Dave be careful -- S/360 Model 67 has VM in the late 1960's - TSS and it's brother MTS, both rely on it. The 67 is a Model 65 with a Data Address Translation unit (DAT box) - is supplied by a 8 x 32 bit TLB which is in a cabinet that t'ed off the main CPU and is about the same size en entire Vax 780 which would follow 10 years later.
Think about that for a minute -- an 8 word TLB. At Intel we regularly examine the different sizes of the different parts of the memory system. Core 7 (aka Nehalem of a few years ago) has a two-level TLB: the first level of TLB is shared between data and instructions. The level 1 data TLB now stores 64 entries for small pages (4K) or 32 for large pages (2M/4M), while the level 1 instruction TLB stores 128 entries for small pages (the same as with Core 2) and seven for large pages. The second level is a unified cache that can store up to 512 entries and operates only with small pages. Also it is also interesting to consider that while the AT&T folks came off of Multics, a number of us university types that would work on earlier Unix came from TSS and MTS (one 360/67). In fact, TSS is still the only system I ever used that lived in the debugger as your command system - which I always thought was a cool idea. As for what started this thread. I think it is interesting that the long term successful architectures in the market did have a excellent compatibility stories. IBM with system 360 certainly set a high bar, and DEC has nothing to be ashamed of, the different DEC lines, particularly the Vax, did a great job here. In truth, probably the best of pure compatibility story has to be Intel. The H/L registers of the 4004 are still there ;-) Seriously, the INTEL*64 is from an computer science standpoint, not an architecture you would create from scratch. But Intel has completely understood the economics of SW compatibility. Also, if you peeked inside a modern processor, you would discover they are dataflow engines and put together with all of the modern computer science; but there is about a 5% silicon tax paid for compatibility. Clearly, my siblings at Intel believe it's worth tax and the customers seem to keep wanting it. Clem On Tue, Feb 16, 2016 at 8:15 AM, Dave Wade <dave.g4...@gmail.com> wrote: > > -----Original Message----- > > From: Simh [mailto:simh-boun...@trailing-edge.com] On Behalf Of Wilm > > Boerhout > > Sent: 16 February 2016 11:58 > > To: simh@trailing-edge.com > > Subject: Re: [Simh] VAX/VMS > > > > Johnny Billquist schreef op 16-2-2016 om 12:49: > > > > > > No, it is not. Talk to IBM about S/360... :-) And there are some VAXen > > S/360 compatibility is only forward, and only to a certain point. S/360 > and S/370 are both 24-bit addressing and fairly compatible, but S/370 > (Mostly) has Virtual Memory as standard. > > Then came the "great divide" S/370XA. XA mode has 31-bit addressing and > different I/O instructions. Some of the XA boxes will work is S/370 mode, > but many won't. > > More recently IBM moved to 64-bit hardware. Again some will boot in 31-bit > mode but more recent boxes need a 64-bit OS. > > So the earliest incarnations of "OS", which were I guess "MFT" which is > basically a fixed number of partitions will run on later machines until you > get to systems which will only run in 31bit mode. (XA Mode). > > OS/VS2 and its siblings MVS (This is the free version), MVS/SP (The paid > for version) will only run on S/370 or later, not on 360, as they need > Virtual Memory and it stops working at the same point as MFT when 31 bit > only machines appear. There are also issues of Virtual Memory Page size > which may stop MVS (the free version working) working on some hardware > (there are patches to work round this). > > You also have issues over disk (DASD in IBM speak) support. So whilst MFT > was written for a 1996 S/360 it would in theory run on an P390E from 1996 > so 30 years of computability. However, it would need older disks, which the > P/390E cannot support. > > Of course these changes are really only to do with programs that run in > supervisor state. User mode programs generally will run unchanged from 1966 > through to the present day, and the latest zOS a descendant of MVS will > still run 24-bit applications. I am pretty sure that until a few years > many commercial sites, so mostly Cobol, still used the older "free" > Fortran-66 compiler for the odd Fortran job. > > > > on which V7.3 will definitely not run. How about rtVAX for example. > > > > > I stand corrected. Please note that I had a marketing job once. It > sticks... > > ... I also believe that some of the in-compatibility in IBM kit is to > drive the hardware->software->Hardware->Software upgrade chain and keep the > dollars rolling in... > > Dave G4UGM > > > > > > _______________________________________________ > > Simh mailing list > > Simh@trailing-edge.com > > http://mailman.trailing-edge.com/mailman/listinfo/simh > > _______________________________________________ > Simh mailing list > Simh@trailing-edge.com > http://mailman.trailing-edge.com/mailman/listinfo/simh >
_______________________________________________ Simh mailing list Simh@trailing-edge.com http://mailman.trailing-edge.com/mailman/listinfo/simh