error: unknown type name '__Int8x8_t' 背景: 编译arm版本的auto apollo代码遇到的编译错误。目标机是Nvidia Jetson Xiavier设备。apollo代码中用到cudda代码,包含了arm_neon.h文件报出的错误。 前提: 在gcc4.8版本能够顺利编译过。到gcc 5以后,arm_neon.h代码结构发生变化,就编译不过,网上看到好多用到opencv...
编译arm版本,找不到一些neon的基础类型例如 '__Int8x8_t'等,原因是头文件没用clang自己的。 见:https://www.cnblogs.com/Forwarderz/p/12656058.html https://github.com/clangd/clangd/issues/1588 现在我是用"C:/Microsoft/AndroidNDK64/android-ndk-r16b/toolchains/llvm/prebuilt/windows-x86_64/lib64/...
ARM compiler keeps reporting "unknown type name '__Int8x8_t'" Could you please provide more information about compiler command and clangd version with reduced code snippet if possible? lichjian changed the title arm 编译器一直报 unknown type name '__Int8x8_t' arm compiler reporting unknown typ...
/home/orangepi/opt/MNN/source/backend/cpu/CPUSoftMaxInt8.cpp:226:56: error: cannot convert ‘uint16x8_t’ to ‘int16x8_t’ 226 | int8x8_t mask = vmovn_s16(vcgeq_s16(input_diff_s16, diff_min_s16)); | ~~~^~~~ | | | uint16x8_t In file included from /home/orangepi/o...
问Tensorflow Lite arm64错误:无法转换‘const int8x8_t？ENMicrocontrollers (MCUs) are the tiny ...