Functions — github.com/Maratyszcza/NNPACK

CodeHub MCPcopy Create free account

hub / github.com/Maratyszcza/NNPACK / functions

Functions873 in github.com/Maratyszcza/NNPACK

⨍ Functions 873◇ Types & classes 270

↓ 8 callersFunctionread_perf_counter

bench/perf_counter.h:48

↓ 8 callersMethodtestSXGEMM

test/testers/gemm-ukernel.h:100

↓ 7 callersFunctionpsimd_ifft8_real_f32

src/psimd/fft/real.h:121

↓ 7 callersFunctionwinograd_f6k3_input_transform

src/neon/winograd/f6x6k3x3.h:9

↓ 7 callersFunctionwinograd_f6k3_kernel_transform

src/psimd/winograd/f6x6k3x3.h:81

↓ 7 callersFunctionwinograd_f6k3_output_transform

src/psimd/winograd/f6x6k3x3.h:156

↓ 7 callersFunctionwinograd_f6k3_output_transform

src/scalar/winograd/f6x6k3x3.h:155

↓ 6 callersFunctionallocate_memory

include/nnpack/system.h:95

↓ 6 callersFunctionconv4_relu

* AlexNet conv4 ReLU layer * channels = 256 * image size = 13x13 */

test/models/alexnet.h:127

↓ 6 callersFunctionneon_transpose4x4_inplace_f32

src/neon/transpose.h:6

↓ 6 callersMethodpoolingHeight

test/testers/pooling.h:140

↓ 6 callersMethodpoolingWidth

test/testers/pooling.h:144

↓ 6 callersFunctionprint_options_help

bench/convolution.c:233

↓ 6 callersFunctionprint_options_help

bench/ugemm.c:97

↓ 6 callersFunctionprint_options_help

bench/gemm.c:107

↓ 6 callersFunctionread_memory

bench/memread.c:3

↓ 6 callersFunctionrelease_memory

include/nnpack/system.h:118

↓ 6 callersFunctionsimd_exp

src/x86_64-fma/vecmath/exp.py:32

↓ 5 callersFunctionmaxf

include/nnpack/utils.h:9

↓ 5 callersFunctionpsimd_ifft16_soa_f32

src/psimd/fft/soa.h:257

↓ 5 callersFunctionpsimd_ifft8_soa_f32

src/psimd/fft/soa.h:94

↓ 5 callersFunctionround_up

include/nnpack/utils.h:25

↓ 5 callersFunctionscalar_ifft16_soa

src/scalar/fft/soa.h:542

↓ 5 callersFunctionscalar_ifft8_soa

src/scalar/fft/soa.h:130

↓ 5 callersMethodtestForwardAndInverseAos

* Validates that complex 1D forward FFT + inverse FFT with array-of-structure layout is an identity transformation. */

test/testers/fourier.h:81

↓ 5 callersMethodtestForwardAosSamples

* Validates that output of forward complex 1D FFT with array-of-structure layout matches reference samples. */

test/testers/fourier.h:66

↓ 5 callersFunctionwinograd_f6k3_kernel_transform

src/scalar/winograd/f6x6k3x3.h:80

↓ 4 callersFunctionfft16fc

src/ref/fft/complex.h:220

↓ 4 callersFunctionfft8_bitreverse

src/x86_64-fma/fft-aos.py:37

↓ 4 callersFunctionfft8fc

src/ref/fft/complex.h:112

↓ 4 callersFunctionifft16fc

src/ref/fft/complex.h:321

↓ 4 callersFunctionifft8fc

src/ref/fft/complex.h:161

↓ 4 callersFunctioninterleave

(sequence_a, sequence_b)

src/x86_64-fma/common.py:210

↓ 4 callersFunctionnnp_convolution_input_gradient

include/nnpack.h:589

↓ 4 callersFunctionnnp_convolution_kernel_gradient

include/nnpack.h:612

↓ 4 callersFunctionnnp_convolution_output

Backward compatible implementations for nnp_convolution_*, if we are in C++ mode.

include/nnpack.h:565

↓ 4 callersFunctionnnp_relu_output

src/relu-output.c:50

↓ 4 callersFunctionnnp_relu_output__reference

src/ref/relu-output.c:26

↓ 4 callersFunctionprint_options_help

bench/pooling.c:69

↓ 4 callersFunctionprint_options_help

bench/transform.c:346

↓ 4 callersFunctionprint_options_help

bench/fully-connected.c:112

↓ 4 callersFunctionpsimd_ifft16_dualreal_f32

src/psimd/fft/dualreal.h:144

↓ 4 callersFunctionpsimd_ifft8_dualreal_f32

src/psimd/fft/dualreal.h:113

↓ 4 callersFunctionscalar_ifft16_dualreal

src/scalar/fft/dualreal.h:156

↓ 4 callersFunctionscalar_ifft16_real

src/scalar/fft/real.h:159

↓ 4 callersFunctionscalar_ifft8_dualreal

src/scalar/fft/dualreal.h:127

↓ 4 callersFunctionscalar_ifft8_real

src/scalar/fft/real.h:119

↓ 4 callersMethodtestInputTransform

* Validates that 1D winograd transform for the input matches the linear transformation defined by a matrix. */

test/testers/winograd.h:86

↓ 4 callersMethodtestInputTransform2D

* Validates that 2D winograd transform for the input matches the linear transformation defined by a matrix. */

test/testers/winograd.h:93

↓ 4 callersMethodtestKernelTransform

* Validates that 1D winograd transform for the kernel matches the linear transformation defined by a matrix. */

test/testers/winograd.h:100

↓ 4 callersMethodtestKernelTransform2D

* Validates that 2D winograd transform for the kernel matches the linear transformation defined by a matrix. */

test/testers/winograd.h:107

↓ 4 callersMethodtestOutputTransform

* Validates that 1D winograd transform for the output matches the linear transformation defined by a matrix. */

test/testers/winograd.h:114

↓ 4 callersMethodtestOutputTransform2D

* Validates that 2D winograd transform for the output matches the linear transformation defined by a matrix. */

test/testers/winograd.h:121

↓ 4 callersFunctionvalidate_convolution_arguments

include/nnpack/validation.h:9

↓ 3 callersFunctionas_float

src/x86_64-fma/exp.c:57

↓ 3 callersFunctionas_uint32

src/x86_64-fma/exp.c:47

↓ 3 callersFunctionconv5_relu

* VGG model A conv5 ReLU layer: * channels = 512 * image size = 28x28 */

test/models/vgg-a.h:143

↓ 3 callersFunctionconv8_relu

* VGG model A conv8 ReLU layer: * channels = 512 * image size = 14x14 */

test/models/vgg-a.h:191

↓ 3 callersFunctionfc7_relu

* OverFeat (Fast model) fc7 ReLU layer * channels = 4096 */

test/models/overfeat-fast.h:188

↓ 3 callersFunctionfft32fc

src/ref/fft/complex.h:440

↓ 3 callersFunctionfft4fc

src/ref/fft/complex.h:52

↓ 3 callersFunctionifft32fc

src/ref/fft/complex.h:657

↓ 3 callersFunctionifft4fc

src/ref/fft/complex.h:79

↓ 3 callersFunctionmalloc_with_alignment

bench/convolution.c:17

↓ 3 callersFunctionopen_performance_counter

bench/perf_counter.c:634

↓ 3 callersFunctionprint_options_help

bench/relu.c:83

↓ 3 callersFunctionpsimd_fft16_real_f32

src/psimd/fft/real.h:51

↓ 3 callersFunctionpsimd_fft16_soa_f32

src/psimd/fft/soa.h:155

↓ 3 callersFunctionpsimd_fft8_real_f32

src/psimd/fft/real.h:9

↓ 3 callersFunctionpsimd_fft8_soa_f32

src/psimd/fft/soa.h:38

↓ 3 callersFunctionscalar_fft16_soa

src/scalar/fft/soa.h:245

↓ 3 callersFunctionscalar_fft8_soa

src/scalar/fft/soa.h:7

↓ 3 callersMethodtestComplexToReal

* Validates that complex-to-real 1D FFT produces the same output as complex FFT with array-of-structures layout. */

test/testers/fourier.h:219

↓ 3 callersMethodtestDualComplexToReal

* Validates that dual-sequence complex-to-real 1D FFT produces the same output as two complex-to-real FFTs. */

test/testers/fourier.h:329

↓ 3 callersMethodtestDualRealToComplex

* Validates that dual-sequence real-to-complex 1D FFT produces the same output as two real-to-complex FFTs. */

test/testers/fourier.h:279

↓ 3 callersMethodtestRealToComplex

* Validates that real-to-complex 1D FFT produces the same output as complex FFT with array-of-structures layout. */

test/testers/fourier.h:167

↓ 3 callersFunctionvalidate_fully_connected_arguments

include/nnpack/validation.h:76

↓ 3 callersFunctionvmulsubq_lane0_f32

include/nnpack/arm_neon.h:97

↓ 3 callersFunctionvmulsubq_lane1_f32

include/nnpack/arm_neon.h:105

↓ 3 callersFunctionwinograd_f6k3_input_transform

src/psimd/winograd/f6x6k3x3.h:10

↓ 3 callersFunctionwinograd_f6k3_input_transform

src/scalar/winograd/f6x6k3x3.h:8

↓ 2 callersFunction_mm256_exp_ps

src/x86_64-fma/exp.c:7

↓ 2 callersFunctiondisable_perf_counter

bench/perf_counter.h:40

↓ 2 callersFunctionenable_perf_counter

bench/perf_counter.h:32

↓ 2 callersFunctionfft2fc

src/ref/fft/complex.h:32

↓ 2 callersFunctionfft8_bitreverse

src/x86_64-fma/fft16x16.py:12

↓ 2 callersFunctionifft2fc

src/ref/fft/complex.h:40

↓ 2 callersFunctioninit_performance_counters

bench/perf_counter.c:646

↓ 2 callersMethodkernelSize

test/testers/winograd.h:35

↓ 2 callersFunctionload_ymm_variable

src/x86_64-fma/fft16x16.py:16

↓ 2 callersFunctionmax__psimd

src/psimd/softmax.c:21

↓ 2 callersFunctionmax__scalar

src/psimd/softmax.c:13

↓ 2 callersFunctionmax__scalar

src/scalar/softmax.c:9

↓ 2 callersFunctionmedian_profile

bench/median.c:54

↓ 2 callersFunctionnnp_convolution_output__reference

src/ref/convolution-output.c:58

↓ 2 callersFunctionnnp_fully_connected_inference

src/fully-connected-inference.c:45

↓ 2 callersFunctionnnp_fully_connected_inference_f16f32

src/fully-connected-inference.c:76

↓ 2 callersFunctionnnp_fully_connected_output

src/fully-connected-output.c:221

↓ 2 callersFunctionnnp_fully_connected_output_f32__reference

src/ref/fully-connected-output.c:51

↓ 2 callersFunctionnnp_max_pooling_output

src/pooling-output.c:141

← previous next →101–200 of 873, ranked by callers