Daniel Lemire

selectFrom and rearrange (vectorized lookup tables) are going to be slow

2

When parsing strings with SIMD instructions, vectorized table lookup like `vpshufb` (x64) or `vtbl` (NEON) are critically important. They are very cheap (often run in 1 cycle) and very powerful....

fix: add parse_string_if_needed function

6

This is a competing PR vs https://github.com/simdjson/simdjson/pull/2211 by @CarlosEduR The idea is that we avoid copying strings to a string buffer when we do not need to. @CarlosEduR uses a...

Issue https://github.com/cpm-cmake/CPM.cmake/issues/577 is difficult to debug once triggered given that it is undocumented. It is likely to affect virtually all developers who rely on PATCHES with CPM_SOURCE_CACHE unset. If there...

Initial work related to issue https://github.com/fastfloat/fast_float/issues/148 It does not yet build.

https://tc39.es/proposal-arraybuffer-base64/spec/#sec-frombase64 cc @constellation

Currently our 'detect encodings' functions do several passes over the day. A single pass is sufficient. It is not difficult to implement, but doing it well requires being crafty and...

allows you to benchmark against 1 chosen URL.

3

This modifies slightly one of our benchmark program so that we can narrowly benchmark URLs one by one if needed.

more extractor

9

This is a small variation on PR https://github.com/simdjson/simdjson/pull/2247 I am observing a very significant performance regression compared to manually provided code. This is with GCC 12 on a recent x64...

add std::span support (optional, C++20)

5

For people who compile using C++20, we could add optional `std::span` throughout our API *when* C++20 is detected. This is not very difficult, but it would require extra documentation and...

For more convenience. (The additional build time is trivial, these headers compile very fast.)