henry@utzoo.UUCP (Henry Spencer) (12/11/85)
A friend pointed out something interesting: there is a job ad in net.jobs, from National, asking for... RISC people! CMOS chip designers, computer architects, compiler people, Unix-porting people. Sounds serious. Just how is this going to affect the existing 32000 line? Compatibility is obviously out, or they wouldn't be hiring compiler and Unix people. Has National decided to make a fresh start and abandon the mistakes of the past? (Undoubtedly the existing products will be "fully supported"... for a little while...)
jans@mako.UUCP (Jan Steinman) (12/13/85)
In article <6217@utzoo.UUCP> henry@utzoo.UUCP (Henry Spencer) writes: >A friend pointed out something interesting: there is a job ad in net.jobs, >from National, asking for... RISC people!... > >Just how is this going to affect the existing 32000 line? Compatibility is >obviously out, or they wouldn't be hiring compiler and Unix people. I don't see why "compatibility is obviously out". A RISC design has higher potential for compatibility than a conventional design, through the use of emulators. A tightly-coupled emulator could provide binary compatibility, possibly with improved performance, yet allow radical improvements in performance by re-compiling the source code into RISC instructions. -- :::::: Artificial Intelligence Machines --- Smalltalk Project :::::: :::::: Jan Steinman Box 1000, MS 60-405 (w)503/685-2956 :::::: :::::: tektronix!tekecs!jans Wilsonville, OR 97070 (h)503/657-7703 ::::::
chongo@nsc.UUCP (Landon Noll) (12/15/85)
In article <6217@utzoo.UUCP> henry@utzoo.UUCP (Henry Spencer) writes: >A friend pointed out something interesting: there is a job ad in net.jobs, >from National, asking for... RISC people! CMOS chip designers, computer >architects, compiler people, Unix-porting people. Sounds serious. Yes, such a project exists. It is at the point where we are searching for additional people to join the project. The project is quite serious. >Just how is this going to affect the existing 32000 line? This new project does NOT signal a dropping of the 32000 line. The 32000 line is going well. In fact work towards the the next generation cpu is going strong. Yeilds are up, and the `3*3**' runs a solid Unix. New yummo 32xxx chips are in the pipes. National just happens to be a multi-nation company that is big enough to do more than one thing and chew bubble gum at the same time! :-} chongo <ns128064 in 1994> /\oo/\
henry@utzoo.UUCP (Henry Spencer) (12/16/85)
> >Just how is this going to affect the existing 32000 line? Compatibility is > >obviously out, or they wouldn't be hiring compiler and Unix people. > > I don't see why "compatibility is obviously out". A RISC design has higher > potential for compatibility than a conventional design, through the use of > emulators. A tightly-coupled emulator could provide binary compatibility, > possibly with improved performance, yet allow radical improvements in > performance by re-compiling the source code into RISC instructions. My point was not that you couldn't get some degree of compatibility in the same way that pdp11 programs run on the VAX or 7094 programs ran on the right model of 360. My point was that the underlying machine clearly is incompatible, i.e. it will not be a 32000. (Hard to see how you could make a 32000 RISC anyway.) -- Henry Spencer @ U of Toronto Zoology {allegra,ihnp4,linus,decvax}!utzoo!henry