National Semiconductor - NIBL - UK Vintage Radio Repair and Restoration Discussion Forum

Moonferret · 11th Oct 2023, 9:55 am

Hi Folks!

There seems to be a lot of interest here for National Semiconductor's SC/MP processor and NIBL language. I recently picked up a NS IMP-16 computer that came with a preliminary version of NIBL, dated 11-5-76 (Guessing US date format) It appears to be in NS' RLM format and was probably generated by the SC/MP cross assembler on the IMP-16. Sadly, I don't have the source or the cross assembler but I figured this might be of interest to somebody here.

Cheers,
Dave

Phil__G · 11th Oct 2023, 12:32 pm

Hi Moonferret, that is a puzzle, its 8k where NIBL is 4k, and unfortunately the image seems to be corrupted, maybe the dreaded 'bit-rot' or maybe it isnt what it says on the tin, it starts out with 160 hex HALTs, there are a lot of blank 00 areas though I can see a few keywords, some of which arent NIBL as we know it... 'PLIST', 'CLEAR', etc. Not sure what it is, will have a better look later. What type of chip is it, a 2764??
Ta for posting though, its all interesting stuff!

Moonferret · 11th Oct 2023, 12:48 pm

It's a direct read from the paper tape and it does seem correct. However, it is in RLM format so it's going to need a little work to decode to a plain binary file. The additional keywords are interesting though.

Cheers,
Dave

Silvester · 11th Oct 2023, 4:09 pm

Odd format (lots of leading zeros and fields of code) but I have been able to extract something which looks very similar to NIBL on my Elektor SC/MP. The .dis text file is from my own disassembler, .bin file is raw data (page 0).

Realtime · 12th Oct 2023, 5:17 pm

Quote:

Originally Posted by Moonferret

It appears to be in NS' RLM format and was probably generated by the SC/MP cross assembler on the IMP-16.

Is there a definition of RLM format? I've read it means Relocatable Load Module but can't find any detail of how it's constructed (or is it more about how it's represented on paper tape)?

Silvester · 12th Oct 2023, 8:06 pm

FWIW the format of the file appears to be repeated sections of:

<NUL> <STX> <length> <checksum> <data> <CR> <LF>

Strip all leading NUL's (0x00) until STX (0x02) encountered
<length> byte bits 5-0 is size of <data> section (b7/b6 include extra info)
<checksum> byte is sum of bytes in <data> modulo 256

The data section starts with two bytes: high/low for file (address?) position followed by code (usually 28 bytes per full line)

The length byte appears to set bit 6 for normal lines, bit 7 and 6 set for termination/EOF.

Moonferret · 12th Oct 2023, 8:34 pm

The RLM format is described in chapter 6.4 of the IMP-16 programming manual here...

http://www.bitsavers.org/components/...pgmg_Nov73.pdf

Looks like it was originally intended for data stored on punched cards and later adapted for paper tape.

Cheers,
Dave

Phil__G · 12th Oct 2023, 8:38 pm

Quote:

Originally Posted by Phil__G

Hi Moonferret, that is a puzzle, its 8k where NIBL is 4k, and unfortunately the image seems to be corrupted, maybe the dreaded 'bit-rot' or maybe it isnt what it says on the tin, it starts out with 160 hex HALTs, there are a lot of blank 00 areas though I can see a few keywords, some of which arent NIBL as we know it... 'PLIST', 'CLEAR', etc. Not sure what it is, will have a better look later. What type of chip is it, a 2764??
Ta for posting though, its all interesting stuff!

Sorry I had assumed a .bin file was a literal binary from a ROM chip, hence my comments

At 0170 theres a 'start of text' character (02) followed by two 16-bit (record length?) then some SC/MP code - 08 (usual throw-away opcode at zero) C4 10 36 etc... which is indeed NIBL code. Its hard to read manually with all the CR/LF and nulls but it does look like NIBL-P.
Taking the first record, the length (if it is that) 5E D7 00 00 is a peculiar number, whether you take it as big or little endian - you'd expact something magic like 32, 64, 4096 etc.
Peculiar and intriguing...

Michael Haardt · 14th Oct 2023, 11:11 pm

Moonferret: I could recover the source from changing the earliest version I had. Would you mind if I include the image from the papertape in my NIBL collection and if it is ok, give credit to who?

It looks like the variable layout was quite different in this version, there are some bugs and the random number generator worked differently.

Phil__G · 14th Oct 2023, 11:28 pm

Have you decoded the RLM Michael?

Michael Haardt · 14th Oct 2023, 11:37 pm

No, but Silvester did so a few posts ago. I used that binary.

Michael Haardt · 23rd Oct 2023, 8:30 am

I updated nibl.tar.gz at

http://www.moria.de/tech/scmp/software/

to include the restored source for the pre-release. If you are curious what exactly has been changed, now a diff of the source can answer that.

11th Oct 2023, 12:32 pm	#2
Phil__G Heptode Join Date: Mar 2011 Location: North Yorkshire, UK. Posts: 911	Re: National Semiconductor - NIBL Hi Moonferret, that is a puzzle, its 8k where NIBL is 4k, and unfortunately the image seems to be corrupted, maybe the dreaded 'bit-rot' or maybe it isnt what it says on the tin, it starts out with 160 hex HALTs, there are a lot of blank 00 areas though I can see a few keywords, some of which arent NIBL as we know it... 'PLIST', 'CLEAR', etc. Not sure what it is, will have a better look later. What type of chip is it, a 2764?? Ta for posting though, its all interesting stuff! Last edited by Phil__G; 11th Oct 2023 at 12:39 pm.

11th Oct 2023, 12:48 pm	#3
Moonferret Triode Join Date: Dec 2011 Location: Preston, Lancashire, UK. Posts: 11	Re: National Semiconductor - NIBL It's a direct read from the paper tape and it does seem correct. However, it is in RLM format so it's going to need a little work to decode to a plain binary file. The additional keywords are interesting though. Cheers, Dave

12th Oct 2023, 8:06 pm	#6
Silvester Triode Join Date: Feb 2014 Location: Redditch, Worcestershire Posts: 27	Re: National Semiconductor - NIBL FWIW the format of the file appears to be repeated sections of: <NUL> <STX> <length> <checksum> <data> <CR> <LF> Strip all leading NUL's (0x00) until STX (0x02) encountered <length> byte bits 5-0 is size of <data> section (b7/b6 include extra info) <checksum> byte is sum of bytes in <data> modulo 256 The data section starts with two bytes: high/low for file (address?) position followed by code (usually 28 bytes per full line) The length byte appears to set bit 6 for normal lines, bit 7 and 6 set for termination/EOF. __________________ David

12th Oct 2023, 8:34 pm	#7
Moonferret Triode Join Date: Dec 2011 Location: Preston, Lancashire, UK. Posts: 11	Re: National Semiconductor - NIBL The RLM format is described in chapter 6.4 of the IMP-16 programming manual here... http://www.bitsavers.org/components/...pgmg_Nov73.pdf Looks like it was originally intended for data stored on punched cards and later adapted for paper tape. Cheers, Dave

14th Oct 2023, 11:11 pm	#9
Michael Haardt Tetrode Join Date: May 2021 Location: Titz, Germany. Posts: 71	Re: National Semiconductor - NIBL Moonferret: I could recover the source from changing the earliest version I had. Would you mind if I include the image from the papertape in my NIBL collection and if it is ok, give credit to who? It looks like the variable layout was quite different in this version, there are some bugs and the random number generator worked differently.

14th Oct 2023, 11:28 pm	#10
Phil__G Heptode Join Date: Mar 2011 Location: North Yorkshire, UK. Posts: 911	Re: National Semiconductor - NIBL Have you decoded the RLM Michael?

14th Oct 2023, 11:37 pm	#11
Michael Haardt Tetrode Join Date: May 2021 Location: Titz, Germany. Posts: 71	Re: National Semiconductor - NIBL No, but Silvester did so a few posts ago. I used that binary.

23rd Oct 2023, 8:30 am	#12
Michael Haardt Tetrode Join Date: May 2021 Location: Titz, Germany. Posts: 71	Re: National Semiconductor - NIBL I updated nibl.tar.gz at http://www.moria.de/tech/scmp/software/ to include the restored source for the pre-release. If you are curious what exactly has been changed, now a diff of the source can answer that.