dev: optimized bitshifts by using a lookup table for powers of two

[augustin-v]

Pull Request type

Please check the type of change your PR introduces:

• Bugfix
• Feature
• Code style update (formatting, renaming)
• Refactoring (no functional changes, no API changes)
• Build-related changes
• Documentation content changes
• [✅ ] Other (please describe): dev

What is the current behavior?

The bit shifting operations in the existing implementation are not optimized, leading to potential inefficiencies.

Resolves: #905

What is the new behavior?

• Introduced a new implementation for bit shifting using a lookup table for powers of two.
• Enhanced efficiency by minimizing multiplications and leveraging early returns for large shift values.

Does this introduce a breaking change?

• Yes
• [✅ ] No

---

This change is 

[enitrat]

Great!

For curiosity, if you're interested, you can do some profiling to see how many steps are saved when doing bitshifts. The biggest difference should be when shifting 255 bits as it saves a lot of steps by skipping the expensive exponenciation

https://foundry-rs.github.io/starknet-foundry/snforge-advanced-features/profiling.html

edit: actually, tests don't pass because POW_2 only gives powers of two up to 128. You could introduce a new POW_2_256 variable with powers of two defined up to 2**255

[enitrat]

The easiest way would be to use python to generate cairo code, like so:

def generate_powers_of_two():
    print("pub const POW_2_256: [")
    print("    u256")
    print("    ; 256] = [")

    for i in range(256):
        value = 2 ** i
        hex_value = f"0x{value:x}"
        print(f"    {hex_value},")

    print("];")

if __name__ == "__main__":
    generate_powers_of_two()

[augustin-v]

Thank you for the review, I'll add a new constant as instructed and try profiling as well!

[enitrat]

hi @augustin-v it seems that the changes brought in scope of these PR highlight a memory-management related issue in Cairo Native which crashes our CI runner, thus I will have to keep it open until it is fixed

[augustin-v]

Hi @enitrat, understood. Thank you for the update, and please let me know if there's anything I can do to assist with resolving the issue

[enitrat]

hey @augustin-v i ran some profiling on test_execute_from_outside_eip1559 - which is just an empty execution, so it only performs validation steps:

Before:

After:

It's nearly 10x better :)

[augustin-v]

@enitrat wow it improved much more than I thought it would, bitshift operations looking good