coffee/fast_float
1
0
Fork 0
mirror of https://github.com/fastfloat/fast_float.git synced 2026-09-14 22:52:47 +08:00
Code Issues Packages Projects Releases Wiki Activity
1,454 Commits 54 Branches 58 Tags 7.2 MiB
Commit Graph

39 Commits

Author SHA1 Message Date
HedgehogInTheCPP
a883161d7e cleanup 2026-08-17 17:47:32 +03:00
HedgehogInTheCPP
3ba6e192ee lindt unfck 2026-08-14 15:09:38 +03:00
HedgehogInTheCPP
a847788226 There's research on the very beginning stage. 
Added improvements and modernize SSE code and enable vectorization for char.
Add more detection of x86 architecture and allow user, for example if MSVC is used set the level manually: used new macro FASTFLOAT_X86_SIMD and available only if FASTFLOAT_SSE_PATCH_REWORK is enabled.
Rename FASTFLOAT_NEON to FASTFLOAT_ARM_NEON for consistency.
 2026-08-13 23:10:11 +03:00
IRainman
71bd4e811f cleanup. 2026-04-30 15:39:27 +03:00
IRainman
c4a979fe0e tests cleanup 2026-01-21 04:26:49 +03:00
IRainman
a9d84bbcb2 style cleanup. 2026-01-21 03:14:46 +03:00
IRainman
61c7b53ea2 small update for benchmark. 2025-12-30 20:42:48 +03:00
IRainman
b7fb05b843 * Added additional compile option FASTFLOAT_TABLE_HACK_CHAR_DIGIT_LUT_DISABLED for improve cache usage in high load. 
* Small optimization in code generation for auto vectorization.
 2025-12-25 14:04:04 +03:00
IRainman
ac1e4dd805 Merge branch 'main' of https://github.com/fastfloat/fast_float 2025-12-24 22:15:34 +03:00
Daniel Lemire
c54e4a7aba updating the deps 2025-12-22 16:05:51 -05:00
Daniel Lemire
0fa058eebb adding bench_ip 
Co-authored-by: Shikhar <shikharish05@gmail.com>
 2025-12-22 11:52:48 -05:00
IRainman
b79c3e5c0a Benchmark are updated. 2025-12-09 22:40:28 +03:00
IRainman
588623e076 * type usage fix for better performance in any hardware. 2025-12-08 23:59:29 +03:00
IRainman
1febc3a070 Fix compilation for older standards 2025-05-05 20:19:46 +03:00
IRainman
d67876e2f7 Merge branch 'main' of https://github.com/IRainman/fast_float 2025-04-30 02:22:10 +03:00
IRainman
4e230e8d24 benchmarks are updated. 2025-04-30 02:19:05 +03:00
IRainman
7bd3c54864 benchmarks are updated. 2025-04-30 01:09:29 +03:00
IRainman
88b3887b52 benchmark cleanup. 2025-04-11 18:15:54 +03:00
IRainman
f8625b6416 fix warnings. 2025-04-10 18:45:05 +03:00
IRainman
8212e9e919 fix warnings in the benchmark. 2025-04-08 01:14:45 +03:00
IRainman
3faba016af Remove PVS. 2025-04-07 23:14:49 +03:00
IRainman
34df2fc76b Tests are updated. 2025-03-28 17:54:06 +03:00
IRainman
6687e734b0 Tests are updated. 2025-03-28 17:49:40 +03:00
IRainman
01e9d35455 Tests updated and fixed. 2025-03-28 16:21:59 +03:00
IRainman
1e3a135b8d reduce register pressure. 2025-03-24 11:45:33 +03:00
IRainman
9688b3bdf7 improvements in benchmark. 
added support for free license for PVS-Studio for open source project (needs recheck by team, I found some strange errors in library, but not fully understand it)
 2025-03-24 11:31:44 +03:00
IRainman
ffd3590a42 benchmarks are improved: back to the cycle metering because when USING_COUNTERS is enabled it's consumed to much resources. 2025-03-24 10:35:55 +03:00
IRainman
3dd3712782 lint 2025-03-12 20:11:11 +03:00
IRainman
681346fde1 Merge branch 'main' of https://github.com/fastfloat/fast_float into fastfloat-main 2025-03-12 17:00:35 +03:00
Daniel Lemire
c0affad8b4
Update benchmark.cpp 2025-03-11 09:19:28 -04:00
Daniel Lemire
7597ca61aa supporting benchmarks under Windows 2025-03-10 19:44:18 -04:00
IRainman
3e86e9a18e #### 
# reading C:/Projects/fast_float/build/benchmarks/data/canada.txt
####
# read 111126 lines
ASCII volume = 1.82777 MB
fastfloat (64)                          :   233.01 MB/s (+/- 2.0 %)    14.17 Mfloat/s      70.59 ns/f
fastfloat (32)                          :   221.31 MB/s (+/- 1.5 %)    13.46 Mfloat/s      74.32 ns/f
UTF-16 volume = 3.65553 MB
fastfloat (64)                          :   460.78 MB/s (+/- 1.4 %)    14.01 Mfloat/s      71.39 ns/f
fastfloat (32)                          :   439.76 MB/s (+/- 2.1 %)    13.37 Mfloat/s      74.80 ns/f
####
# reading C:/Projects/fast_float/build/benchmarks/data/mesh.txt
####
# read 73019 lines
ASCII volume = 0.536009 MB
fastfloat (64)                          :   131.38 MB/s (+/- 0.4 %)    17.90 Mfloat/s      55.87 ns/f
fastfloat (32)                          :   123.03 MB/s (+/- 0.4 %)    16.76 Mfloat/s      59.67 ns/f
UTF-16 volume = 1.07202 MB
fastfloat (64)                          :   259.29 MB/s (+/- 1.5 %)    17.66 Mfloat/s      56.62 ns/f
fastfloat (32)                          :   243.71 MB/s (+/- 1.8 %)    16.60 Mfloat/s      60.24 ns/f

c:\Projects\fast_float\build\benchmarks\Release>
 2025-03-10 05:40:27 +03:00
Daniel Lemire
6f0049a2e7 lint 2025-03-09 15:14:52 -04:00
Daniel Lemire
864e790d22 allowing custom file parsing 2025-03-09 14:27:07 -04:00
IRainman
a22dfc6fcf benchmark are updated, added AST parser emulation for FASTFLOAT_ONLY_POSITIVE_C_NUMBER_WO_INF_NAN. 
Now benchmark only measure parameters for fast_float::from_chars and nothing else.
Copy-past fix.
 2025-03-09 19:55:54 +03:00
IRainman
5097243903 Fix compilation of benchmarks 2025-03-09 04:07:29 +03:00
Daniel Lemire
be9a8453c4 more iterations 2024-12-13 22:54:04 -05:00
Daniel Lemire
8832c532b8 lint 2024-11-25 11:59:48 -05:00
Daniel Lemire
3f2cd66c1c adding actual benchmarks to the project 2024-11-25 11:59:03 -05:00