[Frontend] Replace IR with bytecode

[erick-xanadu]

Context: Writing uncompressed human readable IR to disk takes longer to parse and dump than using compressed MLIR bytecode. Benchmark results discussed offline.

Description of the Change: Use compressed MLIR bytecode when sending a module to the compiler.

Benefits: Less compilation time (depending on how big the IR produced is)

Possible Drawbacks:

Related GitHub Issues:

[sc-78851]

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 81.44%. Comparing base (4925d37) to head (bdaed9e).
Report is 1 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #1311   +/-   ##
=======================================
  Coverage   81.43%   81.44%           
=======================================
  Files          74       74           
  Lines        8172     8174    +2     
  Branches      861      861           
=======================================
+ Hits         6655     6657    +2     
  Misses       1460     1460           
  Partials       57       57           

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

---

🚨 Try these New Features:

• Flaky Tests Detection - Detect and resolve failed and flaky tests

[dime10]

Thanks Erick! Would also to see whether this makes a difference in compile time.

Btw does this affect the 0_* stage in intermediate files or not really?

[erick-xanadu]

Btw does this affect the 0_* stage in intermediate files or not really?

No.