optimize: use SIMD to accelerate json.Compact

bytedance / sonic

A blazingly fast JSON serializing & deserializing library

Apache License 2.0

6.95k stars 339 forks source link

optimize: use SIMD to accelerate json.Compact #414

Open liuq19 opened 1 year ago

liuq19 commented 1 year ago

Compat: now is use the encoding/json API, it is slow and we can use SIMD in native C to accelerate it.

ngqinzhe commented 10 months ago

Hi @liuq19, interested in working on this issue. I'm assuming this is relevant to what you want to achieve here. If this is the correct direction, I will start working on it.

liuq19 commented 10 months ago

Yeah. I think it is a good solution, welcome to try it. And we could use shuffle to compute the space mask. https://github.com/bytedance/sonic/blob/main/native/fastbytes.c#L36

liuq19 commented 5 months ago

hello, is there any update?