From Newsgroup: comp.arch

All you need is mov.
All you need is mov.
All you need is mov, mov.
Mov is all you need.

Some more thought will have to go into rewriting the lyrics
of "All You Need Is Love" completely.

So somebody proved that mov on x86 is Turing-complete, and the
same guy who put up the assembly hall of shame wrote a compiler
which translates C into only mov instructions. Some people have
a twisted mind... https://github.com/xoreaxeaxeax/movfuscator
--
This USENET posting was made without artificial intelligence,
artificial impertinence, artificial arrogance, artificial stupidity,
artificial flavorings or artificial colorants.
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.arch

Thomas Koenig wrote:

All you need is mov.
All you need is mov.
All you need is mov, mov.
Mov is all you need.

Some more thought will have to go into rewriting the lyrics
of "All You Need Is Love" completely.

So somebody proved that mov on x86 is Turing-complete, and the
same guy who put up the assembly hall of shame wrote a compiler
which translates C into only mov instructions. Some people have
a twisted mind... https://github.com/xoreaxeaxeax/movfuscator

Any MOVE is Turing Complete, IFF said MOVE has sufficient side-effects!

Remember the Weitek FPU? All the instructions were memory-mapped onto a
range of memory addresses, so that the memory source/destination was the opcode.

Terje
--
- <Terje.Mathisen at tmsw.no>
"almost all programming can be viewed as an exercise in caching"
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.arch

On 10/08/2026 13:51, Terje Mathisen wrote:

Thomas Koenig wrote:

All you need is mov.
All you need is mov.
All you need is mov, mov.
Mov is all you need.

Some more thought will have to go into rewriting the lyrics
of "All You Need Is Love" completely.

So somebody proved that mov on x86 is Turing-complete, and the
same guy who put up the assembly hall of shame wrote a compiler
which translates C into only mov instructions.-a Some people have
a twisted mind... https://github.com/xoreaxeaxeax/movfuscator

Any MOVE is Turing Complete, IFF said MOVE has sufficient side-effects!

Remember the Weitek FPU? All the instructions were memory-mapped onto a range of memory addresses, so that the memory source/destination was the opcode.

There is a commercial microcontroller family called the MAXQ from Maxim
that has only one instruction - a MOV. I don't think it was ever very popular, but you can still buy them (though they are marked "not
recommended for new designs").

--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.arch

On 8/9/2026 7:55 AM, Thomas Koenig wrote:

All you need is mov.
All you need is mov.
All you need is mov, mov.
Mov is all you need.

Some more thought will have to go into rewriting the lyrics
of "All You Need Is Love" completely.

So somebody proved that mov on x86 is Turing-complete, and the
same guy who put up the assembly hall of shame wrote a compiler
which translates C into only mov instructions. Some people have
a twisted mind... https://github.com/xoreaxeaxeax/movfuscator

This seems like another variant of a single instruction set computer.

en.wikipedia.org/wiki/One-instruction_set_computer
--
- Stephen Fuld
(e-mail address disguised to prevent spam)
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.arch

On 10 Aug 2026, Terje Mathisen wrote
(in article <115cdvo$33ljl$1@dont-email.me>):

Thomas Koenig wrote:

All you need is mov.
All you need is mov.
All you need is mov, mov.
Mov is all you need.

Some more thought will have to go into rewriting the lyrics
of "All You Need Is Love" completely.

So somebody proved that mov on x86 is Turing-complete, and the
same guy who put up the assembly hall of shame wrote a compiler
which translates C into only mov instructions. Some people have
a twisted mind... https://github.com/xoreaxeaxeax/movfuscator

Any MOVE is Turing Complete, IFF said MOVE has sufficient side-effects!

Remember the Weitek FPU? All the instructions were memory-mapped onto a
range of memory addresses, so that the memory source/destination was the opcode.

Terje

Turing's ACE design implemented the more complicated operations (e.g. *)
by moving data to/from a pseudo address that veiled a functional unit,
ca. 1946.
--
Bill Findlay

--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.arch

Thomas Koenig <tkoenig@netcologne.de> wrote:

All you need is mov.
All you need is mov.
All you need is mov, mov.
Mov is all you need.

Some more thought will have to go into rewriting the lyrics
of "All You Need Is Love" completely.

So somebody proved that mov on x86 is Turing-complete, and the
same guy who put up the assembly hall of shame wrote a compiler
which translates C into only mov instructions. Some people have
a twisted mind... https://github.com/xoreaxeaxeax/movfuscator

It is clear that sequence of mov-s plus a single goto is Turing
complete. But how to get a loop without cheating and using only
mov? In 16-bit real mode one can use assigment to code segment
register or simply rely on address wraparound. But that is
obsolete. Honest solution would use user level instrutions in
64-bit mode, I do not see how one could get loop here using
only mov.

OTOH 32-bit ARM 'mov' can do assignment to program counter, but
loads and stores to memory are not called 'mov'. However, little
creative renaming of instructions give Turing complete 'mov'.
--
Waldek Hebisch
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.arch

Stephen Fuld <sfuld@alumni.cmu.edu.invalid> wrote:

On 8/9/2026 7:55 AM, Thomas Koenig wrote:

All you need is mov.
All you need is mov.
All you need is mov, mov.
Mov is all you need.

Some more thought will have to go into rewriting the lyrics
of "All You Need Is Love" completely.

So somebody proved that mov on x86 is Turing-complete, and the
same guy who put up the assembly hall of shame wrote a compiler
which translates C into only mov instructions. Some people have
a twisted mind... https://github.com/xoreaxeaxeax/movfuscator

This seems like another variant of a single instruction set computer.

en.wikipedia.org/wiki/One-instruction_set_computer

Only if one views 'mov' as a single instruction. But there are
several different encodings doing somewhat different things. If one
allows complex instructions, then IBM-360 is a single instruction
computer, with single instruction being 'Execute'. Other
instructions may be viewed just as alternative encodings of special
cases of 'Execute', allowing better code density.
--
Waldek Hebisch
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.arch

Thomas Koenig <tkoenig@netcologne.de> wrote:

All you need is mov.
All you need is mov.
All you need is mov, mov.
Mov is all you need.

Some more thought will have to go into rewriting the lyrics
of "All You Need Is Love" completely.

So somebody proved that mov on x86 is Turing-complete, and the
same guy who put up the assembly hall of shame wrote a compiler
which translates C into only mov instructions. Some people have
a twisted mind... https://github.com/xoreaxeaxeax/movfuscator

I remember a joke paper a while back that asserted a simple
1-instruction program was sufficient to test all the circuitry
of a then-modern computer.

It was a mov that moved itself backwards in memory and jumped on itself.

The paper was mostly caveats. Turned out while the hardware
would accept the instruction no assembler available had the syntax
to describe it. Then it turned out no loader available could put
the damn bit-pattern in the last word of memory as required for
the test. And no known computer had a front panel that allowed you
to start a program on the last word of memory.

Other than that it was a great idea. :)

Yeah. PDP-11.
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.arch

On Wed, 12 Aug 2026 23:47:53 -0000 (UTC), R Kym Horsell wrote:

The paper was mostly caveats. Turned out while the hardware would
accept the instruction no assembler available had the syntax to
describe it.

The other caveats may make sense, but that one doesnrCOt. Where was
there an assembler that *didnrCOt* allow the insertion of arbitrary byte/word/whatever values in the instruction stream?
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.arch

Waldek Hebisch <antispam@fricas.org> schrieb:

Thomas Koenig <tkoenig@netcologne.de> wrote:

All you need is mov.
All you need is mov.
All you need is mov, mov.
Mov is all you need.

Some more thought will have to go into rewriting the lyrics
of "All You Need Is Love" completely.

So somebody proved that mov on x86 is Turing-complete, and the
same guy who put up the assembly hall of shame wrote a compiler
which translates C into only mov instructions. Some people have
a twisted mind... https://github.com/xoreaxeaxeax/movfuscator

It is clear that sequence of mov-s plus a single goto is Turing
complete. But how to get a loop without cheating and using only
mov?

Page faults. I did write "twisted mind" above...
--
This USENET posting was made without artificial intelligence,
artificial impertinence, artificial arrogance, artificial stupidity,
artificial flavorings or artificial colorants.
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.arch


antispam@fricas.org (Waldek Hebisch) posted:

Thomas Koenig <tkoenig@netcologne.de> wrote:

All you need is mov.
All you need is mov.
All you need is mov, mov.
Mov is all you need.

Some more thought will have to go into rewriting the lyrics
of "All You Need Is Love" completely.

So somebody proved that mov on x86 is Turing-complete, and the
same guy who put up the assembly hall of shame wrote a compiler
which translates C into only mov instructions. Some people have
a twisted mind... https://github.com/xoreaxeaxeax/movfuscator

It is clear that sequence of mov-s plus a single goto is Turing
complete. But how to get a loop without cheating and using only
mov? In 16-bit real mode one can use assigment to code segment
register or simply rely on address wraparound. But that is
obsolete. Honest solution would use user level instrutions in
64-bit mode, I do not see how one could get loop here using
only mov.

MOV IP,#top-of-loop

OTOH 32-bit ARM 'mov' can do assignment to program counter, but
loads and stores to memory are not called 'mov'. However, little
creative renaming of instructions give Turing complete 'mov'.

--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.arch

MitchAlsup <user5857@newsgrouper.org.invalid> schrieb:

antispam@fricas.org (Waldek Hebisch) posted:

Thomas Koenig <tkoenig@netcologne.de> wrote:

All you need is mov.
All you need is mov.
All you need is mov, mov.
Mov is all you need.

Some more thought will have to go into rewriting the lyrics
of "All You Need Is Love" completely.

So somebody proved that mov on x86 is Turing-complete, and the
same guy who put up the assembly hall of shame wrote a compiler
which translates C into only mov instructions. Some people have
a twisted mind... https://github.com/xoreaxeaxeax/movfuscator

It is clear that sequence of mov-s plus a single goto is Turing
complete. But how to get a loop without cheating and using only
mov? In 16-bit real mode one can use assigment to code segment
register or simply rely on address wraparound. But that is
obsolete. Honest solution would use user level instrutions in
64-bit mode, I do not see how one could get loop here using
only mov.

MOV IP,#top-of-loop

That instruction is not required by the MOV-only compiler (and it
is not called mov in x86 parlance).

It uses page faults on invalid addresses instead.
--
This USENET posting was made without artificial intelligence,
artificial impertinence, artificial arrogance, artificial stupidity,
artificial flavorings or artificial colorants.
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.arch

On 2026-08-10, Stephen Fuld <sfuld@alumni.cmu.edu.invalid> wrote:

On 8/9/2026 7:55 AM, Thomas Koenig wrote:

All you need is mov.
All you need is mov.
All you need is mov, mov.
Mov is all you need.

Some more thought will have to go into rewriting the lyrics
of "All You Need Is Love" completely.

So somebody proved that mov on x86 is Turing-complete, and the
same guy who put up the assembly hall of shame wrote a compiler
which translates C into only mov instructions. Some people have
a twisted mind... https://github.com/xoreaxeaxeax/movfuscator

This seems like another variant of a single instruction set computer.

en.wikipedia.org/wiki/One-instruction_set_computer

I was about to post about Subleq right now...
--- Synchronet 3.22a-Linux NewsLink 1.2