Add automated benchmarking of the interpretor, possibly as a Github action

[LakshyAAAgrawal]

The first step to optimizing is measurement, and this benchmark would provide a good metric to optimize for

[LakshyAAAgrawal]

Potential solution:

1. Use a benchmark based on https://github.com/sharkdp/hyperfine
2. Each individual program will be treated as a separate benchmark to ensure comparison across commits as benchmarks grow
3. Create 2 GitHub actions, one for pull-requests and one for master-branch-push
   a. Perform relative benchmarking as described in: https://labs.quansight.org/blog/2021/08/github-actions-benchmarks/ for pull-requests
   b. Perform absolute benchmarking and result maintenance on master branch push using https://github.com/marketplace/actions/continuous-benchmark

[LakshyAAAgrawal]

https://github.com/dandavison/chronologer

[LakshyAAAgrawal]

https://www.graalvm.org/graalvm-as-a-platform/language-implementation-framework/Profiling/

[LakshyAAAgrawal]

https://www.graalvm.org/graalvm-as-a-platform/language-implementation-framework/Optimizing/

[LakshyAAAgrawal]

INRIA/spoon#4084

[LakshyAAAgrawal]

A Hyperfine based CAL Benchmark is under experimentation at https://github.com/LakshyAAAgrawal/streamblocks-graalvm/blob/master/.github/workflows/push_to_master.yml and the results available at https://lakshyaaagrawal.github.io/streamblocks-graalvm/dev/bench/hyperfine/

The workflow aims to stabilize the results over multiple jobs, however there is still a lot of regression.