Actual behavior of SEUSSF and SEUSSR

moony · 2021-01-14 00:21

These instructions are documented to relocate and periodically invert the bits of their argument, but their actual behavior is undocumented. Anyone know what algorithm they follow?
If not, posting how they both behave with the attached inputs (which are given in binary) is also welcome.

EDIT: Oop- the inputs are 16bit. You should be able to see how to extend them to 32-bit, though, it's a pretty simple pattern.

Kaio · 2021-01-14 11:40

It's a pity that the documentation is still not complete. But I think you could figure it out by writing some code using PNut and debug. It would be appreciated if you would share your result with others here.

Wuerfel_21 · 2021-01-14 13:14

I'm fairly certain I saw a diagram of what these do at some point... Can't seem to find it right now.

moony · 2021-01-14 13:59

@Kaio my Propeller 2 has not yet arrived. If this isn't answered when it arrives, I'll document it myself.

Wuerfel_21 · 2021-01-14 14:03

Well, I found something. spinsim source code tells us that

static int32_t seuss(int32_t value, int32_t forward)
{
    uint32_t i, x;
    uint8_t bitnum[32] = {
        11,  5, 18, 24, 27, 19, 20, 30, 28, 26, 21, 25,  3,  8,  7, 23,
        13, 12, 16,  2, 15,  1,  9, 31,  0, 29, 17, 10, 14,  4,  6, 22};

    if (forward)
    {
        x = 0x354dae51;
        for (i = 0; i < 32; i++)
        {
            if (value & (1 << i))
                x ^= (1 << bitnum[i]);
        }
    }
    else
    {
        x = 0xeb55032d;
        for (i = 0; i < 32; i++)
        {
            if (value & (1 << bitnum[i]))
                x ^= (1 << i);
        }
    }
    return x;
}

Kaio · 2021-01-14 14:33

Thank you @Wuerfel_21. It looks like that these instructions can be used to encode and decode a value.

Tubular · 2021-01-14 21:45

Wuerfel_21 wrote: »

I'm fairly certain I saw a diagram of what these do at some point... Can't seem to find it right now.

I remember it too, i think maybe @cgracey used it in one of his online video presentations?

I think it had red and green lines showing which bits were inverted and which not, and mapped their source and destination bits

Wuerfel_21 · 2021-01-14 21:48

Tubular wrote: »

I think it had red and green lines showing which bits were inverted and which not, and mapped their source and destination bits

yes, that one.

rogloh · 2021-01-14 22:06

That diagram was probably the one shown in one of my threads here....

https://forums.parallax.com/discussion/comment/1479691/#Comment_1479691

Tubular · 2021-01-14 22:32

Yes!! Collective forumista-brain beats google search once again

Wuerfel_21 · 2021-01-14 22:33

Oh, rogloh-sensei, master of the ancient art of ForumSearch-Fu, I see that I still have much to learn from you.....

Oh well, let's repost it here:

Actual behavior of SEUSSF and SEUSSR

Comments