admin/simdjson-simdjson
1
0
Fork 0
mirror of https://github.com/simdjson/simdjson synced 2026-06-08 17:27:07 +00:00
Code Issues Packages Projects Releases Wiki Activity

Runtime dispatch

Browse Source 
* Attempt 1 - fn targeting

GCC won't work with templates with different targets, need to specialize all the way up the call stack.

* Compiles properly with cmake. Does not with the Makefile.

* Compilation works with Makefile

* instruction_set changes to architecture

* some aesthetic changes

* fix amalgation and tests + aesthetic changes

* This now compiles and passes tests under CLANG

* Minor correction.

* Trying to make it work on ARM

* Adding missing namespace

* Missing bracket

* Fixing minor compilation issues.

* Getting parse to use runtime dispatch

* Fixing amalgamation script.

* Making sure that NEON is supported.

* Fixing typo

* Merging https://github.com/lemire/simdjson/pull/229

* Manual merge of
https://github.com/lemire/simdjson/pull/229
by @jkeiser  (second part)

* Trying another way.

* Removing the paral.

* Fixing the make file

* Let us make the practice run long enough.

* Resolved the awful slowness.

* Cleaning the README.md

* With runtime dispatching, we should not need flags anymore.

* Changing isa detection file's name + fixing typos.
This commit is contained in:
Daniel Lemire
2019-07-28 22:46:33 -04:00
committed by ioioioio
parent bcabdfc1ae
commit eba02dc1b9
39 changed files with 4409 additions and 3663 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/jsonminifier.cpp
+24
View File
@@ -63,6 +63,30 @@ size_t jsonminify(const unsigned char *bytes, size_t howmany,
#include <cstring>
namespace simdjson {
// some intrinsics are missing under GCC?
#ifndef __clang__
#ifndef _MSC_VER
static __m256i inline _mm256_loadu2_m128i(__m128i const *__addr_hi,
__m128i const *__addr_lo) {
__m256i __v256 = _mm256_castsi128_si256(_mm_loadu_si128(__addr_lo));
return _mm256_insertf128_si256(__v256, _mm_loadu_si128(__addr_hi), 1);
}
static inline void _mm256_storeu2_m128i(__m128i *__addr_hi, __m128i *__addr_lo,
__m256i __a) {
__m128i __v128;
__v128 = _mm256_castsi256_si128(__a);
_mm_storeu_si128(__addr_lo, __v128);
__v128 = _mm256_extractf128_si256(__a, 1);
_mm_storeu_si128(__addr_hi, __v128);
}
#endif
#endif
// a straightforward comparison of a mask against input.
static uint64_t cmp_mask_against_input_mini(__m256i input_lo, __m256i input_hi,
__m256i mask) {