int main(void)
{
unsigned char state = 0x01; //Set initial state to 1
int input;
DDRB = 0xff; //Set PORTB as output
DDRD = 0x00; //Set PORTD as input
PORTD = 0xff;
PORTB = state; //Set LEDs to initial state
for(;;) {
while ((input = input_check()) == 0){} //Wait for button press
if (input == 5) { //According to the input
state = SW5(state); //do the corresponding job
PORTB = state;
} else if (input == 4) {
state = SW4(state);
PORTB = state;
} else if (input == 3) {
state = SW3(state);
PORTB = state;
} else if (input == 2) {
state = SW2(state);
PORTB = state;
} else {
state = SW1(state);
PORTB = state;
}
}
return -1;
}
static void intra_predict_vert_4x4_msa(const uint8_t *src, uint8_t *dst,
int32_t dst_stride) {
uint32_t src_data;
src_data = LW(src);
SW4(src_data, src_data, src_data, src_data, dst, dst_stride);
}
static void intra_predict_128dc_4x4_msa(uint8_t *dst, int32_t dst_stride) {
uint32_t out;
const v16i8 store = __msa_ldi_b(128);
out = __msa_copy_u_w((v4i32)store, 0);
SW4(out, out, out, out, dst, dst_stride);
}
static WEBP_INLINE void DC4(uint8_t* dst, const uint8_t* top) {
uint32_t dc = 4;
int i;
for (i = 0; i < 4; ++i) dc += top[i] + top[-5 + i];
dc >>= 3;
dc = dc | (dc << 8) | (dc << 16) | (dc << 24);
SW4(dc, dc, dc, dc, dst, BPS);
}
static void intra_predict_horiz_4x4_msa(const uint8_t *src, uint8_t *dst,
int32_t dst_stride) {
uint32_t out0, out1, out2, out3;
out0 = src[0] * 0x01010101;
out1 = src[1] * 0x01010101;
out2 = src[2] * 0x01010101;
out3 = src[3] * 0x01010101;
SW4(out0, out1, out2, out3, dst, dst_stride);
}
static WEBP_INLINE void VE4(uint8_t* dst, const uint8_t* top) { // vertical
const uint64_t val_m = LD(top - 1);
const v16u8 A = (v16u8)__msa_insert_d((v2i64)A, 0, val_m);
const v16u8 B = SLDI_UB(A, A, 1);
const v16u8 C = SLDI_UB(A, A, 2);
const v16u8 AC = __msa_ave_u_b(A, C);
const v16u8 B2 = __msa_ave_u_b(B, B);
const v16u8 R = __msa_aver_u_b(AC, B2);
const uint32_t out = __msa_copy_s_w((v4i32)R, 0);
SW4(out, out, out, out, dst, BPS);
}
static void intra_predict_dc_tl_4x4_msa(const uint8_t *src, uint8_t *dst,
int32_t dst_stride) {
uint32_t val0;
v16i8 store, data = { 0 };
v8u16 sum_h;
v4u32 sum_w;
val0 = LW(src);
data = (v16i8)__msa_insert_w((v4i32)data, 0, val0);
sum_h = __msa_hadd_u_h((v16u8)data, (v16u8)data);
sum_w = __msa_hadd_u_w(sum_h, sum_h);
sum_w = (v4u32)__msa_srari_w((v4i32)sum_w, 2);
store = __msa_splati_b((v16i8)sum_w, 0);
val0 = __msa_copy_u_w((v4i32)store, 0);
SW4(val0, val0, val0, val0, dst, dst_stride);
}
static WEBP_INLINE void LD4(uint8_t* dst, const uint8_t* top) {
const uint64_t val_m = LD(top);
const v16u8 A = (v16u8)__msa_insert_d((v2i64)A, 0, val_m);
const v16u8 B = SLDI_UB(A, A, 1);
const v16u8 C1 = SLDI_UB(A, A, 2);
const v16u8 C = (v16u8)__msa_insert_b((v16i8)C1, 6, top[7]);
const v16u8 AC = __msa_ave_u_b(A, C);
const v16u8 B2 = __msa_ave_u_b(B, B);
const v16u8 R0 = __msa_aver_u_b(AC, B2);
const v16u8 R1 = SLDI_UB(R0, R0, 1);
const v16u8 R2 = SLDI_UB(R1, R1, 1);
const v16u8 R3 = SLDI_UB(R2, R2, 1);
const uint32_t val0 = __msa_copy_s_w((v4i32)R0, 0);
const uint32_t val1 = __msa_copy_s_w((v4i32)R1, 0);
const uint32_t val2 = __msa_copy_s_w((v4i32)R2, 0);
const uint32_t val3 = __msa_copy_s_w((v4i32)R3, 0);
SW4(val0, val1, val2, val3, dst, BPS);
}
static void intra_predict_dc_4x4_msa(const uint8_t *src_top,
const uint8_t *src_left, uint8_t *dst,
int32_t dst_stride) {
uint32_t val0, val1;
v16i8 store, src = { 0 };
v8u16 sum_h;
v4u32 sum_w;
v2u64 sum_d;
val0 = LW(src_top);
val1 = LW(src_left);
INSERT_W2_SB(val0, val1, src);
sum_h = __msa_hadd_u_h((v16u8)src, (v16u8)src);
sum_w = __msa_hadd_u_w(sum_h, sum_h);
sum_d = __msa_hadd_u_d(sum_w, sum_w);
sum_w = (v4u32)__msa_srari_w((v4i32)sum_d, 3);
store = __msa_splati_b((v16i8)sum_w, 0);
val0 = __msa_copy_u_w((v4i32)store, 0);
SW4(val0, val0, val0, val0, dst, dst_stride);
}
