Things I've done with GNU software
This is a partial list of things I've done with or contributed to GNU software from 1987-1992. Upon reflection, it looks like I didn't have much of a life ;-)
- June 1987: Downloaded GCC 1.0
- June 1987: Ported GCC to NS 32032; 20% better performance than National Semiconductor's proprietary compiler
- July 1987: Enhanced NS 32032 port--now 40% better performance than National Semiconductor's proprietary compiler
- August 1987: Announced plans to write GNU C++ compiler--world's first native-code C++ compiler
- August 1987: Ported GCC to Sun3 FPA (a floating point accelerator card based on a Weitek chipset); 10% better performance than Sun's compiler on a Los Alamos benchmark
- September 1987: Ported GCC and GDB to Intel 386
- October 1987: Ported GCC to Motorola 88000; 10% better performance than Motorola's proprietary compiler
- November 1987: Enhanced GDB with C++ features; first native-code C++ debugger
- December 1987: First release of GNU C++ (version 1.12)
- February 1988: Ported GCC to SPARC; 10% worse performance than Sun's compiler
- June 1988: GCC/G++ work gets me a ticket to graduate school at job at Stanford and a summer job in France (21 releases in 7 months!)
- July 1988: Ported GDB to SPARC (thanks to Sun France for donating the machine)
- August 1988: Profiled in French magazine "Science & Vie Junior" as "Le Chevalier Informatique"
- September 1988: GCC/G++ version 1.37--first really stable version
- April 1989: Wrote GNU Instruction Scheduler for CS 343 at Stanford; performance on SPARC catches up with Sun's compiler
- May 1989: Co-implemented GNU Branch Scheduler; performance on SPARC exceeds Sun's compiler
- June 1989: First "official" business plans for open source-based company
- July-October 1989: company remains anonymous as we search for a name not yet registered with the California Department of Corporations
- November 1989: Cygnus Support is founded
- February 1990: Cygnus signs first contract to provide commercial support for open source software!
- April 1990: GCC Version 2.0 development begins
- June 1990: Ported GCC to Philips VLIW processor
- July 1990: Ported GCC to i960--first embedded systems port
- August 1990: Implemented Exception Handling in GNU C++; presented results at Usenix C++ conference
- 1991: Lots of GCC/G++ merge work
- February 1992: GCC 2.0 released
- June 1992: Ported GCC to Sparc V9 (aka UltraSPARC)
By mid-1992, Cygnus had a GCC, a G++ team, and a GDB team. Aside from
the odd bugfix, I didn't do much development after that. I got to watch
as others became part of the most scalable development project I've ever seen:
- Q1 1992: P1 release, 2 native toolchains
- Q2 1992: P2 release, 5 native toolchains
- Q3 1992: P3 release, 15 host/target combinations
- Q4 1992: P4 release, 25 ...
- Q1 1993: Q1 release, 35 ...
- Q2 1993: Q2 release, 50 ...
- Q3 1993: Q3 release, 65 ...
- Q4 1993: Q4 release, 75 ...
- Q1 1994: R1 release, 85 ...
- Q2 1994: R2 release, 95 ...
- Q3 1994: R3 release, 110 ...
- Q4 1994: R4 release, 120 ...
- Q1 1995: 95q1 release, 125 ...
- Q2 1995: 95q2 release, 135 ...
- Q3 1995: 95q3 release, 145 ...
- Q4 1995: 95q4 release, 150+ ...
- Q1 1996: 96q1 release, 150+ ...
- Q2 1996: 96q2 release, 150+ ...
- Q3 1996: 96q3 release, 150+ ...
- Q4 1996: 96q4 release, 150+ ...
- Q1 1997: 97r1 release, 150+ ...
- Q3 1997: 97r2 release, 150+ ...
- Q2 1998: 98r1 release, 175+ ...
- Current release
No other company on the planet has software support for so many
configurations of anything.