Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!usc!rpi!think.com!sdd.hp.com!wuarchive!uunet!europa.asd.contel.com!noc.sura.net!haven.umd.edu!socrates.umd.edu!mike From: mike@socrates.umd.edu (mike santangelo (UNIX/VMS Sys Staff)) Newsgroups: comp.arch Subject: Re: HP-PA and CISC emulation (was Re: Will NeXT survive?) Message-ID: <1991May08.170019.23878@socrates.umd.edu> Date: 8 May 91 17:00:19 GMT References: <1991Apr29.144421.19819@oakhill.sps.mot.com> <1991May1.160128.1367@sono.uucp> <8283@uceng.UC.EDU> <7628@auspex.auspex.com> <8324@uceng.UC.EDU> Organization: University of Maryland, University College Lines: 61 rang@cs.wisc.edu (Anton Rang) writes: >In article <8324@uceng.UC.EDU> dmocsny@minerva.che.uc.edu (Daniel Mocsny) writes: >>(Does HP-PA do this right now? If so, I am very impressed. I would be >>much more impressed if it could also run the large existing libraries >>of CISC binaries at full speed, but that would be asking quite a bit >>:-) > I seem to recall that the high-end HP-PA machines run HP/3000 >binaries (under MPE) faster than the HP/3000 series itself ever did. >But I could be wrong. I don't know if this is done with a full >software emulator, or with a binary->binary translator, etc. > Anton > >+---------------------------+------------------+-------------+----------------+ >| Anton Rang (grad student) | rang@cs.wisc.edu | UW--Madison | "VMS Forever!" | >+---------------------------+------------------+-------------+----------------+ The HP-PA based HP3000 systems use a very sophisticated emulation system which makes use of something HP calls "millicode". These are tiny little HP-PA based subroutines that they used to emulate the old HP3000 "classic" design instructions, stack and register archiecture, etc. Beleive me, the "classic" HP3000 architecture was as different from HP PA as different can be. The fastest "classic" HP3000 system had 128KB of cache in the SPU and ran at a 75ns microinstruction clock. I am told that the 67ns Model 935 based HP-PA (1.0) processor will run emulation-mode for the "classic" binaries just as fast as the 75ns "classic" HP3000 (a Series 70). We have two old "classic" HP3000 systems and recently purchased an HP3000/960 (HP-PA) system, roughly 25 MIPS (37ns clock). When you have that much computing power, emulation is no joke, it screams. AND, *EVERYTHING* works. I was absolutely amazed at the level of emulation, even some privileged code works (misc sys utilities)! All normal application oriented object code starts up and takes off. We were absolutely amazed at how flawless it all is. And we run some pretty tricky stuff that really made use of the old MPE 5 features specifically (XDS, Message Files), you name a nasty trick we probably used it (barring priviledge code). You have to consider though that HP *HAD* to make the emulation as perfect as possible. Even the latest versions of MPE XL STILL USE some of the old object code IN THE OPERATING SYSTEM. MPE XL 2.05 still used the compatibility mode spooling software from the old MPE 5 systems for printing, and it worked fine. And just one other note, compiling our applications software (student registration, finance, accounts receivable, all these things were in COBOL and used extra-data segments out the ying yang along with several other things along these lines) to "native mode" was almost as painless, most things re-compiled and ran without any tweaking. My hats off to HP for this amazing magic show, it seems to fool just about EVERYTHING we throw at it from our old "classic" HP3000 systems. -- Mike Santangelo (mike@socrates.umd.edu) UNIX / VMS Systems Manager