Improve performance for async functions

[JulienLecoq]

In some async functions, you're using redundant code that makes the execution code slower (I assume that this is also valid for the deno sdk). I didn't new at first that this redundant code could make the execution slower, but I tested it and here is my experience/results.

In some async functions, you're using the following pattern (from /lib/services/account.js):

async get() {
  let path = '/account';
  let payload = {};

  return await this.client.call('get', path, {
      'content-type': 'application/json',
  }, payload);
}

This pattern contains an unnecessary "async" and "await" notation. This could be simplified by:

get() {
  let path = '/account';
  let payload = {};

  return this.client.call('get', path, {
      'content-type': 'application/json',
  }, payload);
}

From this pov, I decided to benchmark the 2 solutions with the following script:

async function measure_async_func(cb) {
    const start = performance.now()

    await cb()

    const stop = performance.now()
    const timeInSeconds = (stop - start) / 1000;
    const timeRounded = timeInSeconds.toFixed(3)

    return timeRounded
}

function printExecutionTime(funcName, executionTime) {
    console.log(`Time taken to execute function ${funcName}(): ${executionTime} seconds`)
}

function getNumber() {
    return new Promise((resolve, _) => {
        resolve(10)
    })
}

function getNumber1() {
    return getNumber()
}

async function getNumber2() {
    return await getNumber()
}

async function getNumbers1(numbersToGenerate) {
    for (let i = 0; i < numbersToGenerate; i++) {
        await getNumber1()
    }

}

async function getNumbers2(numbersToGenerate) {
    for (let i = 0; i < numbersToGenerate; i++) {
        await getNumber2()
    }
}

async function runTests() {
    const numbersToGenerate = 10_000_000

    const executionTime1 = await measure_async_func(() => getNumbers1(numbersToGenerate))
    printExecutionTime("getNumbers1", executionTime1)

    const executionTime2 = await measure_async_func(() => getNumbers2(numbersToGenerate))
    printExecutionTime("getNumbers2", executionTime2)

    const ratio = executionTime2 / executionTime1
    console.log(`getNumbers1() is ${ratio} faster than getNumbers2()`)
}

runTests()

The results are (with Node v17.4.0):

Time taken to execute function getNumbers1(): 0.928 seconds
Time taken to execute function getNumbers2(): 1.613 seconds
getNumbers1() is 1.738146551724138 faster than getNumbers2()

I haven't tested with the sdk itself, but with a simple script, the second solution (without "async" and "await" notation) seems 1.74 times faster than the first one. I should have tested with the sdk itself, but I don't really have time for that rn.

[jatinmark]

i want to work on this issue , can you assign it to me.