ARM-software/ComputeLibrary

Issues

• How to set priority of the application running on GPU?

  #1111 opened 10 days ago by qianfei11
  1

• NEPooling3dLayer performance issue

  #1107 opened 6 months ago by alvoron
  8

• How to use the provided SVE GEMM code？

  #1142 opened a month ago by yohuna77777
  3

• Does this provide support for SVE2 on ARM Cortex A76?

  #1144 opened a month ago by Abhranta
  1

• Build does not work with GCC 15+

  #1139 opened 3 months ago by pinskia
  9

• f16 convolution gives the same performance as f32

  #1130 opened 4 months ago by alvoron
  3

• Could not Compile ARM Compute library on Raspberry Pi

  #1143 opened 2 months ago by pegasus-git
  3

• gemm+silu fused operator is not supported in ACL

  #1125 opened 4 months ago by TianyuLi0
  1

• WOA native building error (Exception 0xC0000005 with CpuElementwiseUnaryKernel.cpp)

  #1137 opened 3 months ago by xengpro
  14

• Accuracy problem with F16 NEFullyConnectedLayer

  #1112 opened 5 months ago by allnes
  2

• NEDeconvolutionLayer f16 performance issue

  #1129 opened a month ago by alvoron
  4

• Why is 1D convolution on CPU via NEConvolutionLayer so slow?

  #1119 opened a month ago by poltomo
  7

• CMake lacks `install` target

  #1132 opened 4 months ago by BwL1289
  11

• NEGEMMLowpMatrixMultiplyCore: set_pretranspose_A & set_pretranspose_B support

  #1127 opened 4 months ago by eshoguli
  1

• The result of CLWinogradConvolutionLayer is incorrect.How should I do?

  #1103 opened 2 months ago by superHappy-yo
  1

• Possible bug in `CpuIm2ColKernel.cpp`

  #1140 opened 2 months ago by VladislavZavadskyy
  1

• Regarding Winograd Convolution accuracy

  #1138 opened 2 months ago by vineel96
  19

• Build fails with android-ndk-r27b toolkit

  #1141 opened 2 months ago by culhatsker
  1

• NEGEMM needs `configure` calling for each `run`

  #1136 opened 3 months ago by alvoron
  5

• NEGEMMLowpMatrixMultiplyCore: GEMMLowpOutputStageInfo fusing to speed up inference

  #1120 opened 2 months ago by eshoguli
  1

• NEGEMMLowpMatrixMultiplyCore: QASYMM8 src1 & QASYMM8_SIGNED src2 support

  #1124 opened 3 months ago by eshoguli
  4

• Poor Thread Scaling Behaviour of SGEMM

  #1118 opened 3 months ago by FabianSchuetze
  1

• libarm_compute-static.a export OpenCL API which is unnecessary

  #1128 opened 3 months ago by YiGe-MediaTek
  7

• When used msvc clang-cl, kernel a64_fp16_4x4_3x3 isn't built

  #1135 opened 3 months ago by allnes
  6

• NEGEMMLowpMatrixMultiplyCore: performance issue int8 vs fp16

  #1131 opened 4 months ago by eshoguli
  2

• Compute Library v24.04 is no longer usable on Cortex-A72 if compile with multiisa support

  #1133 opened 3 months ago by malfet
  4

• NEScale f16 performance issue

  #1134 opened 3 months ago by alvoron
  2

• Problem with graph_alexnet.cpp

  #1094 opened 3 months ago by sdavejones
  7

• NEGEMMLowpMatrixMultiplyCore: why QASYMM8 sources are not supported for F32 output

  #1122 opened 4 months ago by eshoguli
  2

• NEConvolutionLayer Segmentation Fault

  #1126 opened 4 months ago by poltomo
  6

• Question about quantization gemm example

  #1102 opened 4 months ago by allnes
  0

• Accuracy problem for NEPoolingLayer - PoolingType::AVG - RoundingType::CEIL - DataLayout::NCHW

  #1113 opened 4 months ago by allnes
  2

• -Werror=maybe-uninitialized in src/core/NEON/kernels/arm_gemm/quantized.cpp

  #1121 opened 4 months ago by robertbcalhoun
  3

• NEMeanStdDevNormalizationLayer returns nans for f16 tensors #2

  #1114 opened 4 months ago by alvoron
  2

• NEGEMMConvolutionLayer is just returning zeros

  #1123 opened 4 months ago by poltomo
  2

• problem with graph_vgg16.cpp

  #1105 opened 5 months ago by RavikumarLav
  2

• File not found conv1_weights.npy in Example graph_resnet50.cpp in computeLibrary

  #1104 opened 5 months ago by RavikumarLav
  1

• Arm NN compute library build failure on Raspberry PI 4

  #1110 opened 5 months ago by pauljainta
  7

• How to parallelize the SGEMM example across many threads?

  #1117 opened 5 months ago by FabianSchuetze
  1

• How can I benchmark GEMMs with `arm_compute_benchmark`?

  #1115 opened 5 months ago by FabianSchuetze
  1

• why L1_cache_size and L2_cache_size are constant value

  #1101 opened 5 months ago by dervon
  1

• Unable to execute MobilenetV2 on Mali-G710 GPU.

  #1108 opened 5 months ago by somasundaram1702
  4

• Depthwise convolution fp16 performance drop

  #1109 opened 6 months ago by alvoron
  4

• CPUInfo::get().has_fp16() returns true on RaspberryPi 4

  #1096 opened 9 months ago by alvoron
  3

• NEMeanStdDevNormalizationLayer returns nans for f16 tensors

  #1095 opened 6 months ago by alvoron
  3

• Uninitialized Regs used causes unintended behavior

  #1093 opened 7 months ago by Nitesh8998
  3

• Compiling error on Raspberry Pi 5:

  #1097 opened 7 months ago by wachhu
  3

• Unit test failure CPU/UNIT/Context/CpuCapabilities on WoA

  #1098 opened 8 months ago by morgolock
  1

• Segfault when running DWC tests on WoA

  #1099 opened 8 months ago by morgolock
  1

• Fix ACL WoA native build compiler errors

  #1100 opened 8 months ago by morgolock
  1