Currently we (Michael Yaroslavtsev and myself, drop q's and x's before mailing) do not attempt (yet) to emulate special and I/O registers due to the lack of documentation; so we have no way to run the O/S kernel and have to limit our goals to emulation of operating environment that the O/S provides to the user tasks for now.
2/19/2008. The development continues on SourceForge.
2/1/2008. ALGOL-BESM6 works.
3/19/2006. The long-standing bug in the implementation of the division instruction has been fixed.
2/2001. We are getting ready for a major release that has quite a few new features, including some supervisor mode functionality for system calls and working compilers, namely Pascal and Algol-60; and all 3 Fortran compilers capable of compiling Whetstone.
12/98 - 1/99. The work has started to make the supervisor mode functional.
4/29/97. The tape with O.S. sources which I brought from Moscow, Russia where I spent my vacation, has been read despite its low quality thanks to Peter Zaitcev, and some kernel modules have been successfully compiled.
4/22/97. The sources of the emulator, some test programs and the images of the O.S. disk (#2053) and the programming system disk (#2048) are available. This is not a public release, i.e. no README or INSTALL files. Caveat emptor.
4/1/97 - 4/17/97. The tape and the BESM-6 manuals were detained by U.S. Customs for Intellectual Property Rights review.
2/24/97. Most compilers (Algol, Fortran, Assembler; but not Pascal, unfortunately) work now.
2/14/97. One (out of 3 existing in the system) FORTRAN compiler (incidentally, the newest and with highest level of optimisation) works, i.e. produces an object module which then gets linked and executed correctly. The reasons of failure of other compilers are under investigation.
2/13/97. The job control system "Dubna" starts (prints a banner), but fails later.
2/7/97. The standard ALU test "almost works" (except the division instruction whose results, obtained by using the native IEEE f.p. division, sometimes differ from expected ones by one least significant bit). An attempt to simulate the BESM-6 non-restoring division algorithm yielded similar results.
Several interactive programs work, i.e. accept commands and execute the simplest ones. The main obstacle now are system calls that BESM-6 OS had hundreds, and kernel memory structure (some programs read kernel variables directly).
2/2/97. The first interactive program (sysadm services) worked.
1/29/97. Thanks to Sergey V. Kartashoff, the 800 bpi 9-track tapes with the operating system and compilers have been read successfully, and the development of the BESM-6 emulator continues.
Early 90's. The emulation of BESM-6 became conceivable. Serge Vakulenko wrote the first version of the emulator.
The BESM-6 pages have been accessed 36636 times.