Subject: Re: http://www.netbsd.org/Library/Hardware/Machines/DEC/vax/full.html
To: Anders Magnusson <ragge@ludd.luth.se>
From: Johnny Billquist <bqt@Update.UU.SE>
List: port-vax
Date: 02/03/1998 18:39:24
> >
> > -TU81 (and in fact all "big tapes"): I think the TU81 had an own
> > interface card that spoke TMSCP with the CPU. Capacity is hard
> > to tell, it depends on the writing density, the length of the
> > tape :) These were 9-track tape drives, the TU80 could do 1600 bpi,
> > I don't know what a TU79/80/81+ could do...
> >
> TU80 didn't speak TMSCP, but the same as TS05 et.al. Max 1600 bpi.
Correct. Actually it's only 1600 bpi.
> TU81/81+ speak MSCP through a special interface. 6250 bpi.
1600/6250 bpi. TMSCP. How special the interface might be I don't know.
> TA78/81/90/91/92 uses a SDI cable and are connected to a HSC.
> TA78 and TA81 was 1/2" 6250 tapes, TA90 was a tape-robot that could handle
> about 200MB, TA91 400MB and TA92 many GB/tape. These three machines
> was built by IBM but labeled DEC.
Note that TU and TA tapes normally are the same mechanism, but with
different interfaces.
The TU78 is a massbus drive, while the TA78 was TMSCP.
TU81 and TA81 are both TMSCP, but the TA81 has a normal SDI cable, while
the TU81 has a ig D-sub.
I'm not sure any other of the TAs had similar TUs.
> Other tapes to know about:
> TS11: 1600 bpi, Massbus/Unibus.
Actually only Unibus.
> TU45: 1600 bpi, Massbus. Maybe Unibus, don't know.
> TU77: 1600 bpi, Massbus, autoload.
TU45 and TU77, were, along with TU16 and TE16, pure massbus drives, using
either the TM02 or the TM03 formatter.
All capable of 800/1600 bpi.
> TU78: 6250 bpi, Massbus, autoload.
1600/6250 actually.
> > 6200, introduced in '88, with BI and the CVAX chip, used SMP, the third
> > digit in the numbering refers to the number of CPUs
> > 2.8 VUPs, 256 max memory, 1K@80ns on-chip and 256KB@160ns on-board cache
> >
> > 6000-300 (in '89): CVAX+ chipset, introduced the XMI bus
> > 3.8 VUPs, memory and cache as with the 6200
> >
> Both 6200 and 6300 have XMI bus. The CPU cards can be intermixed :-)
>
> > 6000-400 (same year): with the Rigel chipset and vector-processing
> > abilities
> > 13 VUPs, max 512 MB RAM, 16 ns cycle time, 45-90 MFLOPS
> >
> Was vector processing optional? Our machine here doesn't have vector
> functionality. Was it a extra card? It's only about 6 VUPs per CPU,
> the -500 is 13 VUPs.
I haven't heard of it either, and I don't think our has it.
~6 VUPS is the currect number.
Johnny
Johnny Billquist || "I'm on a bus
|| on a psychedelic trip
email: bqt@update.uu.se || Reading murder books
pdp is alive! || tryin' to stay hip" - B. Idol