Bob Supnik bob at supnik.org:
> Bolting on vectors to the existing 3900 would not be all that hard, > because the core floating point routines are available, but it would be > another "never existed in the real world" hybrid. That's already been > done once, when the 3900's memory capacity was extended to 512MB Not to mention the ungodly 32-processor MicroVAX 3900. ;-)A monster in the dark with glowing red eyes. * * * Curiously enough though, there were rumors about the existence of a 64-processor VAX. Some of these relate to VAX M31 (codenamed Andromeda) http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=4859 http://ieeexplore.ieee.org/search/searchresult.jsp?queryText=vax%20m31&newsearch=true which was supposed to be expandable up to 64 processors, but as net rumors say, has actually grown up to 32-way machine only. I vaguely recall however seeing somewhere (at comp.os.vms?) a comment by Steve Lionel about him working with parallel algorithms on a 64-processor VAX SMP machine (running custom OS), and VMS people coming down from the East coast to test the stability of VMS on it, whatever this means (e.g. perhaps partitioned down to 32 processors, or running custom kernel able to handle 64 CPUs ?). My memory of this comment are dim and I may well mis-remember the particulars, however VAX/VMS 7.3 does actually run fine on 32 processors (scalability issues aside), except for the bug in the MONITOR utility making it able to pick up only the first 31 processors in its screens. http://oboguev.net/vax_mp/screen-39320.html
_______________________________________________ Simh mailing list [email protected] http://mailman.trailing-edge.com/mailman/listinfo/simh
