Skip to content

Commit babfa8b

Browse files
committed
documenting benchmarks
1 parent ac33f96 commit babfa8b

File tree

2 files changed

+25
-2
lines changed

2 files changed

+25
-2
lines changed

.github/pull_request_template.md

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -1,3 +1,4 @@
1+
We have benchmarks, please consider running them: see our README for details.
12

23
Our CI tests check formatting automating. If such a test fails, please consider running the bash script:
34

README.md

Lines changed: 24 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -429,8 +429,7 @@ abseil : 430.45 MB/s (+/- 2.2 %) 20.52 Mfl
429429
fastfloat : 1042.38 MB/s (+/- 9.9 %) 49.68 Mfloat/s
430430
```
431431

432-
See <https://github.com/lemire/simple_fastfloat_benchmark> for our benchmarking
433-
code.
432+
See the [Benchmarking](#benchmarking) Section for instructions on how to run our benchmarks.
434433

435434
## Video
436435

@@ -486,6 +485,29 @@ You may directly download automatically generated single-header files:
486485

487486
<https://github.com/fastfloat/fast_float/releases/download/v7.0.0/fast_float.h>
488487

488+
## Benchmarking
489+
490+
The project has its own benchmarks with realistic data inputs. Under Linux or macOS,
491+
you can use it as follows if your system supports C++17:
492+
493+
```
494+
cmake -B build -D FASTFLOAT_BENCHMARKS=ON
495+
cmake --build build
496+
./build/benchmarks/realbenchmark
497+
```
498+
499+
Importantly, by default, the benchmark is built in Release mode.
500+
501+
The instructions are similar under Windows.
502+
503+
Under Linux and macOS, it is recommended to run the benchmarks in a privileged manner to get access
504+
to hardware performance counters. You may be able to do so with the `sudo` command
505+
in some cases:
506+
507+
```
508+
sudo ./build/benchmarks/realbenchmark
509+
```
510+
489511
## Packages
490512

491513
* The fast_float library is part of the [Conan package

0 commit comments

Comments
 (0)