wave8_transpose_2_bf1()

FASTLED_FORCE_INLINE FL_IRAM FL_OPTIMIZE_FUNCTION void fl::detail::wave8_transpose_2_bf1	(	const u8	lanes[2],
		u8	W0,
		u8	W1,
		u8	output[2 *sizeof(Wave8Byte)] )

BF1 for 2-lane Wave8.

Output: 8 symbols × 2 bytes = 16 bytes. Each byte packs 4 pulses × 2 lanes bit-interleaved (lane 1 in even bit positions, lane 0 in odd). High byte (output[s*2 + 0]) holds bf1-pulse indices 0..3 with q=0 → pulse 3 and q=3 → pulse 0; low byte holds bf1-pulse indices 4..7 with q=0 → pulse 7 and q=3 → pulse 4. Layout matches wave8_transpose_2's LUT-based packing (see FL_WAVE8_SPREAD_TO_16 / kTranspose4_16_LUT).

Definition at line 417 of file wave8.hpp.

                                                             {
    u8 d_mask[8];
    u8 m0_mask[8];
    const u8 D_byte = W0 ^ W1;
    for (int p = 0; p < 8; ++p) {
        const int shift = 7 - p;
        d_mask[p] = ((D_byte >> shift) & 1) ? 0xFFu : 0x00u;
        m0_mask[p] = ((W0 >> shift) & 1) ? 0xFFu : 0x00u;
    }
    for (int s = 0; s < 8; ++s) {
        const int bit_idx = 7 - s;
        const u8 b0 = static_cast<u8>((lanes[0] >> bit_idx) & 1u);
        const u8 b1 = static_cast<u8>((lanes[1] >> bit_idx) & 1u);
        u8 byte_hi = 0;
        u8 byte_lo = 0;
        for (int q = 0; q < 4; ++q) {
            // Reference `wave8_transpose_2` lands chipset-byte bit (4+q) at
            // high-byte bit position (2*q) for lane 1 (even) / (2*q+1) for
            // lane 0 (odd). Chipset-byte bit (4+q) corresponds to bf1
            // pulse index (3 - q) because bf1 numbers pulse p ↔ chipset bit
            // (7 - p). Similarly the low byte uses chipset bits 0..3, which
            // map to bf1 pulses 7..4.
            const int p_hi = 3 - q;
            const int p_lo = 7 - q;
            const u8 m0_p_hi = static_cast<u8>(m0_mask[p_hi] & 1u);
            const u8 d_p_hi  = static_cast<u8>(d_mask[p_hi] & 1u);
            const u8 m0_p_lo = static_cast<u8>(m0_mask[p_lo] & 1u);
            const u8 d_p_lo  = static_cast<u8>(d_mask[p_lo] & 1u);
            const u8 v0_hi = static_cast<u8>(m0_p_hi ^ (b0 & d_p_hi));
            const u8 v1_hi = static_cast<u8>(m0_p_hi ^ (b1 & d_p_hi));
            const u8 v0_lo = static_cast<u8>(m0_p_lo ^ (b0 & d_p_lo));
            const u8 v1_lo = static_cast<u8>(m0_p_lo ^ (b1 & d_p_lo));
            byte_hi |= static_cast<u8>((v1_hi << (2 * q)) | (v0_hi << (2 * q + 1)));
            byte_lo |= static_cast<u8>((v1_lo << (2 * q)) | (v0_lo << (2 * q + 1)));
        }
        output[s * 2 + 0] = byte_hi;
        output[s * 2 + 1] = byte_lo;
    }
}

References fl::W0, and fl::W1.

Referenced by fl::wave8Transpose_2_bf1().

Here is the caller graph for this function: