performance.md

Documentation

• Bytecode
• Configuration
• Performance
• Comments //
• Conditions if else end
• Constants true false HIGH LOW INPUT OUTPUT
• Cycles for while next break continue
• Functions function locals return done
• Macros macro
• Numeric variables @ @[]
• Operators + - * / % == != > >= < <= && || & | ^ >> << ++ -- ~ not
• Strings : :[]
• System functions adc read args char cursor delay file close file open file read file write pipe close pipe open pipe read pipe write include index input io open io read io write mem millis number numeric print random restart serial open serial read serial write size stop string
• Unary operators ++ --

Performance

The table below contains the results of the benchmarks developed to compare 20 different programs implemented in BIPLAN and Python:

Language	Bytecode size	Compilation	Interpretation	Startup
BIPLAN	1.12KB	431.58ms	2366.70ms	7ms
Python	8.17KB	1631.59ms	3172.99ms	48ms

1. The BIP bytecode format is 7.32 times more compact than python's pyc
2. biplan compiles 3.7 times faster than python3
3. biplan starts up 6.8 times faster than python3
4. biplan interprets 1.3 times faster than python3

BIPLAN is more efficient than Python and many other interpreted languages. It is particularry well suited for applications that require very fast startup time, efficiency, minimal resource footprint and small binary sizes.

When biplan is compiled using the target fast it is around 30 times faster than python and will compile to bip bytecode that is around 7 times more compact than pyc. To determine the performance of biplan I have implemented benchmarks that compare the compilation time and footprint of 20 different programs implemented in BIPLAN and Python, this is the output on my machine:

 ______________________________________________________________________
|                                       |                |             |
| Benchmark: Compilation time           | biplan compile | python3     |
|_______________________________________|________________|_____________|
|                                       |                |             |
|  1. Hello world!                      | 18.07ms        | 210.54ms    |
|  2. For loop incrementing global      | 17.62ms        | 69.06ms     |
|  3. For loop nested if                | 17.87ms        | 69.90ms     |
|  4. For prints index                  | 17.66ms        | 69.60ms     |
|  5. While loop incrementing global    | 25.53ms        | 79.40ms     |
|  6. While increment and break         | 18.96ms        | 68.59ms     |
|  7. Read elapsed time cyclically      | 17.86ms        | 69.72ms     |
|  8. Get size of string                | 18.02ms        | 71.99ms     |
|  9. Check if numeric                  | 34.97ms        | 69.32ms     |
| 10. Basic math operations             | 23.40ms        | 82.58ms     |
| 11. Bitwise operations                | 17.75ms        | 70.24ms     |
| 12. Bitshift operations               | 18.43ms        | 69.51ms     |
| 13. Function call within condition    | 19.29ms        | 69.93ms     |
| 14. Function calls another function   | 19.00ms        | 70.61ms     |
| 15. Function leap year                | 17.71ms        | 70.12ms     |
| 16. Write character of string         | 19.17ms        | 69.50ms     |
| 17. Read character of string          | 17.78ms        | 69.37ms     |
| 18. For computing prime numbers       | 37.35ms        | 67.91ms     |
| 19. Non-recursive fibonacci           | 17.93ms        | 70.75ms     |
| 20. Factorial 20                      | 17.30ms        | 74.12ms     |
| 21. Pipes                             | 19.91ms        | 68.81ms     |
|_______________________________________|________________|_____________|
                                          431.58ms         1631.59ms

biplan is 3.78 times faster than python3 compile.

 _________________________________________________________________
|                                      |             |            |
| Benchmark: Bytecode size             |  bip size   |  pyc size  |
|______________________________________|_____________|____________|
|                                      |             |            |
|  1. Hello world!                     | 26B         | 279B       |
|  2. For loop incrementing global     | 21B         | 328B       |
|  3. For loop nested if               | 26B         | 303B       |
|  4. For prints index                 | 18B         | 299B       |
|  5. While loop incrementing global   | 21B         | 308B       |
|  6. While increment and break        | 20B         | 272B       |
|  7. Read elapsed time cyclically     | 21B         | 361B       |
|  8. Get size of string               | 40B         | 307B       |
|  9. Check if numeric                 | 110B        | 323B       |
| 10. Basic math operations            | 60B         | 381B       |
| 11. Bitwise operations               | 53B         | 369B       |
| 12. Bitshift operations              | 30B         | 303B       |
| 13. Function call within condition   | 75B         | 464B       |
| 14. Function calls another function  | 92B         | 607B       |
| 15. Function leap year               | 72B         | 536B       |
| 16. Write character of string        | 68B         | 323B       |
| 17. Read character of string         | 71B         | 322B       |
| 18. For computing prime numbers      | 106B        | 520B       |
| 19. Non-recursive fibonacci          | 61B         | 565B       |
| 20. Factorial 20                     | 47B         | 491B       |
| 21. Pipes                            | 105B        | 702B       |
|______________________________________|_____________|____________|
| Total footprint                      | 1.12KB      | 8.17KB     |
|______________________________________|_____________|____________|

Footprint: bip is 7.32 times smaller than pyc.

When biplan is compiled using the target fast it interprets bip bytecode 1.3 times faster than python. To determine the performance of BIPLAN I have implemented benchmarks that compare the execution time of 20 different programs implemented in BIPLAN and Python, this is the output on my machine:

 ________________________________________________________________________________
|                                       |             |             |            |
| Benchmark: Interpretation time        |  BIPLAN     |  Python     | Iterations |
|_______________________________________|_____________|_____________|____________|
|                                       |             |             |            |
|  1. Hello world!                      | 81.55ms     | 135.89ms    | 100000     |
|  2. For loop incrementing global      | 9.62ms      | 71.17ms     | 100000     |
|  3. For loop nested if                | 18.44ms     | 74.04ms     | 100000     |
|  4. For prints index                  | 73.19ms     | 158.94ms    | 100000     |
|  5. While loop incrementing global    | 18.86ms     | 69.60ms     | 10000      |
|  6. While increment and break         | 24.60ms     | 68.55ms     | 10000      |
|  7. Read elapsed time cyclically      | 15.63ms     | 87.61ms     | 100000     |
|  8. Get size of string                | 13.14ms     | 76.14ms     | 100000     |
|  9. Check if numeric                  | 54.20ms     | 71.54ms     | 100000     |
| 10. Basic math operations             | 41.68ms     | 122.57ms    | 100000     |
| 11. Bitwise operations                | 39.43ms     | 113.40ms    | 100000     |
| 12. Bitshift operations               | 22.83ms     | 77.33ms     | 100000     |
| 13. Function call within condition    | 46.56ms     | 86.55ms     | 100000     |
| 14. Function calls another function   | 58.78ms     | 98.93ms     | 100000     |
| 15. Function leap year                | 75.26ms     | 97.41ms     | 100000     |
| 16. Write character of string         | 22.30ms     | 83.99ms     | 100000     |
| 17. Read character of string          | 316.75ms    | 227.03ms    | 100000     |
| 18. For computing prime numbers       | 746.80ms    | 725.51ms    | 100000     |
| 19. Non-recursive fibonacci           | 534.39ms    | 482.48ms    | 100000     |
| 20. Factorial 20                      | 152.60ms    | 244.22ms    | 100000     |
|_______________________________________|_____________|_____________|____________|
                                          2366.70ms     3172.99ms

BIPLAN is 1.34 times faster than Python.