Run code on one big cloud machine.

An example for running one function call on a bigger machine with func_cpu and func_ram.

Sometimes you do not want thousands of small machines. You want one bigger machine, with more CPUs and more RAM, to run one heavy step.

In this example we run a single function call on a bigger machine by passing func_cpu and func_ram.

Before you start

Make sure you have already:

1. installed Burla: pip install burla

2. connected your machine: burla login

3. started your cluster in the Burla dashboard

Step 1: Define a function you want to run on a bigger machine

This example reads a file and returns a small summary.

from pathlib import Path

from burla import remote_parallel_map


def summarize_text_file(file_path):
    text = Path(file_path).read_text()
    return {"file_path": file_path, "character_count": len(text)}

Step 2: Run it once, but with more CPU and RAM

remote_parallel_map always takes a list of inputs. To run your function once, pass a list with one item.

func_cpu is the number of CPUs, and func_ram is RAM in GB.

file_path = "/workspace/shared/big-text-file.txt"

results = remote_parallel_map(
    summarize_text_file,
    [file_path],
    func_cpu=8,
    func_ram=64,
)

print(results[0])

This runs summarize_text_file(file_path) on one remote machine that has up to 8 CPUs and 64GB of RAM available for that function call.