Add support for Faiss SQFP16 Quantization and enable AVX2 Optimization

[naveentatikonda]

Description

This PR includes the following changes:

• Adds a new encoder sq of type fp16 to the Faiss engine which quantizes 32 bit float vectors into 16 bit float values for both hnsw and ivf methods using the Faiss SQFP16 Scalar Quantizer which optimizes memory upto 48% with a very minimal loss of precision. The UX is shown below:

    "encoder":{
      "name":"sq",
      "parameters":{
        "type": "fp16"
      }
    }

• Enabling Faiss AVX2 SIMD optimization, which helps to boost the overall performance on x86 architecture (reduce the query latencies and improve indexing throughput but the vector dimension must be multiple of 8).
• Added a patch script to fix CI build on Linux.
• Update the Faiss submodule to the commit 0013c70 to include the FAISS_ALWAYS_INLINE directives which improves the performance of Scalar Quantizer. Soon, the submodule will be updated to include the SIMD NEON changes after fixing the conflict issues wrt to Multi vector.
• Added a temporary patch file which includes the Faiss SIMD NEON optimization (vector dimension must be multiple of 8) to Scalar Quantizer for ARM architecture

Issues Resolved

#1138

Check List

• New functionality includes testing.
  • All tests pass
• New functionality has been documented.
  • New functionality has javadoc added
• Commits are signed as per the DCO using --signoff

By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.
For more information on following Developer Certificate of Origin and signing off your commits, please check here.

[codecov]

Codecov Report

Attention: 4 lines in your changes are missing coverage. Please review.

Comparison is base (fe592f5) 84.90% compared to head (8e4f404) 85.02%.
Report is 1 commits behind head on main.

Files	Patch %	Lines
.../main/java/org/opensearch/knn/index/Parameter.java	69.23%	3 Missing and 1 partial ⚠️

Additional details and impacted files

@@             Coverage Diff              @@
##               main    #1421      +/-   ##
============================================
+ Coverage     84.90%   85.02%   +0.11%     
- Complexity     1273     1277       +4     
============================================
  Files           167      167              
  Lines          5186     5207      +21     
  Branches        491      493       +2     
============================================
+ Hits           4403     4427      +24     
+ Misses          574      572       -2     
+ Partials        209      208       -1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[jmazanec15]

Enabling Faiss AVX2 SIMD optimization, which helps to boost the overall performance on x86 architecture (reduce the query latencies and improve indexing throughput but the vector dimension must be multiple of 8).

One concern I have with this change is that this could break bwc for users running machines without avx2. What is the plan for handling this?

Update the Faiss submodule.

What did we update the submodule to and why? Is this documented somewhere? I see there is mention of temporary patches - is there a comment on GH explaining this?

[jmazanec15]

Nice job. We should add a validation that with the index description we pass, we correctly create an SQ index (https://github.com/opensearch-project/k-NN/tree/main/jni/tests).

Also, its nice we didnt have to modify our jni wrapper at all.

[jmazanec15]

Why wrap some variables in ${} and not WIN32? Lets be consistent here. Also, can you confirm that WIN32 will refer to the var and not a constant?

[naveentatikonda]

updated the variable. WIN32 is a variable which is True on a Windows machine.

[jmazanec15]

Right but why doesnt it need to be wrapped in ${} like ${WIN32}?

[naveentatikonda]

Not sure. I took this reference, tried and it worked
https://stackoverflow.com/a/17845462

[jmazanec15]

LGTM nice job!

[opensearch-trigger-bot]

The backport to 2.x failed:

The process '/usr/bin/git' failed with exit code 1

To backport manually, run these commands in your terminal:

# Fetch latest updates from GitHub
git fetch
# Create a new working tree
git worktree add .worktrees/backport-2.x 2.x
# Navigate to the new working tree
cd .worktrees/backport-2.x
# Create a new branch
git switch --create backport/backport-1421-to-2.x
# Cherry-pick the merged commit of this pull request and resolve the conflicts
git cherry-pick -x --mainline 1 5f9511b6407691543b1ea0bedd30f3a11f076236
# Push it to GitHub
git push --set-upstream origin backport/backport-1421-to-2.x
# Go back to the original working tree
cd ../..
# Delete the working tree
git worktree remove .worktrees/backport-2.x

Then, create a pull request where the base branch is 2.x and the compare/head branch is backport/backport-1421-to-2.x.