Today (2020年04月24日) I perform tests for chosen solutions for big and small arrays. I tested them on MacOs High Sierra 10.13.6 on Chrome 81.0, Safari 13.1, Firefox 75.0.

• surprisingly for small arrays non-in-place solutions based on slice and reduce (D,E,F) are usually 10x-100x faster than in-place solutions
• for big arrays the in-place-solutions based on splice (AI,BI,CI) was fastest (sometimes ~100x - but it depends of array size)
• for small arrays BI solution was slowest
• for big arrays E solution was slowest

enter image description here

Tests was divided into two groups: in-place solutions (AI,BI,CI) and non-in-place solutions (D,E,F) and was perform for two cases

• test for array with 10 elements - you can run it HERE
• test for array with 1.000.000 elements - you can run it HERE

Tested code is presented in below snippet

Example results for small array on chrome are below

enter image description here

Today (2020年04月24日) I perform tests for chosen solutions for big and small arrays. I tested them on macOS v10.13.6 (High Sierra) on Chrome 81.0, Safari 13.1, and Firefox 75.0.

• surprisingly for small arrays, non-in-place solutions based on slice and reduce (D,E,F) are usually 10x-100x faster than in-place solutions
• for big arrays the in-place-solutions based on splice (AI, BI, and CI) was fastest (sometimes ~100x - but it depends on the array size)
• for small arrays the BI solution was slowest
• for big arrays the E solution was slowest

Enter image description here

Tests were divided into two groups: in-place solutions (AI, BI, and CI) and non-in-place solutions (D, E, and F) and was performed for two cases:

• test for an array with 10 elements - you can run it here
• test for an array with 1,000,000 elements - you can run it here

Tested code is presented in the below snippet:

Example results for a small array on Google Chrome are below:

Enter image description here

Performance

Solutions & Performance

jsfiddle