Re: 'Realistic' accounting values
On 1/18/21 11:00 PM, Mark L. Gaubatz
via groups.io wrote:
Interesting rates – especially the
commercial rates. At these rates, my work on a single
five-month project alone at the time would have had a direct
cost of roughly $7MM, and over $640MM in resources for QA (we
were running 16 engines, and sometimes more, 24x7, on machines
significantly faster than the 3090 at the time).
What kind of machine??
Were you the commercial rate?? Academic billing can be funny.
I know at Clarkson in late 1970's, there two type of charges on
the S/630-65J under OS/360: "funny" and "real" money.?? Most
charges for routine use were funny, but physical supplies like
paper and cards were real.
In the mid-1980's, they got a baby supercomputer (Alliant?) in and they
only allowed "real" charges on it (basically, funded grants).? I
believe it was? under utilized, which implies no one had grant
money for computer time.? :-)
--
Drew Derbyshire
"The juvenile sea squirt wanders through the sea searching for a
suitable rock or hunk of coral to cling to and make its home for life.
For this task, it has a rudimentary nervous system. When it finds its
spot and takes root, it doesn't need its brain anymore so it eats it!
(It's rather like getting tenure.)" -- Daniel Dennett
|
Re: 'Realistic' accounting values
Interesting rates – especially the commercial rates. At these rates, my work on a single five-month project alone at the time would have had a direct cost of roughly $7MM, and over $640MM in resources for QA (we were running 16 engines, and sometimes more, 24x7, on machines significantly faster than the 3090 at the time). ? ? ?
toggle quoted message
Show quoted text
From: H390-MTS@groups.io <H390-MTS@groups.io> On Behalf Of Mike Alexander Sent: Monday, January 18, 2021 8:36 PM To: H390-MTS@groups.io Subject: Re: [H390-MTS] 'Realistic' accounting values ? Here are two PDFs that give the final rates on both MTS machines at Michigan. These are the contents of *RATES and *RATES.UB as of 1996. They mention a file *ALLRATES that is said to contain rate history, but this file seems to have been destroyed some time before 1996. Too bad, it would be interesting. Mike
|
Re: 'Realistic' accounting values
Here are two PDFs that give the final rates on both MTS machines at Michigan. These are the contents of *RATES and *RATES.UB as of 1996. They mention a file *ALLRATES that is said to contain rate history, but this file seems to have been destroyed some time before 1996. Too bad, it would be interesting.
Mike
On 18 Jan 2021, at 15:44, Mark L. Gaubatz via groups.io wrote:
toggle quoted message
Show quoted text
Some examples from 1975: ? 360/195 - ?$1800/hour - $0.500000/sec 370/125 - $130/hour - $0.036111/sec (public contract) 370/145 - $60/hour - $0.016667/sec (off hours, as available) 370/158 - $157/hour - $0.043611/sec (off hours) ? Mark dasdman ? ? From: H390-MTS@groups.io <H390-MTS@groups.io> On Behalf Of Yvan Janssens
Sent: Sunday, January 17, 2021 11:23 PM
To: h390-mts@groups.io
Subject: [H390-MTS] 'Realistic' accounting values ? I want to reconfigure my MTS instance to use period accurate/realistic billing settings instead of the zero charges it's currently configured with. Does anyone have a record of what would be considered 'accurate' charges for eg CPU time (without inflation added)?
|
Re: 'Realistic' accounting values
Some examples from 1975: ? 360/195 - ?$1800/hour - $0.500000/sec 370/125 - $130/hour - $0.036111/sec (public contract) 370/145 - $60/hour - $0.016667/sec (off hours, as available) 370/158 - $157/hour - $0.043611/sec (off hours) ? Mark dasdman ? ?
toggle quoted message
Show quoted text
From: H390-MTS@groups.io <H390-MTS@groups.io> On Behalf Of Yvan Janssens Sent: Sunday, January 17, 2021 11:23 PM To: h390-mts@groups.io Subject: [H390-MTS] 'Realistic' accounting values ? I want to reconfigure my MTS instance to use period accurate/realistic billing settings instead of the zero charges it's currently configured with. Does anyone have a record of what would be considered 'accurate' charges for eg CPU time (without inflation added)?
|
'Realistic' accounting values
Hi,
I want to reconfigure my MTS instance to use period accurate/realistic billing settings instead of the zero charges it's currently configured with.
Does anyone have a record of what would be considered 'accurate' charges for eg CPU time (without inflation added)?
Thank you!
-y
|
What, exactly, is Hercules emulating? A dial up TTY attached to a > 2703? A dial up 2741 on a 2703? MTS supports both but needs to know > which it is talking to. The device type in TABLES in MTS needs to be > set appropriately. I'm under-educated here. A single dialin, "TELE2", is about the extent of what I can provide right now. I'll have to dig into the herc code to see what I can learn about what's actually implemented. The UMTABLES file led me to hope that the needed configuration and DSR support for the hardwired decwriters referred to in code comments might still exist in the runnable D6.0A distribution. Being able to e.g. do file transfers with kermit would be useful. De
|
Mike,
The latest Hercules can emulate TTY or 2741 , dial-up or hardwired. There
are extra options for 2741 8-bit mode for APL
Dave
G4UGM
toggle quoted message
Show quoted text
-----Original Message-----
From: H390-MTS@groups.io <H390-MTS@groups.io> On Behalf Of Mike
Alexander
Sent: 19 December 2020 07:53
To: H390-MTS@groups.io
Cc: George Helffrich <ghfbsd@...>
Subject: Re: [H390-MTS] 2703 lines
Those MTS devices are Memorex 1270 devices (that's what MRXA means)
and are not directly compatible with IBM 2703 devices, although similar.
There is one DSR in MTS that supports both types, but it has to know which
type it is talking to. You might get further if you patch the device type to "TTY " or maybe "2741" but probably not. George Helffrich recently got 2741
support working in D4.1 MTS running on simh emulating a 360/67. This isn't really directly relevant, but his experience might shed some light on how to make this work.
What, exactly, is Hercules emulating? A dial up TTY attached to a 2703?
A dial up 2741 on a 2703? MTS supports both but needs to know which it is talking to. The device type in TABLES in MTS needs to be set appropriately.
George also found a bug in the DSR that supports these devices. It has
assembly parameters controlling whether to include support for 2703 and/or
MRXA controllers. However if you turn on both options the 2703 support
doesn't work right. By the time of D4.1 we no longer had any real IBM
2703's and so the DSR was never tested with them. Apparently it didn't work.
George's solution was to reassemble TSFO with support for MRXA turned
off. This made 2703 support work. He also had to fix a few minor bugs in
simh which isn't relevant, but you might well find some incompatibilities
between Hercules 2703 support and MTS.
It would be nice to get this working in D6.0 since a simulated 2741 makes a nice terminal for some purposes.
I've CCed you, George. Let me know if I've misrepresented anything you did.
Mike
On 17 Dec 2020, at 15:38, Dennis Boone wrote:
Folks,
The UM tables file included in the d6 distribution includes:
SA00 IODEVICE ADDRESS=(00C0,10),CUNUMBR=(003),UNIT=SDA
SA10 IODEVICE ADDRESS=(00CA,2),CUNUMBR=(003),UNIT=SDA
LA00 IODEVICE ADDRESS=(00CC,10),CUNUMBR=(003),UNIT=MRXA
LA10 IODEVICE ADDRESS=(00D6,10),CUNUMBR=(003),UNIT=MRXA
LA20 IODEVICE ADDRESS=(00E0,10),CUNUMBR=(003),UNIT=MRXA
LA30 IODEVICE ADDRESS=(00EA,6),CUNUMBR=(003),UNIT=MRXA
SA12 IODEVICE ADDRESS=(00F0,8),CUNUMBR=(003),UNIT=SDA
LA44 IODEVICE ADDRESS=(00F8,6),CUNUMBR=(003),UNIT=MRXA
LA50 IODEVICE ADDRESS=(00FE,2),CUNUMBR=(003),UNIT=MRXA
It'd be nice to have at least a few plain tty type lines available.
Are
these devices included in the D6.0A runnable distribution? Any hints
on how to enable them? I tried adding:
00FE 2703 LPORT=2741 DIAL=IN TERM=TTY LNCTL=TELE2
to hercules.cnf, and giving `MTS LA50` on the console after ipl. But
MTS says:
00029..MTS LA50
00029 MTS "LA50": does not exist.
and hercules' status seems suspect:
HHC02279I 0:00FE 2703 ASYNC STA=DISA CN=NO, EIB=NO OP=IDLE IO[0] open
Thanks,
De
|
Those MTS devices are Memorex 1270 devices (that's what MRXA means) and are not directly compatible with IBM 2703 devices, although similar. There is one DSR in MTS that supports both types, but it has to know which type it is talking to. You might get further if you patch the device type to "TTY " or maybe "2741" but probably not. George Helffrich recently got 2741 support working in D4.1 MTS running on simh emulating a 360/67. This isn't really directly relevant, but his experience might shed some light on how to make this work.
What, exactly, is Hercules emulating? A dial up TTY attached to a 2703? A dial up 2741 on a 2703? MTS supports both but needs to know which it is talking to. The device type in TABLES in MTS needs to be set appropriately.
George also found a bug in the DSR that supports these devices. It has assembly parameters controlling whether to include support for 2703 and/or MRXA controllers. However if you turn on both options the 2703 support doesn't work right. By the time of D4.1 we no longer had any real IBM 2703's and so the DSR was never tested with them. Apparently it didn't work. George's solution was to reassemble TSFO with support for MRXA turned off. This made 2703 support work. He also had to fix a few minor bugs in simh which isn't relevant, but you might well find some incompatibilities between Hercules 2703 support and MTS.
It would be nice to get this working in D6.0 since a simulated 2741 makes a nice terminal for some purposes.
I've CCed you, George. Let me know if I've misrepresented anything you did.
Mike
toggle quoted message
Show quoted text
On 17 Dec 2020, at 15:38, Dennis Boone wrote: Folks,
The UM tables file included in the d6 distribution includes:
SA00 IODEVICE ADDRESS=(00C0,10),CUNUMBR=(003),UNIT=SDA
SA10 IODEVICE ADDRESS=(00CA,2),CUNUMBR=(003),UNIT=SDA
LA00 IODEVICE ADDRESS=(00CC,10),CUNUMBR=(003),UNIT=MRXA
LA10 IODEVICE ADDRESS=(00D6,10),CUNUMBR=(003),UNIT=MRXA
LA20 IODEVICE ADDRESS=(00E0,10),CUNUMBR=(003),UNIT=MRXA
LA30 IODEVICE ADDRESS=(00EA,6),CUNUMBR=(003),UNIT=MRXA
SA12 IODEVICE ADDRESS=(00F0,8),CUNUMBR=(003),UNIT=SDA
LA44 IODEVICE ADDRESS=(00F8,6),CUNUMBR=(003),UNIT=MRXA
LA50 IODEVICE ADDRESS=(00FE,2),CUNUMBR=(003),UNIT=MRXA
It'd be nice to have at least a few plain tty type lines available. Are
these devices included in the D6.0A runnable distribution? Any hints on
how to enable them? I tried adding:
00FE 2703 LPORT=2741 DIAL=IN TERM=TTY LNCTL=TELE2
to hercules.cnf, and giving `MTS LA50` on the console after ipl. But
MTS says:
00029..MTS LA50
00029 MTS "LA50": does not exist.
and hercules' status seems suspect:
HHC02279I 0:00FE 2703 ASYNC STA=DISA CN=NO, EIB=NO OP=IDLE IO[0] open
Thanks,
De
|
Folks,
The UM tables file included in the d6 distribution includes:
SA00 IODEVICE ADDRESS=(00C0,10),CUNUMBR=(003),UNIT=SDA
SA10 IODEVICE ADDRESS=(00CA,2),CUNUMBR=(003),UNIT=SDA
LA00 IODEVICE ADDRESS=(00CC,10),CUNUMBR=(003),UNIT=MRXA
LA10 IODEVICE ADDRESS=(00D6,10),CUNUMBR=(003),UNIT=MRXA
LA20 IODEVICE ADDRESS=(00E0,10),CUNUMBR=(003),UNIT=MRXA
LA30 IODEVICE ADDRESS=(00EA,6),CUNUMBR=(003),UNIT=MRXA
SA12 IODEVICE ADDRESS=(00F0,8),CUNUMBR=(003),UNIT=SDA
LA44 IODEVICE ADDRESS=(00F8,6),CUNUMBR=(003),UNIT=MRXA
LA50 IODEVICE ADDRESS=(00FE,2),CUNUMBR=(003),UNIT=MRXA
It'd be nice to have at least a few plain tty type lines available. Are
these devices included in the D6.0A runnable distribution? Any hints on
how to enable them? I tried adding:
00FE 2703 LPORT=2741 DIAL=IN TERM=TTY LNCTL=TELE2
to hercules.cnf, and giving `MTS LA50` on the console after ipl. But
MTS says:
00029..MTS LA50
00029 MTS "LA50": does not exist.
and hercules' status seems suspect:
HHC02279I 0:00FE 2703 ASYNC STA=DISA CN=NO, EIB=NO OP=IDLE IO[0] open
Thanks,
De
|
Re: 1996 System Release Status?
On Tue, Apr 28, 2020 at 08:22:49PM -0400, Mike Alexander wrote: Unfortunately nothing much has been done about this in quite a while.
The good news is that several of us who have to do the work to get it
out were talking about it the other day. Perhaps something will happen
while we're all stuck at home.
One good thing is that if we get this out it will likely have TCP/IP
support given the correct updates to Hercules. The version I have
includes an FTP client and server as well as network printing to a
Postscript printer attached to my Mac. This is done by adding emulation
of the network interface devices MTS used for this in the past to
Hercules.
Keep pushing, I really hope it gets done. Has any progress been made on the 1996 system release? -- Kevin Bruceville, TX What's the definition of a legacy system? One that works! Errare humanum est, ignoscere caninum.
|
Re: Printer does work on SDL-HYPERION-4.2.x
According to the documentation ?
? a 1403 needs a cctape=(….) , the fcb format is only valid for 3203, 3211. “optprint” does not appear to be valid at all ? Dave ?
toggle quoted message
Show quoted text
From: H390-MTS@groups.io <H390-MTS@groups.io> On Behalf Of marcoxa@...
Sent: 17 October 2020 14:36
To: H390-MTS@groups.io
Subject: Re: [H390-MTS] Printer does work on SDL-HYPERION-4.2.x? Thanks
I have downloaded the "patched" MTSD6.0Aupdated.cnf file from the main web site and just followed instructions.
This is what I have in my .cnf file for SDL 4.2.1. ? ```
000E?? 1403?? Units/PTR2.txt lpp=66 fcb=1:11,4:1,5:5,8:10,14:6,19:4,24:7,34:2,44:6,49:4,54:7,63:8,66:3 optprint # PTR2
# 000E?? 1403?? Units/PTR2.txt nofcbcheck? # PTR2
# 000E?? 1403?? Units/PTR2.txt???????????? # PTR2
```
The three printers are those I tried.? I do not have set up a 3211 at 0000.
The error I get in the first two cases are about the unsupported fcb/nofcbcheck
I am pretty sure I did start HASP in the last test.
All the best
Marco
|
Re: Printer does work on SDL-HYPERION-4.2.x
In what way did they fail? I seem to recall that the options related to loading the FCB were added to Hercules while we were getting MTS up. Perhaps they are not in the version you are using.
Mike
toggle quoted message
Show quoted text
On 17 Oct 2020, at 10:04, marcoxa@... wrote:
Hi
I was able to make the third, vanilla option to work in my setup.? It was HASP not running properly after all.
Still curious why the other two options bomb.
Thanks
All the best
Marco
|
Re: Printer does work on SDL-HYPERION-4.2.x
Hi
I was able to make the third, vanilla option to work in my setup.? It was HASP not running properly after all.
Still curious why the other two options bomb.
Thanks
All the best
Marco
|
Re: Printer does work on SDL-HYPERION-4.2.x
Thanks
I have downloaded the "patched" MTSD6.0Aupdated.cnf file from the main web site and just followed instructions.
This is what I have in my .cnf file for SDL 4.2.1.
?
```
000E?? 1403?? Units/PTR2.txt lpp=66 fcb=1:11,4:1,5:5,8:10,14:6,19:4,24:7,34:2,44:6,49:4,54:7,63:8,66:3 optprint # PTR2
# 000E?? 1403?? Units/PTR2.txt nofcbcheck? # PTR2
# 000E?? 1403?? Units/PTR2.txt???????????? # PTR2
```
The three printers are those I tried.? I do not have set up a 3211 at 0000.
The error I get in the first two cases are about the unsupported fcb/nofcbcheck
I am pretty sure I did start HASP in the last test.
All the best
Marco
|
Re: Printer does work on SDL-HYPERION-4.2.x
Marco, My Hercules MTS config file contains :- 0000 3211 "C:\Users\Public\Hercules\MTS-D6-0A\Prt/PTR1x.txt" lpp=66 fcb=1:4,2:34,3:66,4:19,5:5,6:14,7:54,8:63,10:8,11:1 # PTR1, a 3270 data stream printer 000E 1403 "C:\Users\Public\Hercules\MTS-D6-0A\Prt/PTR2x.txt" lpp=66 cctape=(4=1,34=2,66=3,19=4,5=5,14=6,54=7,63=8,8=10,1=11) # PTR2 This is for Hercules version 4.2.1.9826-SDL-g9b5cbe3d (4.2.1.9826) Both printers appear on the Hercules console. Chris -- <cjar1950@...> ---------------------------------------------------------------------------------------------------------------------------------- On Fri, 16 Oct 2020 15:27:36 -0700 marcoxa@... wrote: Hi
I am trying to get MTS working on SDL-HYPERION-4.2.x (on Windows 10).
I changed the entry for PRT2 to what is suggested for D6.0a, but I get an error starting up and the printer is not started (it does not appear in the Hercules panel).
Using the option with nofcbcheck does not seem to start the printer at all (error at startup and no printer in the Hercules panel).
The third option seems to add the printer, but then
# list myfile *print*
"*PRINT*": is not available.
Enter a new file/device name, "CANCEL", or "HELP".
? cancel
#
Is it this something that should be "fixed" in SDL-HYPERION-4.2.x?
Or there is some other incantation I could try in the configuration file or in the MTS setup?
Thanks
Marco
|
Re: Printer does work on SDL-HYPERION-4.2.x
I've never tried MTS in that version of Hercules, or any version running on Windows. There are many versions of Hercules out there and it's likely that not all of them contain the changes necessary to avoid problems in MTS, but I don't know the details.
However, the error message about PRINT has nothing to do with this. It would imply that you don't have HASP properly started. There should be something in the D6.0 notes about this. I think that you need to run "MTS *HSP" on the console to start it. You may need to also start the main HASP job using "HASP" on the console if it's not already running.
Mike
toggle quoted message
Show quoted text
On 16 Oct 2020, at 18:27, marcoxa@... wrote:
Hi
I am trying to get MTS working on SDL-HYPERION-4.2.x (on Windows 10).
I changed the entry for PRT2 to what is suggested for D6.0a, but I get an error starting up and the printer is not started (it does not appear in the Hercules panel).
Using the option with nofcbcheck does not seem to start the printer at all (error at startup and no printer in the Hercules panel).
The third option seems to add the printer, but then
# list myfile *print*
"*PRINT*": is not available.
Enter a new file/device name, "CANCEL", or "HELP".
? cancel
#
Is it this something that should be "fixed" in SDL-HYPERION-4.2.x?
Or there is some other incantation I could try in the configuration file or in the MTS setup?
Thanks
Marco
|
Printer does work on SDL-HYPERION-4.2.x
Hi I am trying to get MTS working on SDL-HYPERION-4.2.x (on Windows 10). I changed the entry for PRT2 to what is suggested for D6.0a, but I get an error starting up and the printer is not started (it does not appear in the Hercules panel). Using the option with nofcbcheck does not seem to start the printer at all (error at startup and no printer in the Hercules panel). The third option seems to add the printer, but then
# list myfile *print*???????????????????????????? ?
? "*PRINT*": is not available.????????????????????? ?
? Enter a new file/device name, "CANCEL", or "HELP".
? cancel??????????????????????????????????????????? ?
#?????????????????????????????????????????????????? ?
?
Is it this something that should be "fixed" in SDL-HYPERION-4.2.x?
Or there is some other incantation I could try in the configuration file or in the MTS setup?
?
Thanks
?
Marco
???????????????????????????????????????????????
|
Re: 1996 System Release Status?
On Apr 28, 2020, at 9:29 PM, Mike Alexander < mta@...> wrote: On 28 Apr 2020, at 23:33, Ron Frederick wrote:
That’s great to hear, Mike! I think I might have been the author of some of that FTP code, while I was still an undergrad at RPI. It’d be fun to get that working again here at some point. I also did an rlogin implementation and an NFS implementation, though I’m not sure whether that code ever made it to other MTS sites.
I don't remember seeing rlogin or NFS code, but after nearly 25 years I can't be sure. The most likely place to find it would be on the redistribution tapes, but, unfortunately, they were all or mostly lost. [Redistribution tapes were a way for MTS sites to share stuff before the Internet. Everyone would send tapes of interesting stuff to Michigan and we would combine them all onto one tape and send a copy to everyone. There was no attempt to review or integrate anything.]
Yeah - I remember us doing that with FTP, but I’m not sure some of the other Internet apps I wrote ever got shared beyond RPI. Unfortunately, I don’t have any of that code any more. I do still keep in touch with folks at RPI who worked on MTS there, but I’m not sure if anyone preserved any of that code.
|
Re: 1996 System Release Status?
Mike, On Tue, Apr 28, 2020 at 08:22:49PM -0400, Mike Alexander wrote: One good thing is that if we get this out it will likely have TCP/IP
support given the correct updates to Hercules. The version I have
includes an FTP client and server as well as network printing to a
Postscript printer attached to my Mac. This is done by adding emulation
of the network interface devices MTS used for this in the past to
Hercules. It sounds like it will be worth the wait. -- Kevin Bruceville, TX What's the definition of a legacy system? One that works! Errare humanum est, ignoscere caninum.
|
Re: 1996 System Release Status?
On 28 Apr 2020, at 23:33, Ron Frederick wrote:
That’s great to hear, Mike! I think I might have been the author of some of that FTP code, while I was still an undergrad at RPI. It’d be fun to get that working again here at some point. I also did an rlogin implementation and an NFS implementation, though I’m not sure whether that code ever made it to other MTS sites.
I don't remember seeing rlogin or NFS code, but after nearly 25 years I can't be sure. The most likely place to find it would be on the redistribution tapes, but, unfortunately, they were all or mostly lost. [Redistribution tapes were a way for MTS sites to share stuff before the Internet. Everyone would send tapes of interesting stuff to Michigan and we would combine them all onto one tape and send a copy to everyone. There was no attempt to review or integrate anything.]
Mike
|
Marco Antoniotti
Ron Frederick