Function quantize_row_q5_1_reference

ggml-quants.c:558–600

}

void quantize_row_q5_1_reference(const float * restrict x, block_q5_1 * restrict y, int k) {
    const int qk = QK5_1;

    assert(k % qk == 0);

    const int nb = k / qk;

    for (int i = 0; i < nb; i++) {
        float min = FLT_MAX;
        float max = -FLT_MAX;

        for (int j = 0; j < qk; j++) {
            const float v = x[i*qk + j];

            if (v < min) min = v;
            if (v > max) max = v;
        }

        const float d  = (max - min) / ((1 << 5) - 1);
        const float id = d ? 1.0f/d : 0.0f;

        y[i].d = GGML_FP32_TO_FP16(d);
        y[i].m = GGML_FP32_TO_FP16(min);

        uint32_t qh = 0;

        for (int j = 0; j < qk/2; ++j) {
            const float x0 = (x[i*qk + 0    + j] - min)*id;
            const float x1 = (x[i*qk + qk/2 + j] - min)*id;

            const uint8_t xi0 = (uint8_t)(x0 + 0.5f);
            const uint8_t xi1 = (uint8_t)(x1 + 0.5f);

            y[i].qs[j] = (xi0 & 0x0F) | ((xi1 & 0x0F) << 4);

            // get the 5-th bit and store it in qh at the right position
            qh |= ((xi0 & 0x10u) >> 4) << (j + 0);
            qh |= ((xi1 & 0x10u) >> 4) << (j + qk/2);
        }

        memcpy(&y[i].qh, &qh, sizeof(y[i].qh));
    }
}

void quantize_row_q5_1(const float * restrict x, void * restrict y, int k) {
    quantize_row_q5_1_reference(x, y, k);