1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
/******************************************************************************\
* Authors: Iconoclast *
* Release: 2013.11.26 *
* License: CC0 Public Domain Dedication *
* *
* To the extent possible under law, the author(s) have dedicated all copyright *
* and related and neighboring rights to this software to the public domain *
* worldwide. This software is distributed without any warranty. *
* *
* You should have received a copy of the CC0 Public Domain Dedication along *
* with this software. *
* If not, see <http://creativecommons.org/publicdomain/zero/1.0/>. *
\******************************************************************************/
#include "vu.h"
INLINE static void do_mudl(usf_state_t * state, short* VD, short* VS, short* VT)
{
#ifdef ARCH_MIN_ARM_NEON
uint16x4_t vs_low = vld1_u16((const uint16_t*)VS);
uint16x4_t vs_high = vld1_u16((const uint16_t*)VS+4);
uint16x4_t vt_low = vld1_u16((const uint16_t*)VT);
uint16x4_t vt_high = vld1_u16((const uint16_t*)VT+4);
int16x8_t zero = vdupq_n_s16(0);
uint32x4_t lo = vmull_u16(vs_low, vt_low);
uint32x4_t high = vmull_u16(vs_high, vt_high);
uint16x8_t result = vcombine_u16(vshrn_n_u32(lo,16),vshrn_n_u32(high,16));
vst1q_u16(VACC_L, result);
vst1q_s16(VACC_M, zero);
vst1q_s16(VACC_H, zero);
vector_copy(VD, VACC_L);
return;
#else
register int i;
for (i = 0; i < N; i++)
VACC_L[i] = (unsigned short)(VS[i])*(unsigned short)(VT[i]) >> 16;
for (i = 0; i < N; i++)
VACC_M[i] = 0x0000;
for (i = 0; i < N; i++)
VACC_H[i] = 0x0000;
vector_copy(VD, VACC_L); /* no possibilities to clamp */
return;
#endif
}
static void VMUDL(usf_state_t * state, int vd, int vs, int vt, int e)
{
ALIGNED short ST[N];
SHUFFLE_VECTOR(ST, state->VR[vt], e);
do_mudl(state, state->VR[vd], state->VR[vs], ST);
return;
}