ComputeLibrary.git -

Age	Commit message (Expand)	Author
2021-02-25	Update int8 A55 kernels	Georgios Pinitas
2021-02-25	Avoid division with zero	Georgios Pinitas
2021-02-25	Introduce Context opaque object of the new interface	Georgios Pinitas
2021-02-24	Account Ksections correctly for interleaved and hybrid_indirect gemm kernels	Georgios Pinitas
2021-02-24	Use vector resizing instead of temporary assign in gemm_hybrid_indirect	Georgios Pinitas
2021-02-23	Avoid division with zero on CLGEMM lhs/rhs configuration	Georgios Pinitas
2021-02-23	Remove OpenGL ES support	Manuel Bottini
2021-02-23	Avoid nullptr dereference of vector_sum_col	Georgios Pinitas
2021-02-23	Remove Compute Vision Neon support	Michalis Spyrou
2021-02-23	Remove unreachable code in gemm_hybrid_quantized for computing k_block	Georgios Pinitas
2021-02-22	Close loaded library on OpenCL symbols destruction	Georgios Pinitas
2021-02-22	Comply with Trademark rules for use of Neon, Arm and Mali	Sheri Zhang
2021-02-18	Set CLDirectConvolutionLayerKernel NCHW _border_size to input padding	Giorgio Arena
2021-02-17	Regenerate kernels and update A55 versions	Georgios Pinitas
2021-02-15	Fix data layout retention and handling of leftovers when there is no padding	Michele Di Giorgio
2021-02-10	Revert changes on tensor's strides and fix CLDepthwiseConvolution 3x3 Quantized	Giorgio Arena
2021-02-10	Decouple CpuPoolingKernel data type and data layout	Sheri Zhang
2021-02-09	Make Softmax kernels and operator stateless	Michalis Spyrou
2021-02-09	Fix CLDepthwiseConvolutionLayer 3x3 QASYMM8	Giorgio Arena
2021-02-08	Make memset/copy functions state-less	Sheri Zhang
2021-02-08	Make NEON Pooling kernels and functions state-less	Michele Di Giorgio
2021-02-04	Tweak scheduling use of SQDMULH in quantized AVG pooling	Michele Di Giorgio
2021-02-04	Fix CL Compiler frontend failure for ArgMinMax	Giorgio Arena
2021-02-03	Fix OpenCL direct convolution	Gian Marco Iodice
2021-02-03	Fix segfault in fsrcnn.tflite in GpuAcc	Teresa Charlin
2021-02-03	Make CL Pooling kernels and functions state-less	Michele Di Giorgio
2021-02-03	Add WBSM tuning to CLTuner	Manuel Bottini
2021-02-02	Account for eventual padding in Pooling assembly kernels	Michele Di Giorgio
2021-02-01	Compress OpenCL kernel files using zlib for Android	Georgios Pinitas
2021-02-01	Make data_layout an attribute of the Scale function	Michele Di Giorgio
2021-01-27	Make CL Elementwise Unary kernels and functions state-less	Michele Di Giorgio
2021-01-27	Add broadcasting support in NEPixelWiseMultiplicationKernel for FP16	Pablo Marquez Tello
2021-01-27	Force early exit of threads when output_height=1 and channels=1 in assembly p...	Michele Di Giorgio
2021-01-26	Make CLArithmeticSubtraction kernel and function state-less	Michele Di Giorgio
2021-01-26	Remove iostream header file when not needed	Gian Marco Iodice
2021-01-26	Make CLArithmeticAddition kernel and function state-less	Michele Di Giorgio
2021-01-26	Rename functions/classes for elementwise unary operations	Sang-Hoon Park
2021-01-22	Rename functions/classes for elementwise operations	Sang-Hoon Park
2021-01-22	Fix Scale CTS failures	Sheri Zhang
2021-01-22	Check for unsupported armv7a optimized pooling kernels	Michele Di Giorgio
2021-01-22	Update Fp16 heuristic for Mali-G76 and Mali-G77	Gian Marco Iodice
2021-01-22	CTS failures in Android R and Q in GpuAcc in ArgMinMax	Giorgio Arena
2021-01-21	Make CLFloor and CLActivation kernels and functions state-less	Georgios Pinitas
2021-01-21	Check for FP16 CPU support in optimized pooling kernels	Michele Di Giorgio
2021-01-21	Fix softmax SVE compiling failure without SVE2 support	Sheri Zhang
2021-01-21	Make Sub kernel and operator stateless	Sheri Zhang
2021-01-20	Integrate improved pooling layer on NEON	Michele Di Giorgio
2021-01-20	Make all CL Concatenate kernels and functions state-less	Michele Di Giorgio
2021-01-20	Direct convolution fix for quantized data type	Gian Marco Iodice
2021-01-19	Fix doxygen references to new kernels	Michele Di Giorgio