Xu Jun

JIT compilation speed between MIR and LLVM

21

Hi, I'm greatly attracted by the 100x faster compilation speed shown in your post as we need a fast and lightweight JIT compiler in our [wasm-micro-runtime](https://github.com/bytecodealliance/wasm-micro-runtime). Seems the speed comparison...

### Abstract Currently WebAssembly doesn't support runtime dynamic linking (dlopen, dlsym, dlclose). #444 ``` C typedef int (*TestFunc)(int); int main() { int ret; void *handle; TestFunc test_func; handle = dlopen("app.wasm",...

Introducing `struct.get_indirect` and `struct.set_indirect` opcode

16

### Purpose of this issue Hi, in [GC-05-16 meeting](https:github.com/WebAssembly/meetings/blob/main/gc/2023/GC-05-16.md) we have introduced our idea of adding `struct.get_indirect` and `struct.set_indirect` opcode to support language semantics like `TypeScript's interface` in WasmGC, we...

WASI proposal for dynamic typing (WASI-dyntype)

8

This issue is to introduce the idea of proposing `WASI-dyntype` APIs. ## Introduction The type system of WebAssembly is entirely static in nature, precluding the direct compilation of dynamic programming...

question about command timeout

1

Hi, I meet a problem that the chip enter `COMMAND_TIMEOUT` state. This happens when I use `llcc68_write_buffer` with data longer than 4 bytes, the datasheet said that `COMMAND_TIMEOUT` is caused...

4x16s4 fp32-gemm kernel have better performance than default(5x16) kernel for meteor lake

1

XNNPACK by default uses 5x16 fp32-gemm kernel for `x86_fma3`, but we found that 4x16s4 kernel shows better performance on `meteor lake` CPU (`Intel(R) Core(TM) Ultra 7 155H`) | benchmark |...

detect Meteor Lake CPU model

10

This PR add support for detecting Intel Meteor Lake CPU through its `model number` and `extended model`.

enable AVX_VNNI_INT8 instruction for qs8-qc8w-gemm/igemm

1

The new `AVX_VNNI_INT8` instruction can avoid XOR operation in qs8-qc82 gemm/igemm kernel, resulting ~3% performance improvement on mobilenet v1 and v2 int8 models. ``` bash ------------------------------------------------------------------------------------------------------- Benchmark Time CPU Iterations...