diff options
| -rw-r--r-- | docs/00_introduction.dox | 16 | ||||
| -rw-r--r-- | src/core/NEON/kernels/NEDirectConvolutionLayerKernel.cpp | 2 | ||||
| -rw-r--r-- | src/core/NEON/kernels/NEPixelWiseMultiplicationKernel.cpp | 1 |
3 files changed, 11 insertions, 8 deletions
diff --git a/docs/00_introduction.dox b/docs/00_introduction.dox index 2df12edc67..72a0ede29a 100644 --- a/docs/00_introduction.dox +++ b/docs/00_introduction.dox @@ -514,14 +514,16 @@ To cross compile the examples with the Graph API, such as graph_lenet.cpp, you n i.e. to cross compile the "graph_lenet" example for Linux 32bit: - arm-linux-gnueabihf-g++ examples/graph_lenet.cpp utils/Utils.cpp -I. -Iinclude -std=c++11 -mfpu=neon -L. -larm_compute_graph -larm_compute -larm_compute_core -lOpenCL -o graph_lenet -DARM_COMPUTE_CL + arm-linux-gnueabihf-g++ examples/graph_lenet.cpp utils/Utils.cpp utils/GraphUtils.cpp -I. -Iinclude -std=c++11 -mfpu=neon -L. -larm_compute_graph -larm_compute -larm_compute_core -lOpenCL -o graph_lenet -DARM_COMPUTE_CL i.e. to cross compile the "graph_lenet" example for Linux 64bit: - aarch64-linux-gnu-g++ examples/graph_lenet.cpp utils/Utils.cpp -I. -Iinclude -std=c++11 -L. -larm_compute_graph -larm_compute -larm_compute_core -lOpenCL -o graph_lenet -DARM_COMPUTE_CL + aarch64-linux-gnu-g++ examples/graph_lenet.cpp utils/Utils.cpp utils/GraphUtils.cpp -I. -Iinclude -std=c++11 -L. -larm_compute_graph -larm_compute -larm_compute_core -lOpenCL -o graph_lenet -DARM_COMPUTE_CL (notice the only difference with the 32 bit command is that we don't need the -mfpu option and the compiler's name is different) +@note If compiling using static libraries, this order must be followed when linking: arm_compute_graph_static, arm_compute, arm_compute_core + To compile natively (i.e directly on an ARM device) for NEON for Linux 32bit: g++ examples/neon_convolution.cpp utils/Utils.cpp -I. -Iinclude -std=c++11 -mfpu=neon -larm_compute -larm_compute_core -o neon_convolution @@ -541,14 +543,16 @@ To compile natively (i.e directly on an ARM device) the examples with the Graph i.e. to cross compile the "graph_lenet" example for Linux 32bit: - g++ examples/graph_lenet.cpp utils/Utils.cpp -I. -Iinclude -std=c++11 -mfpu=neon -L. -larm_compute_graph -larm_compute -larm_compute_core -lOpenCL -o graph_lenet -DARM_COMPUTE_CL + g++ examples/graph_lenet.cpp utils/Utils.cpp utils/GraphUtils.cpp -I. -Iinclude -std=c++11 -mfpu=neon -L. -larm_compute_graph -larm_compute -larm_compute_core -lOpenCL -o graph_lenet -DARM_COMPUTE_CL i.e. to cross compile the "graph_lenet" example for Linux 64bit: - g++ examples/graph_lenet.cpp utils/Utils.cpp -I. -Iinclude -std=c++11 L. -larm_compute_graph -larm_compute -larm_compute_core -lOpenCL -o graph_lenet -DARM_COMPUTE_CL + g++ examples/graph_lenet.cpp utils/Utils.cpp utils/GraphUtils.cpp -I. -Iinclude -std=c++11 L. -larm_compute_graph -larm_compute -larm_compute_core -lOpenCL -o graph_lenet -DARM_COMPUTE_CL (notice the only difference with the 32 bit command is that we don't need the -mfpu option) +@note If compiling using static libraries, this order must be followed when linking: arm_compute_graph_static, arm_compute, arm_compute_core + @note These two commands assume libarm_compute.so is available in your library path, if not add the path to it using -L To run the built executable simply run: @@ -574,8 +578,8 @@ Here is a guide to <a href="https://developer.android.com/ndk/guides/standalone_ - Generate the 32 and/or 64 toolchains by running the following commands: - $NDK/build/tools/make_standalone_toolchain.py --arch arm64 --install-dir $MY_TOOLCHAINS/aarch64-linux-android-4.9 --stl gnustl - $NDK/build/tools/make_standalone_toolchain.py --arch arm --install-dir $MY_TOOLCHAINS/arm-linux-androideabi-4.9 --stl gnustl + $NDK/build/tools/make_standalone_toolchain.py --arch arm64 --install-dir $MY_TOOLCHAINS/aarch64-linux-android-4.9 --stl gnustl --api 21 + $NDK/build/tools/make_standalone_toolchain.py --arch arm --install-dir $MY_TOOLCHAINS/arm-linux-androideabi-4.9 --stl gnustl --api 21 @attention Due to some NDK issues make sure you use g++ & gnustl for aarch64 and clang++ & gnustl for armv7 diff --git a/src/core/NEON/kernels/NEDirectConvolutionLayerKernel.cpp b/src/core/NEON/kernels/NEDirectConvolutionLayerKernel.cpp index 8642a19f39..60a3a1b636 100644 --- a/src/core/NEON/kernels/NEDirectConvolutionLayerKernel.cpp +++ b/src/core/NEON/kernels/NEDirectConvolutionLayerKernel.cpp @@ -1082,7 +1082,7 @@ public: the third thread [16,24] and the fourth thread [25,31]. The algorithm outer loop iterates over Z, P, Y, X where P is the depth/3rd dimension of each kernel. This order is not arbitrary, the main benefit of this - is that we setup the neon registers containing the kernerl's values only once and then compute each XY using the preloaded registers as opposed as doing this for every XY value. + is that we setup the neon registers containing the kernel's values only once and then compute each XY using the preloaded registers as opposed as doing this for every XY value. The algorithm does not require allocating any additional memory amd computes the results directly in-place in two stages: 1) Convolve plane 0 with kernel 0 and initialize the corresponding output plane with these values. diff --git a/src/core/NEON/kernels/NEPixelWiseMultiplicationKernel.cpp b/src/core/NEON/kernels/NEPixelWiseMultiplicationKernel.cpp index 2c90d9aa22..c622d4ffc2 100644 --- a/src/core/NEON/kernels/NEPixelWiseMultiplicationKernel.cpp +++ b/src/core/NEON/kernels/NEPixelWiseMultiplicationKernel.cpp @@ -30,7 +30,6 @@ #include "arm_compute/core/NEON/NEFixedPoint.h" #include "arm_compute/core/TensorInfo.h" #include "arm_compute/core/Validate.h" -#include "arm_compute/runtime/NEON/functions/NEPixelWiseMultiplication.h" #include <arm_neon.h> #include <climits> |