* destroy_gpu_instance() will internally ensure that all vulkan devices are idle before proceeding with destruction.

* fix moltenvk static linking

* pnnx fix some undefined dtype

* fix ncnn convdw1d dynamic weight loading

with `/openmp:llvm` compile option

remove _sum0=_sum0

* Fix `destroy_gpu_instance` crash

* Additional check and clear

* promote vfpv4 for auto fp16 storage conversion

* always report neon and vfpv4 for arm64

* replace storexxx to vsseg2e32_v_f32m1

* refine transpose

---------

Co-authored-by: Xinyu302 <Xinyu302@users.noreply.github.com>

Signed-off-by: Xilin Wu <wuxilin123@gmail.com>

* fix sigbus error when loading fp16 model on armv7

* apply for bf16

* create layer decoupled
* no more virtual public
* allow build test with shared library
* decouple cpu vulkan
* drop old scripts

Which fixes int8 requantization on risc-v

Signed-off-by: Molly Sophia <mollysophia379@gmail.com>

call omp_set_num_threads with zero num_threads is implementation defined

Co-authored-by: nihui <nihui@users.noreply.github.com>
Co-authored-by: michaelcai <michaelcai@tencent.com>

* disable rtti and exceptions for msvc

* warnings--

* erff

* arch sse2 for 32bit build

* enable rtti for cross compiling