A puzzle for 2013

[dcooper33]

here is another fun one:

Move the value in n7:0 to n7:1 and move the value in n7:1 to n7:0 without using any intermediate register.

Example
before
n7:0 = 105
n7:1 = 93

after
n7:0 = 93
n7:1 = 105

sub n7:0 n7:1 n7:1 // 105 - 93 = 12
sub n7:0 n7:1 n7:0 // 105 - 12 = 93
add n7:0 n7:1 n7:1 // 93 + 12 =105

That was a good one

 

[Eric Nelson]

sub n7:0 n7:1 n7:1 // 105 - 93 = 12
sub n7:0 n7:1 n7:0 // 105 - 12 = 93
add n7:0 n7:1 n7:1 // 93 + 12 =105

That was a good one

Yes, very clever!

A handful of the AD processors have an instruction called SWAP, which allows you to swap data between (consecutive) groups of up to 255 registers. I remember using this instruction to move recipe data around. Handy when you need it. Page 170 in this PDF:

http://www.automationdirect.com/static/manuals/d006userm/ch5.pdf



-Eric

 

[dcooper33]

A handful of the AD processors have an instruction called SWAP, which allows you to swap data between (consecutive) groups of up to 255 registers. I remember using this instruction to move recipe data around. Handy when you need it. Page 170 in this PDF:

http://www.automationdirect.com/static/manuals/d006userm/ch5.pdf



-Eric

Seems like I remember seeing a similar SWAP instruction in the Logix instruction set, but I've never read up on it or used it. If it isn't there, it should be

 

[Eric Nelson]

Seems like I remember seeing a similar SWAP instruction in the Logix instruction set, but I've never read up on it or used it. If it isn't there, it should be

I think SWP just does the high/low byte swap...

In the AD stuff, that instruction is SWAPB



-Eric

 

[TConnolly]

sub n7:0 n7:1 n7:1 // 105 - 93 = 12
sub n7:0 n7:1 n7:0 // 105 - 12 = 93
add n7:0 n7:1 n7:1 // 93 + 12 =105

That was a good one

Another way:
XOR N7:0 N7:1 N7:0
XOR N7:0 N7:1 N7:1
XOR N7:0 N7:1 N7:0
(this method won't have any over/under flow problems)

 

[dcooper33]

Another way:
XOR N7:0 N7:1 N7:0
XOR N7:0 N7:1 N7:1
XOR N7:0 N7:1 N7:0
(this method won't have any over/under flow problems)

Very nice. A little more counter-intuitive, with different intermediate results, but definitely fool-proof