Modified and reimplemented FastLanes library in C

[dbeck]

There is a background and intro document in tsl/src/compression/algorithms/fastlanes/README.md

The short version is: this change brings a C implementation of the FastLanes library to Timescale DB. The original idea is from 'The FastLanes Compression Layout: Decoding >100 Billion Integers per Second with Scalar Code' by Azim Afroozeh and Peter Boncz (https://doi.org/10.14778/3598581.3598587)

Our version departs from the original ideas in a few ways:

• instead of having a single 1024 bit virtual register, I introduce multiple smaller register widths, that I call 'tiered FL', the widths are 8, 16, 32, 64, 128 and 256

• we do not have 1024 bit virtual registers

• our implementation relies on C macros, instead of C++ templates and code generation

The README.md file provides more details about the library.

The changelog file is disabled, as this change doesn't provide any useful functionality to the user on its own.
This change provides the basis for later compression algorithms to be introduced.

Disable-check: force-changelog-file

Co-authored-by: Sven Klemm 31455525+svenklemm@users.noreply.github.com
Signed-off-by: David Beck david.beck.priv@gmail.com

[github-actions]

@melihmutlu, @pnthao: please review this pull request.

Powered by pull-review

[codecov]

Codecov Report

❌ Patch coverage is 94.93671% with 32 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
tsl/test/src/test_fastlanes.c	95.17%	1 Missing and 13 partials ⚠️
...on/algorithms/fastlanes/fastlanes_tier_ffor_impl.h	92.70%	6 Missing and 1 partial ⚠️
...ompression/algorithms/fastlanes/fastlanes_sizing.c	94.38%	5 Missing ⚠️
...on/algorithms/fastlanes/fastlanes_tier_pack_impl.h	94.36%	4 Missing ⚠️
.../compression/algorithms/fastlanes/fastlanes_ffor.c	97.14%	1 Missing ⚠️
.../compression/algorithms/fastlanes/fastlanes_pack.c	97.14%	1 Missing ⚠️

📢 Thoughts on this report? Let us know!

[svenklemm]

should this be on tsl side, where all the other compression stuff lives?

[dbeck]

should this be on tsl side, where all the other compression stuff lives?

I'm hesitating on this. It logically belongs to the compression, but:

• releasing this as TSL licensed code is not elegant as this is based on other people's ideas
• and also the compression uses the src/adt/ stuff from the Apache codebase

[dbeck]

I am not going to fix the rest of codecov gaps, because they are either not reasonable or simply wrong

[akuzm]

I am not going to fix the rest of codecov gaps, because they are either not reasonable or simply wrong

I think it makes sense to add Assert(false) to the unreachable default cases/returns. For partial coverage of test assert functions, probably need to add them to the exception list like I did for assertions #9195

[dbeck]

I am not going to fix the rest of codecov gaps, because they are either not reasonable or simply wrong

I think it makes sense to add Assert(false) to the unreachable default cases/returns. For partial coverage of test assert functions, probably need to add them to the exception list like I did for assertions #9195

Thanks for the pointer, I now added the TestAssert.. stuff too, plus the Assert(false) guards.

[akuzm]

Haven't read the algorithm code in detail, but I think we can proceed with merging it.