roy@phri.UUCP (Roy Smith) (08/04/88)
For months I've been listening to people talking about "4.3-tahoe" but I still havn't seen anybody explain what a "tahoe" is. -- Roy Smith, System Administrator Public Health Research Institute {allegra,philabs,cmcl2,rutgers}!phri!roy -or- phri!roy@uunet.uu.net "The connector is the network"
chris@mimsy.UUCP (Chris Torek) (08/05/88)
In article <3420@phri.UUCP> roy@phri.UUCP (Roy Smith) writes: >For months I've been listening to people talking about "4.3-tahoe" >but I still havn't seen anybody explain what a "tahoe" is. The Tahoe is a series of machines (as the VAX is a series). Tahoes are sold by CCI, Harris, Sperry/Unisys, and by one other company whose name I have forgotten. The original Tahoe was the CCI Power 6/32, a ~6-VAX-MIPS box with (uh oh) VERSAbus-based disk and tape drives. VME-based peripherals are now available for this CPU. The Tahoe has been called a `RISC': Reused Instruction Set Computer. The Tahoe architecture is remarkably similar to that of the VAX; however, many of the special purpose instructions are missing (no branch on bit set and clear interlocked, e.g.), many of the fancier addressing modes are missing; and the basic instruction cycle is much faster (the CCI is mostly TTL, yet runs about as fast as the ECL gate array VAX 8650). (Whether the speed difference can be attributed to leaving out the fancier instructions and addressing modes is a topic for another raging comp.arch controversy :-) .) 4.3BSD-tahoe runs on the original CCI Tahoe and (as far as I know) on the Harris HCX-7. It may soon run on the HCX-9. I believe the Sperry 7000 is identical to the HCX-7, so it should run on that too. 4.3BSD-tahoe also runs on the VAX 11/750, 11/780, 11/785, 8200, 8250, 8600, and 8650, on the MicroVAX-II, and possibly on the VAX 11/725 and 11/730 (this has not been tested). The tape from Berkeley has Tahoe binaries rather than VAX binaries, but the full source tree is present, so installation on a VAX is largely a matter of replacing the old sources and recompiling. -- In-Real-Life: Chris Torek, Univ of MD Comp Sci Dept (+1 301 454 7163) Domain: chris@mimsy.umd.edu Path: uunet!mimsy!chris
klein@lupine.UUCP (Doug Klein ) (08/05/88)
In article <3420@phri.UUCP>, roy@phri.UUCP (Roy Smith) writes: > > For months I've been listening to people talking about "4.3-tahoe" > but I still havn't seen anybody explain what a "tahoe" is. Last time I checked the Chevrolet ads, a '4.3 Tahoe' was an S10 pickup truck with a 4.3 litre V6 engine, and the top of the line interior -:) Doug
jack@leo.UUCP ( Jack Benkual) (08/05/88)
In article <3420@phri.UUCP>, roy@phri.UUCP (Roy Smith) writes: > > For months I've been listening to people talking about "4.3-tahoe" > but I still havn't seen anybody explain what a "tahoe" is. > -- > Roy Smith, System Administrator > Public Health Research Institute > {allegra,philabs,cmcl2,rutgers}!phri!roy -or- phri!roy@uunet.uu.net > "The connector is the network" Tahoe was the project name for the CCI Power 6/32 Superminicomputer available since 1985. It has an architecture very similar to VAX/780 and probably that and the higher performance (8 VAX MIPS in 1985) was the reason that Berkeley released 4.3-Tahoe. ICL ( Clan4-7) and Unisys ( 7000 series) sales it as an OEM. Harris (HCX-7) got the technology and manufacturing rights from CCI. It comes with the different configurations and performance range of 3-15 MIPS. They are probably 1700 systems installed world wide manufactured by CCI. -- Simplify! Simplify! Simplify! | jack Benkual @ uunet!ccicpg!leo!jack | CCI Computers (nee Computer Consoles Inc)
ralphw@ius3.ius.cs.cmu.edu (Ralph Hyre) (08/08/88)
In article <3259@leo.UUCP> jack@leo.UUCP ( Jack Benkual) writes: >In article <3420@phri.UUCP>, roy@phri.UUCP (Roy Smith) writes: >> but I still havn't seen anybody explain what a "tahoe" is. >Tahoe was the project name for the CCI Power 6/32 Superminicomputer >... architecture similar to VAX/780 and probably that ... >and the higher performance (8 VAX MIPS in 1985) was the reason that Berkeley >released 4.3-Tahoe. Well, I also heard (from a combination of Byte and the USENET, I believe) that development work moved from Vaxen to CCI machines because of the EXTREME difficulty of getting needed technical information out of DEC (at the time, they may have changed now). fodder for inews fodder for inews fodder for inews fodder for inews fodder for inews -- - Ralph W. Hyre, Jr. Internet: ralphw@ius2.cs.cmu.edu Phone:(412)268-{2847,3275} CMU-{BUGS,DARK} Amateur Packet Radio: N3FGW@W2XO, or c/o W3VC, CMU Radio Club, Pittsburgh, PA
chris@mimsy.UUCP (Chris Torek) (08/08/88)
>In article <3259@leo.UUCP> jack@leo.UUCP (Jack Benkual) writes: >>... architecture similar to VAX/780 and probably that ... >>and the higher performance ... was the reason that Berkeley >>released 4.3-Tahoe. In article <2620@pt.cs.cmu.edu> ralphw@ius3.ius.cs.cmu.edu (Ralph Hyre) writes: >Well, I also heard (from a combination of Byte and the USENET, I believe) >that development work moved from Vaxen to CCI machines because of the >EXTREME difficulty of getting needed technical information out of DEC >(at the time, they may have changed now). The difficulty is still there; but on the other hand, CCI and Harris have not exactly been shoving information at Berkeley either. The improved performance was certainly one reason for the switch; the fact that Sam Leffler had already done much of the work anyway was another. It is also nice to have two differing architectures around in order to test portability: the new VM will be able to run on a virtual- address-cache architecture right away, since the Tahoe is such a machine. (Incidentally, tweaking the old VAX-oriented code to work on that machine took surprisingly few changes.) -- In-Real-Life: Chris Torek, Univ of MD Comp Sci Dept (+1 301 454 7163) Domain: chris@mimsy.umd.edu Path: uunet!mimsy!chris
ralphw@ius3.ius.cs.cmu.edu (Ralph Hyre) (08/09/88)
In article <12889@mimsy.UUCP> chris@mimsy.UUCP (Chris Torek) writes: [about 4.3-tahoe] >It is also nice to have two differing architectures around in order >to test portability: the new VM will be able to run on a virtual- >address-cache architecture right away, since the Tahoe is such a machine. >(Incidentally, tweaking the old VAX-oriented code to work on that machine >took surprisingly few changes.) One might wonder why they didn't go with Suns, since they have a friend in a high place (Bill Joy), and could also presumably get good deals on them :-) I heard that Sun runs SunOS on Vaxen for internal testing purposes, it would be interesting if Berkeley sported 4.3 on Suns, since it would fix even more portability problems like NULL derefs and such. I can see that advantage in not breaking everything at once, however :-) -- - Ralph W. Hyre, Jr. Internet: ralphw@ius2.cs.cmu.edu Phone:(412)268-{2847,3275} CMU-{BUGS,DARK} Amateur Packet Radio: N3FGW@W2XO, or c/o W3VC, CMU Radio Club, Pittsburgh, PA
chris@mimsy.UUCP (Chris Torek) (08/09/88)
[This discussion should move, but I have no idea to where it should move...] In article <2642@pt.cs.cmu.edu> ralphw@ius3.ius.cs.cmu.edu (Ralph Hyre) writes: >One might wonder [Berkeley] didn't go with Suns [for a second porting >base], since they have a friend in a high place (Bill Joy), and could >also presumably get good deals on them :-) Well, speaking only for myself, I have considered this. The problem is incentive, or rather lack thereof: Porting 4.3BSD to the Sun-2 and Sun-3 would take much effort (think about all those drivers!...), and neither machine is particularly fast; porting to the Sun-4 requires stable Sun-4s (which are now available, but were not a few years ago). Moreover, the return would be fairly small: SunOS 3.5 is close enough to 4.3BSD that we do not suffer `porting shock'. -- In-Real-Life: Chris Torek, Univ of MD Comp Sci Dept (+1 301 454 7163) Domain: chris@mimsy.umd.edu Path: uunet!mimsy!chris
wunder@hp-sde.SDE.HP.COM (Walter Underwood) (08/10/88)
/ hp-sde:comp.arch / chris@mimsy.UUCP (Chris Torek) / The Tahoe architecture is remarkably similar to that of the VAX; ... and the basic instruction cycle is much faster (the CCI is mostly TTL, yet runs about as fast as the ECL gate array VAX 8650). Amusing. The HP9000 model 840 is also TTL and is also about the same as an 8650 (or one of those ECL VAXen, I've really lost track). The series 800 also seems to have about the same performance range as the CCI -- roughly 5 to 14 MIPS. I guess we've all learned a lot about architecture since the mid-70s (when the VAX was designed). wunder