To get the latest version of The Dragon Code: Benchmark
, simply require the project
using Composer:
composer require dragon-code/benchmark --dev
Or manually update require-dev
block of composer.json
and run composer update
console command:
{
"require-dev": {
"dragon-code/benchmark": "^3.0"
}
}
Note
The result of the execution is printed to the console, so make sure you call the code from the console.
use DragonCode\Benchmark\Benchmark;
Benchmark::start()->compare(
fn () => /* some code */,
fn () => /* some code */,
);
Benchmark::start()->compare([
fn () => /* some code */,
fn () => /* some code */,
]);
Benchmark::start()->compare([
'foo' => fn () => /* some code */,
'bar' => fn () => /* some code */,
]);
Result example:
------- --------------------- --------------------
# 0 1
------- --------------------- --------------------
1 0.0101 ms - 64.8Kb 0.0014 ms - 13.7Kb
2 0.0099 ms - 59.5Kb 0.0011 ms - 12.9Kb
3 0.0014 ms - 59.1Kb 0.001 ms - 12.5Kb
4 0.0012 ms - 58.8Kb 0.0011 ms - 12.2Kb
5 0.0012 ms - 58.4Kb 0.0011 ms - 11.8Kb
6 0.0011 ms - 58Kb 0.001 ms - 11.4Kb
7 0.0011 ms - 57.6Kb 0.001 ms - 11Kb
8 0.0011 ms - 57.3Kb 0.001 ms - 10.7Kb
9 0.0012 ms - 56.6Kb 0.001 ms - 10Kb
...
100 0.0011 ms - 14.8Kb 0.0011 ms - 4.1Kb
------- --------------------- --------------------
min 0.001 ms - 14.8Kb 0.001 ms - 4.1Kb
max 0.0101 ms - 64.8Kb 0.0026 ms - 13.7Kb
avg 0.00122 ms - 47.5Kb 0.0016 ms - 4.1Kb
total 0.7998 ms 0.6156 ms
------- --------------------- --------------------
Order - 1 - - 2 -
------- --------------------- --------------------
When measuring the average value among the results, when more than 10 iterations are used, the final data is filtered by peak values. The calculation of the 10% of the lowest and 10% of the highest values is excluded from the total result, thus the final data becomes cleaner and less dependent on any external factors.
By default, the benchmark performs 100 iterations per callback, but you can change this number by calling
the iterations
method:
use DragonCode\Benchmark\Benchmark;
Benchmark::start()
->iterations(5)
->compare(
fn () => /* some code */,
fn () => /* some code */,
);
If the passed value is less than 1, then one iteration will be performed for each callback.
------- --------------------- ---------------------
# 0 1
------- --------------------- ---------------------
1 0.0077 ms - 64.8Kb 0.0015 ms - 57.3Kb
2 0.0023 ms - 59.5Kb 0.0013 ms - 56.5Kb
3 0.0013 ms - 59.1Kb 0.0012 ms - 56.1Kb
4 0.0012 ms - 58.8Kb 0.0011 ms - 55.8Kb
5 0.0011 ms - 58.4Kb 0.0011 ms - 55.4Kb
------- --------------------- ---------------------
min 0.0011 ms - 58.4Kb 0.0011 ms - 55.4Kb
max 0.0077 ms - 64.8Kb 0.0015 ms - 57.3Kb
avg 0.00272 ms - 60.1Kb 0.00124 ms - 56.2Kb
total 0.2453 ms 0.1105 ms
------- --------------------- ---------------------
Order - 2 - - 1 -
------- --------------------- ---------------------
You can also get the number of the current execution iteration from the input parameter:
use DragonCode\Benchmark\Benchmark;
Benchmark::start()
->iterations(5)
->compare(
fn (int $iteration) => /* some code */,
fn (int $iteration) => /* some code */,
);
If you want to see only the summary result of the run time without detailed information for each iteration, then you can
call the withoutData
method, which will display only the
summary information:
use DragonCode\Benchmark\Benchmark;
Benchmark::start()
->withoutData()
->compare([
'foo' => fn () => /* some code */,
'bar' => fn () => /* some code */,
]);
Result example:
------- ------------------- -------------------
# foo bar
------- ------------------- -------------------
min 12.02 ms - 58.4Kb 14.71 ms - 55.4Kb
max 15.66 ms - 64.8Kb 15.67 ms - 57.3Kb
avg 14.65 ms - 60.1Kb 15.17 ms - 56.2Kb
total 73.93 ms 76.31 ms
------- ------------------- -------------------
Order - 1 - - 2 -
------- ------------------- -------------------
Note
If the option to display detailed information is enabled (without using the
withoutData
method) and more than 1000 iterations are requested, then the output of detailed information will be forcibly disabled, since there will be absolutely no point in it with a significantly increasing load on the computer.
By default, the script does not round measurement results, but you can specify the number of decimal places to which rounding can be performed.
For example:
use DragonCode\Benchmark\Benchmark;
Benchmark::start()
->iterations(5)
->round(2)
->compare(
fn () => /* some code */,
fn () => /* some code */,
);
Result example:
------- ------------------ ------------------
# 0 1
------- ------------------ ------------------
1 11.85 ms - 0b 15.22 ms - 0b
2 14.56 ms - 0b 14.94 ms - 0b
3 15.4 ms - 0b 14.99 ms - 0b
4 15.3 ms - 0b 15.24 ms - 0b
5 14.76 ms - 0b 15.14 ms - 0b
------- ------------------ ------------------
min 11.85 ms - 0b 14.94 ms - 0b
max 15.4 ms - 0b 15.24 ms - 0b
avg 14.37 ms - 0b 15.11 ms - 0b
total 73.47 ms 76.03 ms
------- ------------------ ------------------
Order - 1 - - 2 -
------- ------------------ ------------------
In some cases, it becomes necessary to call some action before starting each check cycle so that its time does not fall
into the result of the runtime check.
There is a prepare
method for this:
use DragonCode\Benchmark\Benchmark;
Benchmark::start()
->prepare(fn () => /* some code */)
->compare(
fn () => /* some code */,
fn () => /* some code */,
);
When calling a callback, the name and iteration parameters are passed to it. If necessary, you can use this information inside the callback function.
use DragonCode\Benchmark\Benchmark;
Benchmark::start()
->prepare(fn (mixed $name, int $iteration) => /* some code */)
->compare(
fn () => /* some code */,
fn () => /* some code */,
);
You can also get the number of the current iteration and the result of the execution of the preliminary function from the input parameter:
use DragonCode\Benchmark\Benchmark;
Benchmark::start()
->prepare(fn (mixed $name, int $iteration) => /* some code */)
->compare(
fn (int $iteration, mixed $prepareResult) => /* some code */,
fn (int $iteration, mixed $prepareResult) => /* some code */,
);
------- ------------------ ------------------
# foo bar
------- ------------------ ------------------
1 11.33 ms - 0b 14.46 ms - 0b
2 14.63 ms - 0b 14.8 ms - 0b
3 14.72 ms - 0b 15.02 ms - 0b
4 15.28 ms - 0b 15.04 ms - 0b
N+1 15.03 ms - 0b 15.09 ms - 0b
------- ------------------ ------------------
min 11.33 ms - 0b 14.46 ms - 0b
max 15.28 ms - 0b 15.09 ms - 0b
avg 14.2 ms - 0b 14.88 ms - 0b
total 71.62 ms 75.12 ms
------- ------------------ ------------------
Order - 1 - - 2 -
------- ------------------ ------------------
foo
,bar
- The names of the columns in the passed array. Needed for identification. By default, the array index is used, starting from zero. For example,1, 2, 3,.. N+1
.1
,2
,3
, ...,N+1
- Verification iteration sequence number.11.33 ms
- Execution time of the checked code in one iteration.0b
,6.8Kb
, etc. - The amount of RAM used by the checked code.min
- Minimum code processing time.max
- Maximum code processing time.avg
- The arithmetic mean value among all iterations, taking into account the elimination of 10% of the smallest and 10% of the largest values to obtain a more accurate value through the possible intervention of external factors.total
- The total time and RAM spent on checking all iterations of the code.
This package is licensed under the MIT License.