| /external/XNNPACK/src/f32-vsigmoid/gen/ |
| D | vsigmoid-scalar-rr2-lut2048-p1-div-x2.c | 76 float vf1 = vy1 / vd1; in xnn_f32_vsigmoid_ukernel__scalar_rr2_lut2048_p1_div_x2() local
82 vf1 = 0.0f; in xnn_f32_vsigmoid_ukernel__scalar_rr2_lut2048_p1_div_x2()
89 vf1 = vone - vf1; in xnn_f32_vsigmoid_ukernel__scalar_rr2_lut2048_p1_div_x2()
93 y[1] = vf1; in xnn_f32_vsigmoid_ukernel__scalar_rr2_lut2048_p1_div_x2()
|
| D | vsigmoid-scalar-rr2-lut64-p2-div-x2.c | 79 float vf1 = vy1 / vd1; in xnn_f32_vsigmoid_ukernel__scalar_rr2_lut64_p2_div_x2() local
85 vf1 = 0.0f; in xnn_f32_vsigmoid_ukernel__scalar_rr2_lut64_p2_div_x2()
92 vf1 = vone - vf1; in xnn_f32_vsigmoid_ukernel__scalar_rr2_lut64_p2_div_x2()
96 y[1] = vf1; in xnn_f32_vsigmoid_ukernel__scalar_rr2_lut64_p2_div_x2()
|
| D | vsigmoid-scalar-rr2-p5-div-x2.c | 83 float vf1 = ve1 / vd1; in xnn_f32_vsigmoid_ukernel__scalar_rr2_p5_div_x2() local
89 vf1 = 0.0f; in xnn_f32_vsigmoid_ukernel__scalar_rr2_p5_div_x2()
96 vf1 = vone - vf1; in xnn_f32_vsigmoid_ukernel__scalar_rr2_p5_div_x2()
100 y[1] = vf1; in xnn_f32_vsigmoid_ukernel__scalar_rr2_p5_div_x2()
|
| D | vsigmoid-avx2-rr1-p5-div-x16.c | 80 __m256 vf1 = _mm256_div_ps(ve1, vd1); in xnn_f32_vsigmoid_ukernel__avx2_rr1_p5_div_x16() local
83 vf1 = _mm256_andnot_ps(_mm256_cmp_ps(vz1, vdenorm_cutoff, _CMP_LT_OS), vf1); in xnn_f32_vsigmoid_ukernel__avx2_rr1_p5_div_x16()
86 vf1 = _mm256_blendv_ps(_mm256_sub_ps(vone, vf1), vf1, vx1); in xnn_f32_vsigmoid_ukernel__avx2_rr1_p5_div_x16()
89 _mm256_storeu_ps(y + 8, vf1); in xnn_f32_vsigmoid_ukernel__avx2_rr1_p5_div_x16()
|
| D | vsigmoid-avx2-rr1-p5-nr1fma-x16.c | 87 __m256 vf1 = _mm256_mul_ps(ve1, vr1); in xnn_f32_vsigmoid_ukernel__avx2_rr1_p5_nr1fma_x16() local
90 vf1 = _mm256_andnot_ps(_mm256_cmp_ps(vz1, vdenorm_cutoff, _CMP_LT_OS), vf1); in xnn_f32_vsigmoid_ukernel__avx2_rr1_p5_nr1fma_x16()
93 vf1 = _mm256_blendv_ps(_mm256_sub_ps(vone, vf1), vf1, vx1); in xnn_f32_vsigmoid_ukernel__avx2_rr1_p5_nr1fma_x16()
96 _mm256_storeu_ps(y + 8, vf1); in xnn_f32_vsigmoid_ukernel__avx2_rr1_p5_nr1fma_x16()
|
| D | vsigmoid-avx2-rr1-p5-nr2fma-x16.c | 89 __m256 vf1 = _mm256_mul_ps(ve1, vr1); in xnn_f32_vsigmoid_ukernel__avx2_rr1_p5_nr2fma_x16() local
92 vf1 = _mm256_andnot_ps(_mm256_cmp_ps(vz1, vdenorm_cutoff, _CMP_LT_OS), vf1); in xnn_f32_vsigmoid_ukernel__avx2_rr1_p5_nr2fma_x16()
95 vf1 = _mm256_blendv_ps(_mm256_sub_ps(vone, vf1), vf1, vx1); in xnn_f32_vsigmoid_ukernel__avx2_rr1_p5_nr2fma_x16()
98 _mm256_storeu_ps(y + 8, vf1); in xnn_f32_vsigmoid_ukernel__avx2_rr1_p5_nr2fma_x16()
|
| D | vsigmoid-avx-rr2-p5-div-x16.c | 88 __m256 vf1 = _mm256_div_ps(ve1, vd1); in xnn_f32_vsigmoid_ukernel__avx_rr2_p5_div_x16() local
91 vf1 = _mm256_andnot_ps(_mm256_cmp_ps(vz1, vdenorm_cutoff, _CMP_LT_OS), vf1); in xnn_f32_vsigmoid_ukernel__avx_rr2_p5_div_x16()
94 vf1 = _mm256_blendv_ps(_mm256_sub_ps(vone, vf1), vf1, vx1); in xnn_f32_vsigmoid_ukernel__avx_rr2_p5_div_x16()
97 _mm256_storeu_ps(y + 8, vf1); in xnn_f32_vsigmoid_ukernel__avx_rr2_p5_div_x16()
|
| D | vsigmoid-avx2-rr1-p5-div-x24.c | 93 __m256 vf1 = _mm256_div_ps(ve1, vd1); in xnn_f32_vsigmoid_ukernel__avx2_rr1_p5_div_x24() local
97 vf1 = _mm256_andnot_ps(_mm256_cmp_ps(vz1, vdenorm_cutoff, _CMP_LT_OS), vf1); in xnn_f32_vsigmoid_ukernel__avx2_rr1_p5_div_x24()
101 vf1 = _mm256_blendv_ps(_mm256_sub_ps(vone, vf1), vf1, vx1); in xnn_f32_vsigmoid_ukernel__avx2_rr1_p5_div_x24()
105 _mm256_storeu_ps(y + 8, vf1); in xnn_f32_vsigmoid_ukernel__avx2_rr1_p5_div_x24()
|
| D | vsigmoid-scalar-rr2-lut2048-p1-div-x4.c | 100 float vf1 = vy1 / vd1; in xnn_f32_vsigmoid_ukernel__scalar_rr2_lut2048_p1_div_x4() local
108 vf1 = 0.0f; in xnn_f32_vsigmoid_ukernel__scalar_rr2_lut2048_p1_div_x4()
121 vf1 = vone - vf1; in xnn_f32_vsigmoid_ukernel__scalar_rr2_lut2048_p1_div_x4()
131 y[1] = vf1; in xnn_f32_vsigmoid_ukernel__scalar_rr2_lut2048_p1_div_x4()
|
| /external/XNNPACK/src/f16-vsigmoid/gen/ |
| D | vsigmoid-avx2-rr1-p2-div-x16.c | 71 __m256 vf1 = _mm256_div_ps(ve1, vd1); in xnn_f16_vsigmoid_ukernel__avx2_rr1_p2_div_x16() local
74 vf1 = _mm256_andnot_ps(_mm256_cmp_ps(vz1, vdenorm_cutoff, _CMP_LT_OS), vf1); in xnn_f16_vsigmoid_ukernel__avx2_rr1_p2_div_x16()
77 vf1 = _mm256_blendv_ps(_mm256_sub_ps(vone, vf1), vf1, vx1); in xnn_f16_vsigmoid_ukernel__avx2_rr1_p2_div_x16()
80 _mm_storeu_si128((__m128i*) (o + 8), _mm256_cvtps_ph(vf1, _MM_FROUND_NO_EXC)); in xnn_f16_vsigmoid_ukernel__avx2_rr1_p2_div_x16()
|
| D | vsigmoid-neonfp16arith-rr2-p2-div-x16.c | 72 float16x8_t vf1 = vdivq_f16(ve1, vd1); in xnn_f16_vsigmoid_ukernel__neonfp16arith_rr2_p2_div_x16() local
75 …vf1 = vreinterpretq_f16_u16(vbicq_u16(vreinterpretq_u16_f16(vf1), vcagtq_f16(vx1, vdenorm_cutoff))… in xnn_f16_vsigmoid_ukernel__neonfp16arith_rr2_p2_div_x16()
81 vf1 = vbslq_f16(vm1, vf1, vsubq_f16(vone, vf1)); in xnn_f16_vsigmoid_ukernel__neonfp16arith_rr2_p2_div_x16()
84 vst1q_f16(o, vf1); o += 8; in xnn_f16_vsigmoid_ukernel__neonfp16arith_rr2_p2_div_x16()
|
| D | vsigmoid-avx2-rr1-p2-rcp-x16.c | 74 __m256 vf1 = _mm256_mul_ps(ve1, vr1); in xnn_f16_vsigmoid_ukernel__avx2_rr1_p2_rcp_x16() local
77 vf1 = _mm256_andnot_ps(_mm256_cmp_ps(vz1, vdenorm_cutoff, _CMP_LT_OS), vf1); in xnn_f16_vsigmoid_ukernel__avx2_rr1_p2_rcp_x16()
80 vf1 = _mm256_blendv_ps(_mm256_sub_ps(vone, vf1), vf1, vx1); in xnn_f16_vsigmoid_ukernel__avx2_rr1_p2_rcp_x16()
83 _mm_storeu_si128((__m128i*) (o + 8), _mm256_cvtps_ph(vf1, _MM_FROUND_NO_EXC)); in xnn_f16_vsigmoid_ukernel__avx2_rr1_p2_rcp_x16()
|
| D | vsigmoid-neonfp16arith-rr2-p2-nr1recps-x16.c | 81 float16x8_t vf1 = vmulq_f16(ve1, vr1); in xnn_f16_vsigmoid_ukernel__neonfp16arith_rr2_p2_nr1recps_x16() local
84 …vf1 = vreinterpretq_f16_u16(vbicq_u16(vreinterpretq_u16_f16(vf1), vcagtq_f16(vx1, vdenorm_cutoff))… in xnn_f16_vsigmoid_ukernel__neonfp16arith_rr2_p2_nr1recps_x16()
90 vf1 = vbslq_f16(vm1, vf1, vsubq_f16(vone, vf1)); in xnn_f16_vsigmoid_ukernel__neonfp16arith_rr2_p2_nr1recps_x16()
93 vst1q_f16(o, vf1); o += 8; in xnn_f16_vsigmoid_ukernel__neonfp16arith_rr2_p2_nr1recps_x16()
|
| D | vsigmoid-neonfp16arith-rr2-p2-div-x24.c | 83 float16x8_t vf1 = vdivq_f16(ve1, vd1); in xnn_f16_vsigmoid_ukernel__neonfp16arith_rr2_p2_div_x24() local
87 …vf1 = vreinterpretq_f16_u16(vbicq_u16(vreinterpretq_u16_f16(vf1), vcagtq_f16(vx1, vdenorm_cutoff))… in xnn_f16_vsigmoid_ukernel__neonfp16arith_rr2_p2_div_x24()
95 vf1 = vbslq_f16(vm1, vf1, vsubq_f16(vone, vf1)); in xnn_f16_vsigmoid_ukernel__neonfp16arith_rr2_p2_div_x24()
99 vst1q_f16(o, vf1); o += 8; in xnn_f16_vsigmoid_ukernel__neonfp16arith_rr2_p2_div_x24()
|
| D | vsigmoid-avx2-rr1-p2-div-x24.c | 81 __m256 vf1 = _mm256_div_ps(ve1, vd1); in xnn_f16_vsigmoid_ukernel__avx2_rr1_p2_div_x24() local
85 vf1 = _mm256_andnot_ps(_mm256_cmp_ps(vz1, vdenorm_cutoff, _CMP_LT_OS), vf1); in xnn_f16_vsigmoid_ukernel__avx2_rr1_p2_div_x24()
89 vf1 = _mm256_blendv_ps(_mm256_sub_ps(vone, vf1), vf1, vx1); in xnn_f16_vsigmoid_ukernel__avx2_rr1_p2_div_x24()
93 _mm_storeu_si128((__m128i*) (o + 8), _mm256_cvtps_ph(vf1, _MM_FROUND_NO_EXC)); in xnn_f16_vsigmoid_ukernel__avx2_rr1_p2_div_x24()
|
| D | vsigmoid-neonfp16arith-rr2-p2-nr1fma-x16.c | 81 float16x8_t vf1 = vmulq_f16(ve1, vr1); in xnn_f16_vsigmoid_ukernel__neonfp16arith_rr2_p2_nr1fma_x16() local
84 …vf1 = vreinterpretq_f16_u16(vbicq_u16(vreinterpretq_u16_f16(vf1), vcagtq_f16(vx1, vdenorm_cutoff))… in xnn_f16_vsigmoid_ukernel__neonfp16arith_rr2_p2_nr1fma_x16()
90 vf1 = vbslq_f16(vm1, vf1, vsubq_f16(vone, vf1)); in xnn_f16_vsigmoid_ukernel__neonfp16arith_rr2_p2_nr1fma_x16()
93 vst1q_f16(o, vf1); o += 8; in xnn_f16_vsigmoid_ukernel__neonfp16arith_rr2_p2_nr1fma_x16()
|
| D | vsigmoid-avx2-rr1-p2-rcp-x24.c | 85 __m256 vf1 = _mm256_mul_ps(ve1, vr1); in xnn_f16_vsigmoid_ukernel__avx2_rr1_p2_rcp_x24() local
89 vf1 = _mm256_andnot_ps(_mm256_cmp_ps(vz1, vdenorm_cutoff, _CMP_LT_OS), vf1); in xnn_f16_vsigmoid_ukernel__avx2_rr1_p2_rcp_x24()
93 vf1 = _mm256_blendv_ps(_mm256_sub_ps(vone, vf1), vf1, vx1); in xnn_f16_vsigmoid_ukernel__avx2_rr1_p2_rcp_x24()
97 _mm_storeu_si128((__m128i*) (o + 8), _mm256_cvtps_ph(vf1, _MM_FROUND_NO_EXC)); in xnn_f16_vsigmoid_ukernel__avx2_rr1_p2_rcp_x24()
|
| D | vsigmoid-avx2-rr1-p2-div-x32.c | 91 __m256 vf1 = _mm256_div_ps(ve1, vd1); in xnn_f16_vsigmoid_ukernel__avx2_rr1_p2_div_x32() local
96 vf1 = _mm256_andnot_ps(_mm256_cmp_ps(vz1, vdenorm_cutoff, _CMP_LT_OS), vf1); in xnn_f16_vsigmoid_ukernel__avx2_rr1_p2_div_x32()
101 vf1 = _mm256_blendv_ps(_mm256_sub_ps(vone, vf1), vf1, vx1); in xnn_f16_vsigmoid_ukernel__avx2_rr1_p2_div_x32()
106 _mm_storeu_si128((__m128i*) (o + 8), _mm256_cvtps_ph(vf1, _MM_FROUND_NO_EXC)); in xnn_f16_vsigmoid_ukernel__avx2_rr1_p2_div_x32()
|
| /external/XNNPACK/src/f32-f16-vcvt/gen/ |
| D | vcvt-neon-x8.c | 51 float32x4_t vf1 = vmulq_f32(vabsx1, vscale_to_inf); in xnn_f32_f16_vcvt_ukernel__neon_x8() local
58 vf1 = vmulq_f32(vf1, vscale_to_zero); in xnn_f32_f16_vcvt_ukernel__neon_x8()
65 vf1 = vaddq_f32(vf1, vreinterpretq_f32_u32(vbias1)); in xnn_f32_f16_vcvt_ukernel__neon_x8()
67 …bine_u16(vshrn_n_u32(vreinterpretq_u32_f32(vf0), 13), vshrn_n_u32(vreinterpretq_u32_f32(vf1), 13)); in xnn_f32_f16_vcvt_ukernel__neon_x8()
68 …anth0 = vcombine_u16(vmovn_u32(vreinterpretq_u32_f32(vf0)), vmovn_u32(vreinterpretq_u32_f32(vf1))); in xnn_f32_f16_vcvt_ukernel__neon_x8()
|
| D | vcvt-scalar-bitcast-x2.c | 49 float vf1 = uint32_as_float(vnonsignw1); in xnn_f32_f16_vcvt_ukernel__scalar_bitcast_x2() local
56 vf1 *= vscale_to_inf; in xnn_f32_f16_vcvt_ukernel__scalar_bitcast_x2()
61 vf1 *= vscale_to_zero; in xnn_f32_f16_vcvt_ukernel__scalar_bitcast_x2()
66 vf1 += uint32_as_float(vbias1); in xnn_f32_f16_vcvt_ukernel__scalar_bitcast_x2()
69 const uint32_t vbits1 = float_as_uint32(vf1); in xnn_f32_f16_vcvt_ukernel__scalar_bitcast_x2()
|
| D | vcvt-neon-x16.c | 57 float32x4_t vf1 = vmulq_f32(vabsx1, vscale_to_inf); in xnn_f32_f16_vcvt_ukernel__neon_x16() local
70 vf1 = vmulq_f32(vf1, vscale_to_zero); in xnn_f32_f16_vcvt_ukernel__neon_x16()
82 vf1 = vaddq_f32(vf1, vreinterpretq_f32_u32(vbias1)); in xnn_f32_f16_vcvt_ukernel__neon_x16()
86 …bine_u16(vshrn_n_u32(vreinterpretq_u32_f32(vf0), 13), vshrn_n_u32(vreinterpretq_u32_f32(vf1), 13)); in xnn_f32_f16_vcvt_ukernel__neon_x16()
88 …anth0 = vcombine_u16(vmovn_u32(vreinterpretq_u32_f32(vf0)), vmovn_u32(vreinterpretq_u32_f32(vf1))); in xnn_f32_f16_vcvt_ukernel__neon_x16()
|
| D | vcvt-scalar-bitcast-x3.c | 51 float vf1 = uint32_as_float(vnonsignw1); in xnn_f32_f16_vcvt_ukernel__scalar_bitcast_x3() local
61 vf1 *= vscale_to_inf; in xnn_f32_f16_vcvt_ukernel__scalar_bitcast_x3()
68 vf1 *= vscale_to_zero; in xnn_f32_f16_vcvt_ukernel__scalar_bitcast_x3()
75 vf1 += uint32_as_float(vbias1); in xnn_f32_f16_vcvt_ukernel__scalar_bitcast_x3()
79 const uint32_t vbits1 = float_as_uint32(vf1); in xnn_f32_f16_vcvt_ukernel__scalar_bitcast_x3()
|
| /external/XNNPACK/src/f32-vscaleexpminusmax/gen/ |
| D | avx2-p5-x16.c | 97 __m256 vf1 = _mm256_fmadd_ps(vt1, vp1, vs1); in xnn_f32_vscaleexpminusmax_ukernel__avx2_p5_x16() local
102 vf1 = _mm256_andnot_ps(_mm256_cmp_ps(vx1, vdenorm_cutoff, _CMP_LT_OS), vf1); in xnn_f32_vscaleexpminusmax_ukernel__avx2_p5_x16()
106 vf1 = _mm256_mul_ps(vf1, vscale); in xnn_f32_vscaleexpminusmax_ukernel__avx2_p5_x16()
110 _mm256_storeu_ps(output + 8, vf1); in xnn_f32_vscaleexpminusmax_ukernel__avx2_p5_x16()
|
| /external/deqp-deps/glslang/Test/baseResults/ |
| D | hlsl.intrinsics.negative.comp.out | 11 0:55 Function Definition: ComputeShaderFunction1(vf1;vf1;vf1;vi1; ( temp void)
74 0:55 Function Definition: ComputeShaderFunction1(vf1;vf1;vf1;vi1; ( temp void)
140 Name 21 "ComputeShaderFunction1(vf1;vf1;vf1;vi1;"
241 21(ComputeShaderFunction1(vf1;vf1;vf1;vi1;): 2 Function None 10
|
| /external/angle/third_party/vulkan-deps/glslang/src/Test/baseResults/ |
| D | hlsl.intrinsics.negative.comp.out | 11 0:55 Function Definition: ComputeShaderFunction1(vf1;vf1;vf1;vi1; ( temp void)
74 0:55 Function Definition: ComputeShaderFunction1(vf1;vf1;vf1;vi1; ( temp void)
140 Name 21 "ComputeShaderFunction1(vf1;vf1;vf1;vi1;"
241 21(ComputeShaderFunction1(vf1;vf1;vf1;vi1;): 2 Function None 10
|