Yi-Yen Chung

> /simde/arm/neon/ld1_x4.h:51:24: error: incompatible pointer types assigning to 'const __fp16 *' from 'const simde_float16 *' (aka 'const _Float16 *') > return vld1_f16_x4(ptr); > ^~~ > /simde/arm/neon/ld1q_x2.h:52:25: error: incompatible pointer types...

> @yyctw Have you been generating your test matrices using a real ARM chip, or Qemu? @mr-c I generated the test matrices only using QEMU. I found that: 1. All...

@SakuragiHnmch Based on the discussions in both [ARM Intrinsics don't build on clang](https://github.com/VectorCamp/vectorscan/issues/21) and [Why does march=native not work on Apple M1?](https://stackoverflow.com/questions/65966969/why-does-march-native-not-work-on-apple-m1), could you please try adding `-march=native` or `-mcpu=apple-m1`...

@SakuragiHnmch I don't know why there are inconsistent results between QEMU and the macOS M1 processor. In branch [issue_1099](https://github.com/yyctw/simde/tree/issue_1099), I added a bug macro to skip those functions on macOS....

I think this issue has some weird points: 1. For `cvt_n.c`, I tested it on the Mac server and encountered the same error as @SakuragiHnmch. However, @mr-c didn't get assertions...