Rewrite the script in C or C++

C & C++ — incorrectly specified categories?

95 USD

Client's review of cooperation with Roman Smolyakov

Rewrite the script in C or C++

95 USD

Roman left me with a very pleasant impression - he did everything quickly and efficiently. There was simply no chance for problems 🙂

5.0

Roman Freelancer

Reciprocal Feedback

15 April 2025

Freelancer's review of cooperation with Roman Freelancer

Rewrite the script in C or C++

95 USD

Everything is fine with the client! We will continue to work! Feel free to reach out!

5.0

Roman Smolyakov

Reciprocal Feedback

15 April 2025

Project translated automatically. Log in or register, to view the original

Good day

There is a script written in Python and it is very slow, processing 1 MB of data takes more than 25 seconds

1. The problem in the code (technical):

Specifically — in generate_auto_seed() and rank_two_symbols_optimized()

They take ~90% of the total execution time of the script

This is related to:

iterating over each bit in a long array (0 and 1)

calculating a large rank of permutations (a mathematically intensive operation)

using large numbers (often with binomial coefficients)

We need a specialist who:

Understands the algorithmic essence of ranking permutations,

Is skilled in memory optimization, threading, and large numbers,

Can implement low-level block processing of the 0/1 array.

The main essence is speed - the desired speed is 1-3 seconds

In parallel - work on the seed

Send your application if you can implement this.

Thank you

Proposals 5 Discussions 4

Ziadin M.

0 0

Projects -
Rating -
Rating 235

Budget: 100 USD Deadline: 1 day

Hello!

Your script is slow because of two functions:

generate_auto_seed() — bit enumeration, working with large numbers.

rank_two_symbols_optimized() — permutation rank calculation (heavy combinatorial operations).

What I will do:

1. Algorithmic optimization
I will replace the enumeration with an O(n) algorithm for ranking binary sequences (analogous to the Combinatorial Number System).

I will pre-calculate binomial coefficients in O(n²) once, not in a loop.

2. Speedup by 10-100 times
I will rewrite critical parts in Cython (remove Python overhead).

I will add Numba JIT compilation for math.

If needed — SIMD instructions (AVX2) via C extensions.

3. Parallelism and memory
I will split the data into blocks, process in 4-8 threads (multiprocessing).

I will replace lists with numpy.ndarray (dtype=uint8) to save memory.

4. Guarantees
The first 500 KB will be processed in ≤3 sec in 2 days.

The final result is 1-3 sec/MB (or I will refund the money for the improvements).

Why me?

I recently accelerated similar code for seed generation in cryptography (from 30 sec to 0.8 sec).

I am ready to show a prototype on the first 100 lines of your data for free.

What should I send:

Sources of generate_auto_seed() and rank_two_symbols_optimized().

An example of input data (bit array 0/1).

Precise requirements for the environment (Python 3.10? Linux/Windows?).

Starting today. Waiting for your data!

Respectfully,
Ziqo,
P.S. If you have any doubts, I can do a free analysis of 10-20 lines of code and show how to speed them up

Roman Smolyakov

Winning proposal

29 1

Budget: 95 USD Deadline: 1 day

Good day! I am always available to help. Write to me, we will discuss! And don't forget to check out the portfolio — there is a lot of interesting stuff there!