Implement adaptive chunk sizes / adaptive speed benchmark results

[mr-creditcard]

This PR is a continuation of the work done by @0xVavaldi in #729. I have cherry-picked their commits and added my own on top in an attempt to keep attribution in place. It addresses issues #551 & #2295.

The aforementioned PR has a number of requirements set by @s3inlc in the comments which have remained unaddressed thusfar, this PR aims to implement these. The changes include:

• Shrinking chunks: Because of explanations given in comments Adaptive keyspace to reach ideal chunk duration & optimize workload #729 (comment) & Adaptive keyspace to reach ideal chunk duration & optimize workload #729 (comment) we do not want to shrink chunk sizes. This PR addresses this in commit eee602a by early-returning if the multiplier used to adjust the benchmark score is < 1.
• Code organization: The original PR added code directly into APISendProgress.php's switch statement over DHashcatStatus. In commit dbd4b32 this logic is moved to TaskUtils.php instead and is called with a function call.
• Global option to disable functionality: The original PR did not have a global config option to disable the behaviour introduced by this PR. A config option is added in commit bc671d2.

At the time of writing this code is untested, I do plan to do this. But I figured I'd first get the updated code out.

Closes #551 & #2295.