Compression: support FOR & replace Delta with DeltaFOR

[Lloyd-Pottiger]

What problem does this PR solve?

Issue Number: ref #8982

Problem Summary:

What is changed and how it works?

FOR: https://dbms-arch.fandom.com/wiki/Frame_of_Reference_(Compression_Scheme)

Delta is not a compression algorithm, but Delta + FOR be.

Check List

Tests

• Unit test
• Integration test
• Manual test (add detailed scripts or steps below)
• No code

Side effects

• Performance regression: Consumes more CPU
• Performance regression: Consumes more Memory
• Breaking backward compatibility

Documentation

• Affects user behaviors
• Contains syntax changes
• Contains variable changes
• Contains experimental features
• Changes MySQL compatibility

Release note

None

[CalvinNeo]

lgtm

[JinheLin]

What if frame_of_reference is negative?

[Lloyd-Pottiger]

T can only be an unsigned type, frame_of_reference can not be negative.

[Lloyd-Pottiger]

Now we support signed types, but here is just an optimization, when frame_of_reference = 0, no need to do minus.

And for signed types, there is no need to consider overflow, because

int count = std::numeric_limits<int>::min();
count -= std::numeric_limits<int>::max();
count += std::numeric_limits<int>::max();
std::cout << count << "\n"; // equal to std::numeric_limits<int>::min()

[JaySon-Huang]

/hold

[JaySon-Huang]

I think doing this for loop after the for (; i < count; i += (sizeof(__m256i) / sizeof(T))) can bring better performance. Because we can assume the memory address of dst + 0 should be aligned with the cache line, but it is not always aligned when it starts with dst + misaligned_count

[Lloyd-Pottiger]

Good catch, fixed.

[JaySon-Huang]

LGTM

[ti-chi-bot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: CalvinNeo, JaySon-Huang, JinheLin

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

• OWNERS [CalvinNeo,JaySon-Huang,JinheLin]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[Lloyd-Pottiger]

/unhold