On 4 Jun 2012, at 23:15, Johnny Billquist wrote:
Nice.
Yeah, trouble is because I am using my current RasPi to work on VAX stuff I need a second one to run my PDP-11 sim on now :D C'mon Farnell get it together!!
It would be interesting to hear of a comparison. The "newer" VAXen are really not that bad. But of course, everything is relative...
Like I say, it's not chronically 'slow' it just has a few pregnant pauses. Example: when I an using the TCP/IP configuration script in VMS 7.3 the pauses between menu screens are quite long. Stuff at the DCL terminal on the other hand is very quick (DIR's of long directory lists, etc.). There is always a smidge of additional lag because I am using SSH and the RasPi is having to encode the SSH packets on the fly as it goes. Telnet is much faster so I may switch over to using the DZ11 terminal emulation to work on it. Only problem is that's not the Operator terminal on SimH so I can't do everything through it.
Talk about doing a misread. I saw that as an RSX-11 MultiProcessor system. :-)
:D That'd be nice... but tricky ;)
I would suspect that is because the CPU will still be just waiting for I/O lots of the time. Compilations, as well as running something like SYSGEN in RSX, is really I/O intensive. Not much CPU work in there.
I can tell when the computer is working on the SD card, the LED flashes, and some times it's still loading SD card data (disk images) and sometimes it's thinking really hard :) SYSGENs are snappy except for the time consuming parts (the actually assembly compiles for the Executive, Drivers, etc). The RAM is either DDR or DDR2 so there shouldn't be a speed issue there, the biggest bottleneck is really the SD card.
I am looking at trying to mount an external hard drive or SSD to handle the disk images instead of the SD card. For some ^$%^$% reason you can't mount a disk as a user in Linux (I might be missing something, admittedly) like you can in RSX and VMS (again, more demonstration that UNIX sucks ;)) so I have to futz about as root to do that.
At $35 it was bound to have limitations and for the most part they are livable-with.
--
Mark Benson
http://DECtec.info
Twitter: @DECtecInfo
HECnet: STAR69::MARK
Online Resource & Mailing List for DEC Enthusiasts.
RT-11 is "famous" for self-modifying code. That fact makes it so much
fun to debug when you have a real-time bug. Been there... Done that...
-Steve
-----Original Message-----
From: owner-hecnet at Update.UU.SE
[mailto:owner-hecnet at Update.UU.SE] On Behalf Of Johnny Billquist
Sent: Monday, June 04, 2012 22:52
To: hecnet at Update.UU.SE
Subject: Re: [HECnet] Speaking of IDLE
On 2012-06-05 04:44, Johnny Billquist wrote:
On 2012-06-05 03:24, Steve Davidson wrote:
Paul,
I think that depends on whether or not the idle loop
pattern (for the
console light display) is being used. The lights pattern most
certainly makes use of the WAIT instruction. The SJ monitor is a
completely different beast than FB and friends - no
question about it!
Well, there are more ways to spin a cat... Or spin the
lights in this
case. WAIT is one way to show something on the front panel. Loading
the display register is another.
RT-11 is in fact using the Switch Register at 177570 to
display things
on the front panel.
Look at RMONFB.MAC, around line 4800, to see the code.
(Every time I read the RT-11 sources, I feel a little dirty... No
offense meant for the RT-11 fans around here or anywhere.)
It's all conditionalized on LIGH$T, and there is not even
the option
if you are running the SJ monitor...
Aw, heck. I might as well post the code, since isn't that long...
====
3$:
.IF NE LIGH$T
.ROM DEC LITECT,VALUE=1
BNE 8$
ADD #512.,LITECT
4$: ROL 7$
BNE 5$
COM 7$
5$: BCC 6$
ADD #100,4$
BIC #200,4$
6$: BIT #LIGHT$,CONFG2
BEQ 8$
MOV (PC)+,@(PC)+
7$: .WORD 0,SR
.ENDC
8$:
..NULJ::
====
(And yes, SR is defined as 177570)
(And maybe you understand why I feel dirty when reading RT-11
after this fine example of code... It's cut-and-pasted right
from the source, it really looks just like this.)
Johnny
--
Johnny Billquist || "I'm on a bus
|| on a psychedelic trip
email: bqt at softjar.se || Reading murder books
pdp is alive! || tryin' to stay hip" - B. Idol
On Tue, Jun 05, 2012 at 12:17:23AM +0200, Johnny Billquist wrote:
FYI - Update now have a Cisco box, which will be setup within a few
days to be able to route DECnet over IP, so that will become another
options to hook into HECnet at the Update site.
Sorry that I haven't gotten arround to it sooner. Update is preparing an
exhibiton at the University Museum. Lots of work with that :)
I can't promise getting the Cisco installed this week, but I'll see what
I can do.
I can however recommend anyone being close to uppsala this summer to
take a tour past Gustavianum to look at some nice DEC (and other)
hardware. Here is a hilight of the objects:
DECSYSTEM-2060
PDP-12
PDP-11/70
PDP-8
CRAY YMP EL
DG NOVA 3
peripherals!
Cheers,
Pontus.
On 06/05/2012 08:50 AM, Peter Lothberg wrote:
Something got messed up, I did some clear operations on the permanent database and
now I don't remember how to make it work again.....
Anyone that can give me a hint? Its VMS 5.4 on a uvax2
NCP>purge known lines
All line parameters (Y, N): y
%NCP-I-SUCCESS, Success
Line = QNA-0
%NML-I-RECDELET, Database entry deleted
%NCP-I-SUCCESS, Success
Line = QNA-1
%NML-I-RECDELET, Database entry deleted
How do I put QNA-0 and QNA-1 back in to decnet?
NCP>purge known obje
All parameters (Y, N): y
%NCP-I-SUCCESS, Success
Object = TASK
%NML-I-RECDELET, Database entry deleted
%NCP-I-SUCCESS, Success
Object = MAIL
%NML-I-RECDELET, Database entry deleted
%NCP-I-SUCCESS, Success
Object = PHONE
%NML-I-RECDELET, Database entry deleted
%NCP-I-SUCCESS, Success
Object = NML
%NML-I-RECDELET, Database entry deleted
%NCP-I-SUCCESS, Success
Object = MIRROR
%NML-I-RECDELET, Database entry deleted
%NCP-I-SUCCESS, Success
Object = VPM
%NML-I-RECDELET, Database entry deleted
How do I put those objects back?
Eeeeek! This may be a "shotgun" approach, but to be on the safe side
I'd just run netconfig (@sys$manager:netconfig) and have it generate the
databases anew.
-Dave
--
Dave McGuire, AK4HZ
New Kensington, PA
Something got messed up, I did some clear operations on the permanent database and
now I don't remember how to make it work again.....
Anyone that can give me a hint? Its VMS 5.4 on a uvax2
NCP>purge known lines
All line parameters (Y, N): y
%NCP-I-SUCCESS, Success
Line = QNA-0
%NML-I-RECDELET, Database entry deleted
%NCP-I-SUCCESS, Success
Line = QNA-1
%NML-I-RECDELET, Database entry deleted
How do I put QNA-0 and QNA-1 back in to decnet?
NCP>purge known obje
All parameters (Y, N): y
%NCP-I-SUCCESS, Success
Object = TASK
%NML-I-RECDELET, Database entry deleted
%NCP-I-SUCCESS, Success
Object = MAIL
%NML-I-RECDELET, Database entry deleted
%NCP-I-SUCCESS, Success
Object = PHONE
%NML-I-RECDELET, Database entry deleted
%NCP-I-SUCCESS, Success
Object = NML
%NML-I-RECDELET, Database entry deleted
%NCP-I-SUCCESS, Success
Object = MIRROR
%NML-I-RECDELET, Database entry deleted
%NCP-I-SUCCESS, Success
Object = VPM
%NML-I-RECDELET, Database entry deleted
How do I put those objects back?
--P
40 VUPs.
Nice! ~ the same then.
Oh yes, I'll bet! And keep in mind it's not just the CPU speed but
the memory bandwidth. It really sped up during that era.
True. Even just using the 4000-m90 for simple numerical problems, it's
quite slippery.
Sadly, no VPU though.
I WANT WANT WANT a 6000-500 with a vector board. It's a shame that
VAX 6000s have dried up so quickly. I must've passed up thirty of
them
over the past fifteen years. I'm getting a 6300 next month from a
semi-local guy. That'll be nice, but no vector boards for that model.
Will keep an eye out. I'd love a 6000 also, but doubt I could power it.
Not aware of the vector board. I did do an mp3 decoding routine using
Rigel's VPU back in '99 on a bet for beer. Was the 6000/500 VPU just
Rigels or something else?
Al
On 06/04/2012 11:40 PM, Boyanich, Alastair wrote:
Some of them are downright zippy. I'll have a VAX 7000 and a
4000-700a on HECnet soon! Zoom!
Dave: Nice. What's the 4000-700a pull VUP wise?
40 VUPs.
I was *VERY* shocked
going from a 4000-m300 "Pele" at 8VUPS to the 4000-m90A at 38VUPS. It's
like greased lightening by comparison.
Oh yes, I'll bet! And keep in mind it's not just the CPU speed but
the memory bandwidth. It really sped up during that era.
Sadly, no VPU though.
I WANT WANT WANT a 6000-500 with a vector board. It's a shame that
VAX 6000s have dried up so quickly. I must've passed up thirty of them
over the past fifteen years. I'm getting a 6300 next month from a
semi-local guy. That'll be nice, but no vector boards for that model.
-Dave
--
Dave McGuire, AK4HZ
New Kensington, PA
Some of them are downright zippy. I'll have a VAX 7000 and a
4000-700a on HECnet soon! Zoom!
Dave: Nice. What's the 4000-700a pull VUP wise? I was *VERY* shocked
going from a 4000-m300 "Pele" at 8VUPS to the 4000-m90A at 38VUPS. It's
like greased lightening by comparison. Sadly, no VPU though.
Al.
On 2012-06-05 04:44, Johnny Billquist wrote:
On 2012-06-05 03:24, Steve Davidson wrote:
Paul,
I think that depends on whether or not the idle loop pattern (for the
console light display) is being used. The lights pattern most certainly
makes use of the WAIT instruction. The SJ monitor is a completely
different beast than FB and friends - no question about it!
Well, there are more ways to spin a cat... Or spin the lights in this
case. WAIT is one way to show something on the front panel. Loading the
display register is another.
RT-11 is in fact using the Switch Register at 177570 to display things
on the front panel.
Look at RMONFB.MAC, around line 4800, to see the code.
(Every time I read the RT-11 sources, I feel a little dirty... No
offense meant for the RT-11 fans around here or anywhere.)
It's all conditionalized on LIGH$T, and there is not even the option if
you are running the SJ monitor...
Aw, heck. I might as well post the code, since isn't that long...
====
3$:
.IF NE LIGH$T
.ROM DEC LITECT,VALUE=1
BNE 8$
ADD #512.,LITECT
4$: ROL 7$
BNE 5$
COM 7$
5$: BCC 6$
ADD #100,4$
BIC #200,4$
6$: BIT #LIGHT$,CONFG2
BEQ 8$
MOV (PC)+,@(PC)+
7$: .WORD 0,SR
.ENDC
8$:
..NULJ::
====
(And yes, SR is defined as 177570)
(And maybe you understand why I feel dirty when reading RT-11 after this fine example of code... It's cut-and-pasted right from the source, it really looks just like this.)
Johnny
--
Johnny Billquist || "I'm on a bus
|| on a psychedelic trip
email: bqt at softjar.se || Reading murder books
pdp is alive! || tryin' to stay hip" - B. Idol
On 2012-06-05 03:24, Steve Davidson wrote:
-----Original Message-----
From: owner-hecnet at Update.UU.SE
[mailto:owner-hecnet at Update.UU.SE] On Behalf Of Paul_Koning at Dell.com
Sent: Monday, June 04, 2012 20:58
To: hecnet at Update.UU.SE
Subject: Re: [HECnet] Speaking of IDLE
On Jun 4, 2012, at 6:22 PM, Johnny Billquist wrote:
On 2012-06-04 13:52, Brian Hechinger wrote:
Anyone know how it works? I'd really like to get it
working on Solaris.
Simple. For a VAX, simh tries to detect the instruction
sequence that the OS do in the idle loop. That's why it
depends on which OS you are running on the emulated machine.
For a PDP-11 it is (supposedly) simpler, since hopefully
the idle loop
of the OS uses the PDP-11 WAIT instruction, but that
actually depends
on which OS we're talking about. Not sure that RT-11 do,
for instance.
(I have some vague memory of discussing this with someone a
few years
ago, and coming to the realization that not all PDP-11
software might
be using the WAIT.)
For other hardware and OS combinations, the answers might
differ even more.
The doc says that RT and Unix do it differently (no WAIT). I
haven't seen the Unix code but I did see the one for RT (F/B
version), and indeed, no WAIT instruction there. I'm not
sure why not. RT11 S/J seems to just be full of spin loops,
no central idle of any kind that I can see.
paul
Paul,
I think that depends on whether or not the idle loop pattern (for the
console light display) is being used. The lights pattern most certainly
makes use of the WAIT instruction. The SJ monitor is a completely
different beast than FB and friends - no question about it!
Well, there are more ways to spin a cat... Or spin the lights in this case. WAIT is one way to show something on the front panel. Loading the display register is another.
RT-11 is in fact using the Switch Register at 177570 to display things on the front panel.
Look at RMONFB.MAC, around line 4800, to see the code.
(Every time I read the RT-11 sources, I feel a little dirty... No offense meant for the RT-11 fans around here or anywhere.)
It's all conditionalized on LIGH$T, and there is not even the option if you are running the SJ monitor...
Johnny
--
Johnny Billquist || "I'm on a bus
|| on a psychedelic trip
email: bqt at softjar.se || Reading murder books
pdp is alive! || tryin' to stay hip" - B. Idol
