| /dports/devel/clxclient/clxclient-3.9.2/source/ |
| H A D | scroll.cc | 43 _k0 (0), in X_scroll()
166 k -= _k0; in bpress()
169 else _zz = k + _k0; in bpress()
190 if (_k0 + d <= 0) in bmotion()
192 d = -_k0; in bmotion()
265 XDrawLine (dpy (), win (), dgc (), _k0 + 1, _dw, _k0 + 1, _dw + DW - 2); in redraw()
266 XDrawLine (dpy (), win (), dgc (), _k0 + 1, _dw, _k0 + _dk - 1, _dw); in redraw()
269 XDrawLine (dpy (), win (), dgc (), _k0 + _dk, _dw + DW - 1, _k0 + _dk, _dw + 1); in redraw()
279 XDrawLine (dpy (), win (), dgc (), _dw, _k0 + 1, _dw + DW - 2, _k0 + 1); in redraw()
280 XDrawLine (dpy (), win (), dgc (), _dw, _k0 + 1, _dw, _k0 + _dk - 1); in redraw()
[all …]
|
| /dports/audio/zita-at1/zita-at1-0.6.2/source/ |
| H A D | tmeter.cc | 32 _k0 (86), in Tmeter()
61 XPutImage (dpy (), win (), dgc (), _imag1, _k0 - 2, 0, XM + _k0 - 2, YM, 5 + _k1 - _k0, Y1); in expose()
77 XPutImage (dpy (), win (), dgc (), _imag0, _k0 - 2, 0, XM + _k0 - 2, YM, 5 + _k1 - _k0, Y1); in update()
78 _k0 = k0; in update()
80 XPutImage (dpy (), win (), dgc (), _imag1, _k0 - 2, 0, XM + _k0 - 2, YM, 5 + _k1 - _k0, Y1); in update()
|
| /dports/graphics/vapoursynth-waifu2x-ncnn-vulkan/vapoursynth-waifu2x-ncnn-vulkan-r4/deps/ncnn/src/layer/arm/ |
| H A D | convolution_1x1_fp16s.h | 180 float16x8_t _k0 = vld1q_f16(kptr); in conv1x1s1_sgemm_fp16sa_neon() local
189 _sum0 = vfmaq_laneq_f16(_sum0, _p0, _k0, 0); in conv1x1s1_sgemm_fp16sa_neon()
269 float16x8_t _k0 = vld1q_f16(kptr); in conv1x1s1_sgemm_fp16sa_neon() local
316 float16x8_t _k0 = vld1q_f16(kptr); in conv1x1s1_sgemm_fp16sa_neon() local
342 float16x8_t _k0 = vld1q_f16(kptr); in conv1x1s1_sgemm_fp16sa_neon() local
344 _sum0 = vfmaq_f16(_sum0, _k0, _p0); in conv1x1s1_sgemm_fp16sa_neon()
401 float16x8_t _k0 = vld1q_f16(kptr); in conv1x1s1_sgemm_fp16sa_neon() local
420 float16x8_t _k0 = vld1q_dup_f16(kptr); in conv1x1s1_sgemm_fp16sa_neon() local
422 _sum0 = vfmaq_f16(_sum0, _p0, _k0); in conv1x1s1_sgemm_fp16sa_neon()
446 float16x8_t _k0 = vld1q_f16(kptr); in conv1x1s1_sgemm_fp16sa_neon() local
[all …]
|
| H A D | deconvolution_3x3.h | 49 float32x4_t _k0 = vld1q_f32(k0); in deconv3x3s1_neon() local
72 _out00 = vmlaq_lane_f32(_out00, _v, vget_low_f32(_k0), 0); in deconv3x3s1_neon()
74 float32x4_t _out01 = vmulq_lane_f32(_v, vget_low_f32(_k0), 1); in deconv3x3s1_neon()
85 _out00high = vmla_lane_f32(_out00high, vget_low_f32(_v), vget_high_f32(_k0), 0); in deconv3x3s1_neon()
165 _out00 = vmlaq_lane_f32(_out00, _v, vget_low_f32(_k0), 0); in deconv3x3s1_neon()
169 _out01 = vmlaq_lane_f32(_out01, _v, vget_low_f32(_k0), 1); in deconv3x3s1_neon()
173 _out02 = vmlaq_lane_f32(_out02, _v, vget_high_f32(_k0), 0); in deconv3x3s1_neon()
270 float32x4_t _k0 = vld1q_f32(k0); in deconv3x3s2_neon() local
290 float32x4_t _out00 = vmulq_lane_f32(_v, vget_low_f32(_k0), 0); // 0,2,4,6 in deconv3x3s2_neon()
291 float32x4_t _out01 = vmulq_lane_f32(_v, vget_low_f32(_k0), 1); // 1,3,5,7 in deconv3x3s2_neon()
[all …]
|
| H A D | convolution_2x2.h | 55 float32x4_t _k0 = vld1q_f32(kernel0); in conv2x2s1_neon() local
143 "w"(_k0), // %12 in conv2x2s1_neon()
224 "w"(_k0), // %12 in conv2x2s1_neon()
237 float32x4_t _s0s1 = vmulq_f32(_r00r1, _k0); in conv2x2s1_neon()
289 float32x4_t _k0 = vdupq_n_f32(kernel0[0]); in conv2x2s1_neon() local
350 "w"(_k0), // %8 in conv2x2s1_neon()
401 "w"(_k0), // %8 in conv2x2s1_neon()
|
| H A D | convolution_1x1.h | 864 "w"(_k0), // %22 in conv1x1s1_neon()
1022 "w"(_k0), // %16 in conv1x1s1_neon()
1239 "w"(_k0), // %18 in conv1x1s1_neon()
1356 "w"(_k0), // %18 in conv1x1s1_neon()
1474 "w"(_k0), // %12 in conv1x1s1_neon()
1531 "w"(_k0), // %12 in conv1x1s1_neon()
1653 "w"(_k0), // %12 in conv1x1s1_neon()
1700 "w"(_k0), // %12 in conv1x1s1_neon()
1770 "w"(_k0) // %6 in conv1x1s1_neon()
1796 "w"(_k0) // %6 in conv1x1s1_neon()
[all …]
|
| H A D | convolution_3x3_pack8to1_int8.h | 929 int16x8_t _k0 = vld1q_s16(kptr); in conv3x3s1_winograd42_pack8to1_int8_neon() local
931 _sum0 = vmlal_lane_s16(_sum0, vget_low_s16(_r0), vget_low_s16(_k0), 0); in conv3x3s1_winograd42_pack8to1_int8_neon()
933 _sum2 = vmlal_lane_s16(_sum2, vget_low_s16(_r1), vget_low_s16(_k0), 1); in conv3x3s1_winograd42_pack8to1_int8_neon()
935 _sum0 = vmlal_lane_s16(_sum0, vget_low_s16(_r2), vget_low_s16(_k0), 2); in conv3x3s1_winograd42_pack8to1_int8_neon()
937 _sum2 = vmlal_lane_s16(_sum2, vget_low_s16(_r3), vget_low_s16(_k0), 3); in conv3x3s1_winograd42_pack8to1_int8_neon()
980 int16x8_t _k0 = vld1q_s16(kptr); in conv3x3s1_winograd42_pack8to1_int8_neon() local
982 _sum0 = vmlal_lane_s16(_sum0, vget_low_s16(_r0), vget_low_s16(_k0), 0); in conv3x3s1_winograd42_pack8to1_int8_neon()
984 _sum0 = vmlal_lane_s16(_sum0, vget_low_s16(_r1), vget_low_s16(_k0), 2); in conv3x3s1_winograd42_pack8to1_int8_neon()
1017 int16x8_t _k0 = vld1q_s16(kptr); in conv3x3s1_winograd42_pack8to1_int8_neon() local
1019 _sum0 = vmlal_s16(_sum0, vget_low_s16(_r0), vget_low_s16(_k0)); in conv3x3s1_winograd42_pack8to1_int8_neon()
[all …]
|
| /dports/graphics/waifu2x-ncnn-vulkan/waifu2x-ncnn-vulkan-20210521/src/ncnn/src/layer/arm/ |
| H A D | convolution_1x1_fp16s.h | 180 float16x8_t _k0 = vld1q_f16(kptr); in conv1x1s1_sgemm_fp16sa_neon() local
189 _sum0 = vfmaq_laneq_f16(_sum0, _p0, _k0, 0); in conv1x1s1_sgemm_fp16sa_neon()
269 float16x8_t _k0 = vld1q_f16(kptr); in conv1x1s1_sgemm_fp16sa_neon() local
316 float16x8_t _k0 = vld1q_f16(kptr); in conv1x1s1_sgemm_fp16sa_neon() local
342 float16x8_t _k0 = vld1q_f16(kptr); in conv1x1s1_sgemm_fp16sa_neon() local
344 _sum0 = vfmaq_f16(_sum0, _k0, _p0); in conv1x1s1_sgemm_fp16sa_neon()
401 float16x8_t _k0 = vld1q_f16(kptr); in conv1x1s1_sgemm_fp16sa_neon() local
420 float16x8_t _k0 = vld1q_dup_f16(kptr); in conv1x1s1_sgemm_fp16sa_neon() local
422 _sum0 = vfmaq_f16(_sum0, _p0, _k0); in conv1x1s1_sgemm_fp16sa_neon()
446 float16x8_t _k0 = vld1q_f16(kptr); in conv1x1s1_sgemm_fp16sa_neon() local
[all …]
|
| H A D | deconvolution_3x3.h | 49 float32x4_t _k0 = vld1q_f32(k0); in deconv3x3s1_neon() local
72 _out00 = vmlaq_lane_f32(_out00, _v, vget_low_f32(_k0), 0); in deconv3x3s1_neon()
74 float32x4_t _out01 = vmulq_lane_f32(_v, vget_low_f32(_k0), 1); in deconv3x3s1_neon()
85 _out00high = vmla_lane_f32(_out00high, vget_low_f32(_v), vget_high_f32(_k0), 0); in deconv3x3s1_neon()
165 _out00 = vmlaq_lane_f32(_out00, _v, vget_low_f32(_k0), 0); in deconv3x3s1_neon()
169 _out01 = vmlaq_lane_f32(_out01, _v, vget_low_f32(_k0), 1); in deconv3x3s1_neon()
173 _out02 = vmlaq_lane_f32(_out02, _v, vget_high_f32(_k0), 0); in deconv3x3s1_neon()
270 float32x4_t _k0 = vld1q_f32(k0); in deconv3x3s2_neon() local
290 float32x4_t _out00 = vmulq_lane_f32(_v, vget_low_f32(_k0), 0); // 0,2,4,6 in deconv3x3s2_neon()
291 float32x4_t _out01 = vmulq_lane_f32(_v, vget_low_f32(_k0), 1); // 1,3,5,7 in deconv3x3s2_neon()
[all …]
|
| H A D | convolution_2x2.h | 55 float32x4_t _k0 = vld1q_f32(kernel0); in conv2x2s1_neon() local
143 "w"(_k0), // %12 in conv2x2s1_neon()
224 "w"(_k0), // %12 in conv2x2s1_neon()
237 float32x4_t _s0s1 = vmulq_f32(_r00r1, _k0); in conv2x2s1_neon()
289 float32x4_t _k0 = vdupq_n_f32(kernel0[0]); in conv2x2s1_neon() local
350 "w"(_k0), // %8 in conv2x2s1_neon()
401 "w"(_k0), // %8 in conv2x2s1_neon()
|
| H A D | convolution_1x1.h | 864 "w"(_k0), // %22 in conv1x1s1_neon()
1022 "w"(_k0), // %16 in conv1x1s1_neon()
1239 "w"(_k0), // %18 in conv1x1s1_neon()
1356 "w"(_k0), // %18 in conv1x1s1_neon()
1474 "w"(_k0), // %12 in conv1x1s1_neon()
1531 "w"(_k0), // %12 in conv1x1s1_neon()
1653 "w"(_k0), // %12 in conv1x1s1_neon()
1700 "w"(_k0), // %12 in conv1x1s1_neon()
1770 "w"(_k0) // %6 in conv1x1s1_neon()
1796 "w"(_k0) // %6 in conv1x1s1_neon()
[all …]
|
| H A D | convolution_3x3_pack8to1_int8.h | 929 int16x8_t _k0 = vld1q_s16(kptr); in conv3x3s1_winograd42_pack8to1_int8_neon() local
931 _sum0 = vmlal_lane_s16(_sum0, vget_low_s16(_r0), vget_low_s16(_k0), 0); in conv3x3s1_winograd42_pack8to1_int8_neon()
933 _sum2 = vmlal_lane_s16(_sum2, vget_low_s16(_r1), vget_low_s16(_k0), 1); in conv3x3s1_winograd42_pack8to1_int8_neon()
935 _sum0 = vmlal_lane_s16(_sum0, vget_low_s16(_r2), vget_low_s16(_k0), 2); in conv3x3s1_winograd42_pack8to1_int8_neon()
937 _sum2 = vmlal_lane_s16(_sum2, vget_low_s16(_r3), vget_low_s16(_k0), 3); in conv3x3s1_winograd42_pack8to1_int8_neon()
980 int16x8_t _k0 = vld1q_s16(kptr); in conv3x3s1_winograd42_pack8to1_int8_neon() local
982 _sum0 = vmlal_lane_s16(_sum0, vget_low_s16(_r0), vget_low_s16(_k0), 0); in conv3x3s1_winograd42_pack8to1_int8_neon()
984 _sum0 = vmlal_lane_s16(_sum0, vget_low_s16(_r1), vget_low_s16(_k0), 2); in conv3x3s1_winograd42_pack8to1_int8_neon()
1017 int16x8_t _k0 = vld1q_s16(kptr); in conv3x3s1_winograd42_pack8to1_int8_neon() local
1019 _sum0 = vmlal_s16(_sum0, vget_low_s16(_r0), vget_low_s16(_k0)); in conv3x3s1_winograd42_pack8to1_int8_neon()
[all …]
|
| /dports/benchmarks/vkpeak/vkpeak-20210430/ncnn/src/layer/arm/ |
| H A D | convolution_1x1_fp16s.h | 180 float16x8_t _k0 = vld1q_f16(kptr); in conv1x1s1_sgemm_fp16sa_neon() local
189 _sum0 = vfmaq_laneq_f16(_sum0, _p0, _k0, 0); in conv1x1s1_sgemm_fp16sa_neon()
269 float16x8_t _k0 = vld1q_f16(kptr); in conv1x1s1_sgemm_fp16sa_neon() local
316 float16x8_t _k0 = vld1q_f16(kptr); in conv1x1s1_sgemm_fp16sa_neon() local
342 float16x8_t _k0 = vld1q_f16(kptr); in conv1x1s1_sgemm_fp16sa_neon() local
344 _sum0 = vfmaq_f16(_sum0, _k0, _p0); in conv1x1s1_sgemm_fp16sa_neon()
401 float16x8_t _k0 = vld1q_f16(kptr); in conv1x1s1_sgemm_fp16sa_neon() local
420 float16x8_t _k0 = vld1q_dup_f16(kptr); in conv1x1s1_sgemm_fp16sa_neon() local
422 _sum0 = vfmaq_f16(_sum0, _p0, _k0); in conv1x1s1_sgemm_fp16sa_neon()
446 float16x8_t _k0 = vld1q_f16(kptr); in conv1x1s1_sgemm_fp16sa_neon() local
[all …]
|
| H A D | deconvolution_3x3.h | 49 float32x4_t _k0 = vld1q_f32(k0); in deconv3x3s1_neon() local
72 _out00 = vmlaq_lane_f32(_out00, _v, vget_low_f32(_k0), 0); in deconv3x3s1_neon()
74 float32x4_t _out01 = vmulq_lane_f32(_v, vget_low_f32(_k0), 1); in deconv3x3s1_neon()
85 _out00high = vmla_lane_f32(_out00high, vget_low_f32(_v), vget_high_f32(_k0), 0); in deconv3x3s1_neon()
165 _out00 = vmlaq_lane_f32(_out00, _v, vget_low_f32(_k0), 0); in deconv3x3s1_neon()
169 _out01 = vmlaq_lane_f32(_out01, _v, vget_low_f32(_k0), 1); in deconv3x3s1_neon()
173 _out02 = vmlaq_lane_f32(_out02, _v, vget_high_f32(_k0), 0); in deconv3x3s1_neon()
270 float32x4_t _k0 = vld1q_f32(k0); in deconv3x3s2_neon() local
290 float32x4_t _out00 = vmulq_lane_f32(_v, vget_low_f32(_k0), 0); // 0,2,4,6 in deconv3x3s2_neon()
291 float32x4_t _out01 = vmulq_lane_f32(_v, vget_low_f32(_k0), 1); // 1,3,5,7 in deconv3x3s2_neon()
[all …]
|
| H A D | convolution_2x2.h | 55 float32x4_t _k0 = vld1q_f32(kernel0); in conv2x2s1_neon() local
143 "w"(_k0), // %12 in conv2x2s1_neon()
224 "w"(_k0), // %12 in conv2x2s1_neon()
237 float32x4_t _s0s1 = vmulq_f32(_r00r1, _k0); in conv2x2s1_neon()
289 float32x4_t _k0 = vdupq_n_f32(kernel0[0]); in conv2x2s1_neon() local
350 "w"(_k0), // %8 in conv2x2s1_neon()
401 "w"(_k0), // %8 in conv2x2s1_neon()
|
| H A D | convolution_1x1.h | 864 "w"(_k0), // %22 in conv1x1s1_neon()
1022 "w"(_k0), // %16 in conv1x1s1_neon()
1239 "w"(_k0), // %18 in conv1x1s1_neon()
1356 "w"(_k0), // %18 in conv1x1s1_neon()
1474 "w"(_k0), // %12 in conv1x1s1_neon()
1531 "w"(_k0), // %12 in conv1x1s1_neon()
1653 "w"(_k0), // %12 in conv1x1s1_neon()
1700 "w"(_k0), // %12 in conv1x1s1_neon()
1770 "w"(_k0) // %6 in conv1x1s1_neon()
1796 "w"(_k0) // %6 in conv1x1s1_neon()
[all …]
|
| H A D | convolution_3x3_pack8to1_int8.h | 929 int16x8_t _k0 = vld1q_s16(kptr); in conv3x3s1_winograd42_pack8to1_int8_neon() local
931 _sum0 = vmlal_lane_s16(_sum0, vget_low_s16(_r0), vget_low_s16(_k0), 0); in conv3x3s1_winograd42_pack8to1_int8_neon()
933 _sum2 = vmlal_lane_s16(_sum2, vget_low_s16(_r1), vget_low_s16(_k0), 1); in conv3x3s1_winograd42_pack8to1_int8_neon()
935 _sum0 = vmlal_lane_s16(_sum0, vget_low_s16(_r2), vget_low_s16(_k0), 2); in conv3x3s1_winograd42_pack8to1_int8_neon()
937 _sum2 = vmlal_lane_s16(_sum2, vget_low_s16(_r3), vget_low_s16(_k0), 3); in conv3x3s1_winograd42_pack8to1_int8_neon()
980 int16x8_t _k0 = vld1q_s16(kptr); in conv3x3s1_winograd42_pack8to1_int8_neon() local
982 _sum0 = vmlal_lane_s16(_sum0, vget_low_s16(_r0), vget_low_s16(_k0), 0); in conv3x3s1_winograd42_pack8to1_int8_neon()
984 _sum0 = vmlal_lane_s16(_sum0, vget_low_s16(_r1), vget_low_s16(_k0), 2); in conv3x3s1_winograd42_pack8to1_int8_neon()
1017 int16x8_t _k0 = vld1q_s16(kptr); in conv3x3s1_winograd42_pack8to1_int8_neon() local
1019 _sum0 = vmlal_s16(_sum0, vget_low_s16(_r0), vget_low_s16(_k0)); in conv3x3s1_winograd42_pack8to1_int8_neon()
[all …]
|
| /dports/misc/ncnn/ncnn-20211208/src/layer/arm/ |
| H A D | convolution_1x1_fp16s.h | 180 float16x8_t _k0 = vld1q_f16(kptr); in conv1x1s1_sgemm_fp16sa_neon() local
189 _sum0 = vfmaq_laneq_f16(_sum0, _p0, _k0, 0); in conv1x1s1_sgemm_fp16sa_neon()
269 float16x8_t _k0 = vld1q_f16(kptr); in conv1x1s1_sgemm_fp16sa_neon() local
316 float16x8_t _k0 = vld1q_f16(kptr); in conv1x1s1_sgemm_fp16sa_neon() local
342 float16x8_t _k0 = vld1q_f16(kptr); in conv1x1s1_sgemm_fp16sa_neon() local
344 _sum0 = vfmaq_f16(_sum0, _k0, _p0); in conv1x1s1_sgemm_fp16sa_neon()
401 float16x8_t _k0 = vld1q_f16(kptr); in conv1x1s1_sgemm_fp16sa_neon() local
420 float16x8_t _k0 = vld1q_dup_f16(kptr); in conv1x1s1_sgemm_fp16sa_neon() local
422 _sum0 = vfmaq_f16(_sum0, _p0, _k0); in conv1x1s1_sgemm_fp16sa_neon()
446 float16x8_t _k0 = vld1q_f16(kptr); in conv1x1s1_sgemm_fp16sa_neon() local
[all …]
|
| H A D | deconvolution_3x3.h | 49 float32x4_t _k0 = vld1q_f32(k0); in deconv3x3s1_neon() local
72 _out00 = vmlaq_lane_f32(_out00, _v, vget_low_f32(_k0), 0); in deconv3x3s1_neon()
74 float32x4_t _out01 = vmulq_lane_f32(_v, vget_low_f32(_k0), 1); in deconv3x3s1_neon()
85 _out00high = vmla_lane_f32(_out00high, vget_low_f32(_v), vget_high_f32(_k0), 0); in deconv3x3s1_neon()
165 _out00 = vmlaq_lane_f32(_out00, _v, vget_low_f32(_k0), 0); in deconv3x3s1_neon()
169 _out01 = vmlaq_lane_f32(_out01, _v, vget_low_f32(_k0), 1); in deconv3x3s1_neon()
173 _out02 = vmlaq_lane_f32(_out02, _v, vget_high_f32(_k0), 0); in deconv3x3s1_neon()
270 float32x4_t _k0 = vld1q_f32(k0); in deconv3x3s2_neon() local
290 float32x4_t _out00 = vmulq_lane_f32(_v, vget_low_f32(_k0), 0); // 0,2,4,6 in deconv3x3s2_neon()
291 float32x4_t _out01 = vmulq_lane_f32(_v, vget_low_f32(_k0), 1); // 1,3,5,7 in deconv3x3s2_neon()
[all …]
|
| H A D | convolution_2x2.h | 55 float32x4_t _k0 = vld1q_f32(kernel0); in conv2x2s1_neon() local
143 "w"(_k0), // %12 in conv2x2s1_neon()
224 "w"(_k0), // %12 in conv2x2s1_neon()
237 float32x4_t _s0s1 = vmulq_f32(_r00r1, _k0); in conv2x2s1_neon()
289 float32x4_t _k0 = vdupq_n_f32(kernel0[0]); in conv2x2s1_neon() local
350 "w"(_k0), // %8 in conv2x2s1_neon()
401 "w"(_k0), // %8 in conv2x2s1_neon()
|
| H A D | convolution_1x1.h | 864 "w"(_k0), // %22 in conv1x1s1_neon()
1022 "w"(_k0), // %16 in conv1x1s1_neon()
1239 "w"(_k0), // %18 in conv1x1s1_neon()
1356 "w"(_k0), // %18 in conv1x1s1_neon()
1474 "w"(_k0), // %12 in conv1x1s1_neon()
1531 "w"(_k0), // %12 in conv1x1s1_neon()
1653 "w"(_k0), // %12 in conv1x1s1_neon()
1700 "w"(_k0), // %12 in conv1x1s1_neon()
1770 "w"(_k0) // %6 in conv1x1s1_neon()
1796 "w"(_k0) // %6 in conv1x1s1_neon()
[all …]
|
| /dports/graphics/realsr-ncnn-vulkan/realsr-ncnn-vulkan-20210210/src/ncnn/src/layer/arm/ |
| H A D | convolution_1x1_fp16s.h | 180 float16x8_t _k0 = vld1q_f16(kptr); in conv1x1s1_sgemm_fp16sa_neon() local
189 _sum0 = vfmaq_laneq_f16(_sum0, _p0, _k0, 0); in conv1x1s1_sgemm_fp16sa_neon()
269 float16x8_t _k0 = vld1q_f16(kptr); in conv1x1s1_sgemm_fp16sa_neon() local
316 float16x8_t _k0 = vld1q_f16(kptr); in conv1x1s1_sgemm_fp16sa_neon() local
342 float16x8_t _k0 = vld1q_f16(kptr); in conv1x1s1_sgemm_fp16sa_neon() local
344 _sum0 = vfmaq_f16(_sum0, _k0, _p0); in conv1x1s1_sgemm_fp16sa_neon()
401 float16x8_t _k0 = vld1q_f16(kptr); in conv1x1s1_sgemm_fp16sa_neon() local
420 float16x8_t _k0 = vld1q_dup_f16(kptr); in conv1x1s1_sgemm_fp16sa_neon() local
422 _sum0 = vfmaq_f16(_sum0, _p0, _k0); in conv1x1s1_sgemm_fp16sa_neon()
446 float16x8_t _k0 = vld1q_f16(kptr); in conv1x1s1_sgemm_fp16sa_neon() local
[all …]
|
| H A D | deconvolution_3x3.h | 49 float32x4_t _k0 = vld1q_f32(k0); in deconv3x3s1_neon() local
72 _out00 = vmlaq_lane_f32(_out00, _v, vget_low_f32(_k0), 0); in deconv3x3s1_neon()
74 float32x4_t _out01 = vmulq_lane_f32(_v, vget_low_f32(_k0), 1); in deconv3x3s1_neon()
85 _out00high = vmla_lane_f32(_out00high, vget_low_f32(_v), vget_high_f32(_k0), 0); in deconv3x3s1_neon()
165 _out00 = vmlaq_lane_f32(_out00, _v, vget_low_f32(_k0), 0); in deconv3x3s1_neon()
169 _out01 = vmlaq_lane_f32(_out01, _v, vget_low_f32(_k0), 1); in deconv3x3s1_neon()
173 _out02 = vmlaq_lane_f32(_out02, _v, vget_high_f32(_k0), 0); in deconv3x3s1_neon()
270 float32x4_t _k0 = vld1q_f32(k0); in deconv3x3s2_neon() local
290 float32x4_t _out00 = vmulq_lane_f32(_v, vget_low_f32(_k0), 0); // 0,2,4,6 in deconv3x3s2_neon()
291 float32x4_t _out01 = vmulq_lane_f32(_v, vget_low_f32(_k0), 1); // 1,3,5,7 in deconv3x3s2_neon()
[all …]
|
| H A D | convolution_2x2.h | 55 float32x4_t _k0 = vld1q_f32(kernel0); in conv2x2s1_neon() local
143 "w"(_k0), // %12 in conv2x2s1_neon()
224 "w"(_k0), // %12 in conv2x2s1_neon()
237 float32x4_t _s0s1 = vmulq_f32(_r00r1, _k0); in conv2x2s1_neon()
289 float32x4_t _k0 = vdupq_n_f32(kernel0[0]); in conv2x2s1_neon() local
350 "w"(_k0), // %8 in conv2x2s1_neon()
401 "w"(_k0), // %8 in conv2x2s1_neon()
|
| /dports/astro/geographiclib/GeographicLib-1.52/src/ |
| H A D | PolarStereographic.cpp | 23 , _k0(k0) in PolarStereographic()
29 if (!(isfinite(_k0) && _k0 > 0)) in PolarStereographic()
72 rho *= 2 * _k0 * _a / _c; in Forward()
74 _k0; in Forward()
86 t = rho != 0 ? rho / (2 * _k0 * _a / _c) : in Reverse()
92 _k0; in Reverse()
104 _k0 = 1; in SetScale()
106 _k0 *= k/kold; in SetScale()
|