generic-library/webrtc
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Neon version of rftfsub_128()

Browse Source 
The performance gain on a Nexus 7 reported by audioproc is ~3.3%

The output is bit exact.

BUG=3131
TESTED=trybots and manually on N7
R=bjornv@webrtc.org, cd@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/14819004

Patch from Scott LaVarnway <slavarnw@gmail.com>.

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6628 4adac7df-926f-26a2-2b94-8c16560cd09d
This commit is contained in:
bjornv@webrtc.org 2014-07-08 21:12:23 +00:00
parent 55535d4e58
commit e55641d4f7
1 changed files with 88 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

88
webrtc/modules/audio_processing/aec/aec_rdft_neon.c
View File

@ -178,8 +178,96 @@ static void cftmdl_128_neon(float* a) {
}
}
__inline static float32x4_t reverse_order_f32x4(float32x4_t in) {
// A B C D -> C D A B
const float32x4_t rev = vcombine_f32(vget_high_f32(in), vget_low_f32(in));
// C D A B -> D C B A
return vrev64q_f32(rev);
}
static void rftfsub_128_neon(float* a) {
const float* c = rdft_w + 32;
int j1, j2, k1, k2;
float wkr, wki, xr, xi, yr, yi;
const float32x4_t mm_half = vdupq_n_f32(0.5f);
// Vectorized code (four at once).
// Note: commented number are indexes for the first iteration of the loop.
for (j1 = 1, j2 = 2; j2 + 7 < 64; j1 += 4, j2 += 8) {
// Load 'wk'.
const float32x4_t c_j1 = vld1q_f32(&c[j1]); // 1, 2, 3, 4,
const float32x4_t c_k1 = vld1q_f32(&c[29 - j1]); // 28, 29, 30, 31,
const float32x4_t wkrt = vsubq_f32(mm_half, c_k1); // 28, 29, 30, 31,
const float32x4_t wkr_ = reverse_order_f32x4(wkrt);
const float32x4_t wki_ = c_j1; // 1, 2, 3, 4,
// Load and shuffle 'a'.
// 2, 4, 6, 8, 3, 5, 7, 9
float32x4x2_t a_j2_p = vld2q_f32(&a[0 + j2]);
// 120, 122, 124, 126, 121, 123, 125, 127,
const float32x4x2_t k2_0_4 = vld2q_f32(&a[122 - j2]);
// 126, 124, 122, 120
const float32x4_t a_k2_p0 = reverse_order_f32x4(k2_0_4.val[0]);
// 127, 125, 123, 121
const float32x4_t a_k2_p1 = reverse_order_f32x4(k2_0_4.val[1]);
// Calculate 'x'.
const float32x4_t xr_ = vsubq_f32(a_j2_p.val[0], a_k2_p0);
// 2-126, 4-124, 6-122, 8-120,
const float32x4_t xi_ = vaddq_f32(a_j2_p.val[1], a_k2_p1);
// 3-127, 5-125, 7-123, 9-121,
// Calculate product into 'y'.
// yr = wkr * xr - wki * xi;
// yi = wkr * xi + wki * xr;
const float32x4_t a_ = vmulq_f32(wkr_, xr_);
const float32x4_t b_ = vmulq_f32(wki_, xi_);
const float32x4_t c_ = vmulq_f32(wkr_, xi_);
const float32x4_t d_ = vmulq_f32(wki_, xr_);
const float32x4_t yr_ = vsubq_f32(a_, b_); // 2-126, 4-124, 6-122, 8-120,
const float32x4_t yi_ = vaddq_f32(c_, d_); // 3-127, 5-125, 7-123, 9-121,
// Update 'a'.
// a[j2 + 0] -= yr;
// a[j2 + 1] -= yi;
// a[k2 + 0] += yr;
// a[k2 + 1] -= yi;
// 126, 124, 122, 120,
const float32x4_t a_k2_p0n = vaddq_f32(a_k2_p0, yr_);
// 127, 125, 123, 121,
const float32x4_t a_k2_p1n = vsubq_f32(a_k2_p1, yi_);
// Shuffle in right order and store.
const float32x4_t a_k2_p0nr = vrev64q_f32(a_k2_p0n);
const float32x4_t a_k2_p1nr = vrev64q_f32(a_k2_p1n);
// 124, 125, 126, 127, 120, 121, 122, 123
const float32x4x2_t a_k2_n = vzipq_f32(a_k2_p0nr, a_k2_p1nr);
// 2, 4, 6, 8,
a_j2_p.val[0] = vsubq_f32(a_j2_p.val[0], yr_);
// 3, 5, 7, 9,
a_j2_p.val[1] = vsubq_f32(a_j2_p.val[1], yi_);
// 2, 3, 4, 5, 6, 7, 8, 9,
vst2q_f32(&a[0 + j2], a_j2_p);
vst1q_f32(&a[122 - j2], a_k2_n.val[1]);
vst1q_f32(&a[126 - j2], a_k2_n.val[0]);
}
// Scalar code for the remaining items.
for (; j2 < 64; j1 += 1, j2 += 2) {
k2 = 128 - j2;
k1 = 32 - j1;
wkr = 0.5f - c[k1];
wki = c[j1];
xr = a[j2 + 0] - a[k2 + 0];
xi = a[j2 + 1] + a[k2 + 1];
yr = wkr * xr - wki * xi;
yi = wkr * xi + wki * xr;
a[j2 + 0] -= yr;
a[j2 + 1] -= yi;
a[k2 + 0] += yr;
a[k2 + 1] -= yi;
}
}
void aec_rdft_init_neon(void) {
cft1st_128 = cft1st_128_neon;
cftmdl_128 = cftmdl_128_neon;
rftfsub_128 = rftfsub_128_neon;
}