paradigm locate 5.0 for 80188: insufficient memory?

Hello,
  I have paradigm locate 5.0 that I use to generate rom image for
80C188XL cpu. it works fine with the demo comming with it (the 'sieve'
demo). However, it does not work with my own files, complaining "error
1004: insufficient memory available for paradigm locate". The compiled
executable is not very big, it is about 38 kBytes. I run locate on the
computer with 128MBytes memory and it gives the same error message. I
uploaded the files to (http://www.mscpscsi.com/otherfiles/locateq.zip).
Please take a look.  Thank you.
  
vax, 9000

"vax9000" <VAX9000@gmail.com> wrote in message 
news:1144266805.158285.203310@i40g2000cwc.googlegroups.com...
> Hello,
>  I have paradigm locate 5.0 that I use to generate rom image for
> 80C188XL cpu. it works fine with the demo comming with it (the 'sieve'
> demo). However, it does not work with my own files, complaining "error
> 1004: insufficient memory available for paradigm locate". The compiled
> executable is not very big, it is about 38 kBytes. I run locate on the
> computer with 128MBytes memory and it gives the same error message. I
> uploaded the files to (http://www.mscpscsi.com/otherfiles/locateq.zip).
> Please take a look.  Thank you.
>
> vax, 9000
>
A brief discussion with a colleague of mine who uses Paradigm 5 on 80188 
says that you probably don't have the .cfg files set up correctly with 
respect to allocated space for ROM, RAM and Code. He didn't go into much 
detail other than to say that Paradigm is a hard road and that, if you 
afford it, you should take your kit to Paradigm and sit with them while they 
go through it with you. 

vax9000 escribi�:

> Hello,
>   I have paradigm locate 5.0 that I use to generate rom image for
> 80C188XL cpu. it works fine with the demo comming with it (the 'sieve'
> demo). However, it does not work with my own files, complaining "error
> 1004: insufficient memory available for paradigm locate". The compiled
> executable is not very big, it is about 38 kBytes. I run locate on the
> computer with 128MBytes memory and it gives the same error message. I
> uploaded the files to (http://www.mscpscsi.com/otherfiles/locateq.zip).
> Please take a look.  Thank you.
>   
> vax, 9000
> 

I used Paradigm locate a thousand years ago, with 80188EB devices. IIRC, 
it was an MS-DOS program. I am not sure, but the message

"error 1004: insufficient memory available for paradigm locate"

reminds me of similar errors reported by MS-DOS based programs in 
computers that have in fact *more* available memory than expected 
(Polytron's PVCS and Make, which I still use, comes to my mind).

The solution to these problems is to *reduce* the amount of "real-mode" 
available memory below 512K.

If you are using a W2K machine, you can do this executing command.com 
(not cmd.exe!) and mem.exe several times until you see that the memory 
available to MS-DOS programs is less than 512 K. Then, launch locate.exe 
and cross your fingers. This has saved my day several times.

Ignacio G.T. wrote:
> vax9000 escribi�:
> 
>> Hello,
>>   I have paradigm locate 5.0 that I use to generate rom image for
>> 80C188XL cpu. it works fine with the demo comming with it (the 'sieve'
>> demo). However, it does not work with my own files, complaining "error
>> 1004: insufficient memory available for paradigm locate". The compiled
>> executable is not very big, it is about 38 kBytes. I run locate on the
>> computer with 128MBytes memory and it gives the same error message. I
>> uploaded the files to (http://www.mscpscsi.com/otherfiles/locateq.zip).
>> Please take a look.  Thank you.
>>   vax, 9000
>>
> 
> I used Paradigm locate a thousand years ago, with 80188EB devices. IIRC, 
> it was an MS-DOS program. I am not sure, but the message
> 
> "error 1004: insufficient memory available for paradigm locate"
> 
> reminds me of similar errors reported by MS-DOS based programs in 
> computers that have in fact *more* available memory than expected 
> (Polytron's PVCS and Make, which I still use, comes to my mind).
> 
> The solution to these problems is to *reduce* the amount of "real-mode" 
> available memory below 512K.
> 
> If you are using a W2K machine, you can do this executing command.com 
> (not cmd.exe!) and mem.exe several times until you see that the memory 
> available to MS-DOS programs is less than 512 K. Then, launch locate.exe 
> and cross your fingers. This has saved my day several times.

If this helps (like it did with the late Wordstar), there's
an addressing blunder in the code: It is calculating space
in paragraphs (1 para = 16 bytes) and the calculation thinks
the size as a signed 16 bit entity.

AFAIK, there's no such blunder in Locate.

The first things to look at are the maps from the
DOS linker and Locate. The third thing would be
comparing the Locate control file with the maps.

There are subtle issues in the location process, as
the groups to be located separately must also have
different starting paragraph addresses to enable
Locate to find them from the .exe file.

-- 

Tauno Voipio
tauno voipio (at) iki fi

Ignacio G.T. wrote:
>...
> The solution to these problems is to *reduce* the amount of "real-mode"
> available memory below 512K.
>
> If you are using a W2K machine, you can do this executing command.com
> (not cmd.exe!) and mem.exe several times until you see that the memory
> available to MS-DOS programs is less than 512 K. Then, launch locate.exe
> and cross your fingers. This has saved my day several times.

Thank you for the tip. However, it does not work with paradigm locate.
I ran command.com several times until the available 'triditioanl'
memory was lower than 512K. But still this error message was printed
when I ran "locate bsimrom". 

vax, 9000

My .cfg file was copied from the "sieve" demo. I modified it a little
bit, from 64KB rom to 128KB rom which I used. The same .cfg file worked
well with "sieve" but broke with my own files. I will double check to
see whether it works with different configuration.

Paradigm may be a hard road. But other roads are harder. I have checked
open watcom image generator and jloc and they both have this or other
problems. I bought Paradigm from ebay with $40 with original disks and
manual. I don't know whether Paradigm would honor my ownership but it
does not hurt to try.

vax, 9000

Yes, it seems the time for me to play with the map file and
configuration file. I have wasted a lot of time on this issue and It
may be easier to write my own locator. I will not go that way until I
am desperate.

vax9000 wrote:
> My .cfg file was copied from the "sieve" demo. I modified it a little
> bit, from 64KB rom to 128KB rom which I used. The same .cfg file worked
> well with "sieve" but broke with my own files. I will double check to
> see whether it works with different configuration.
> 
> Paradigm may be a hard road. But other roads are harder. I have checked
> open watcom image generator and jloc and they both have this or other
> problems. I bought Paradigm from ebay with $40 with original disks and
> manual. I don't know whether Paradigm would honor my ownership but it
> does not hurt to try.
> 
> vax, 9000


Did you take into account that the memory model can need
change when exceeding 64 kbytes in an 80188?

-- 

Tauno Voipio
tauno voipio (at) iki fi

Tauno Voipio wrote:
>
> Did you take into account that the memory model can need
> change when exceeding 64 kbytes in an 80188?

The compiled code size is 38KB. I used the "small" model. I changed the
size specification in the .cfg file back to 64KB from 128KB but still
got the "insufficient memory". I have uploaded the new .cfg file to
www.mscpscsi.com/otherfiles/bsimrom.cfg

OK, I solved the "insufficient memory" problem by deleting in the .cfg
file the line "absfile	axe86" which controls the debug support. Thank
you for your answers.

vax, 9000

> 
> -- 
> 
> Tauno Voipio
> tauno voipio (at) iki fi

This problem is solved by deleting the optional debug support control
line "absfile	axe86	" from the .cfg file. Thank you Tom, Tauno and
Ignacio for your help. 

vax, 9000