GCC is 25 years old today Date: Sun, 1 Apr 2012 06:20:20 -0700 (PDT) Organization: Compilers Central Lines: 29 Sender: news@iecc.com Approved: comp.compilers@iecc.com Message-ID: <12-04-003@comp.compilers> References: <12-03-051@comp.compilers> <12-03-053@comp.compilers> <12-03-062@comp.compilers> <12-03-074@comp.compilers> NNTP-Posting-Host: news.iecc.com X-Trace: leila.iecc.com 1333336453 76728 64.57.183.58 (2 Apr 2012 03:14:13 GMT) X-Complaints-To: abuse@iecc.com NNTP-Posting-Date: Mon, 2 Apr 2012 03:14:13 +0000 (UTC) Keywords: GCC, C, history Posted-Date: 01 Apr 2012 23:14:13 EDT X-submission-address: compilers@iecc.com X-moderator-address: compilers-request@iecc.com X-FAQ-and-archives: http://compilers.iecc.com X-Received-Bytes: 2470 Xref: csiph.com comp.compilers:546 > I had an internship at HP in the late 1980s. One of the people there > commented that, when gcc came out, it's code was (IIRC) 30% faster > than that of HPs compilers (for the 68k-based HP9000 computers running > HP/UX); a year later HPs compiler had caught up quite a bit. My guess > is that HP's compiler was PCC-based, and that optimization did not play > a big role in the Unix market until the arrival of RISCs and GCC. I worked on HP's 68K compilers from 1986 to 1992, particularly C and Fortran front ends. These compilers were indeed PCC based. We did optimization improvements including support for Dragon floating point accelerator in HP-UX 6.5 (1988), peephole optimizations, adding new inliner and some loop optimizations. We also updated front ends to support ANSI C and features from upcoming Fortran 90 standard. There was also some work to improve strcmp/strcpy as they played heavily into Dhrystone 1.0. I still have the HP-UX 6.5 T-shirt with release goals of >8 MIPS. The performance emphasis came more from competitive positioning against Sun and correlation with GCC introduction would be coincidental in my recall. PA-RISC, SPARC and MIPS were coming on the scene also with optimization emphasis. We were aware of GCC but more as an interesting curiosity than competitively. I also was at Denver C++ conference in October 1988 when Michael Tiemann presented what was first called GNU C++ but became G++. Mike Vermeulen